机器学习学术速递[11.11]

AI 总结：本文提出基于互正交性的排列式预训练模型合并方法，可在更小宽度下实现线性模式连通性，还能合并不同数据分布训练的模型，合并预测近似原始模型的logit集成结果。

链接：https://arxiv.org/abs/2611.13016

机构：Tohoku University（东北大学）； NTT Computer and Data Science Laboratories（NTT计算机与数据科学实验室）

作者：Akira Ito, Daiki Chijiwa, Atsutoshi Kumagai

英文摘要：Ito et al. (2026) showed that sufficiently wide, independently trained neural networks can exhibit linear mode connectivity (LMC) with softmax temperature calibration, even without permutation. LMC means that a low-loss linear path connects the models' parameters. This behavior is explained by layerwise exponentially weighted connectivity (LEWC), whereby each merged layer outputs an exponentially weighted sum of the original models' corresponding representations, yielding ensemble-like predictions. However, prior work observed LEWC with substantial widening, such as 8192 units per hidden layer for MLPs and $16\times$ width for VGG-11, raising the question of whether such widths are necessary. We therefore search for permutations that encourage reciprocal orthogonality (RO), which suppresses interference between each model's weights and the other model's representations. This allows both models' computations to coexist in the merged network, promoting LEWC at smaller widths. We formulate the search as independent layerwise quadratic assignment problems and approximately solve their continuous relaxations using Frank--Wolfe. Experiments show that our method facilitates LMC with less width expansion. It also merges models trained on different data distributions, where weight matching, which maximizes weight similarity through permutations, struggles. The merged predictions approximate those of a logit ensemble of the original models.

AI 总结：该研究提出基于Broyden方法的并行拟牛顿法，通过对角加低秩近似雅可比块实现非线性状态空间模型的高效并行评估，在低秩交互或雅可比评估昂贵场景下收敛更快。

链接：https://arxiv.org/abs/2611.13523

机构：Stanford University（斯坦福大学）

作者：Ian Christopher Tanoh, Scott W. Linderman

英文摘要：Nonlinear state space models are typically evaluated sequentially, limiting their ability to exploit modern parallel hardware. Recent work has shown that state space models can be evaluated in parallel by reformulating evaluation as a root-finding problem, which can be solved with a parallel form of Newton's method. However, exact Newton methods require expensive multiplication of dense Jacobian matrices, and quasi-Newton methods based on diagonal approximations, while more scalable, are often slow to converge because they neglect interactions across state dimensions. We introduce a novel quasi-Newton approach based on Broyden's method, which captures coupling terms with diagonal-plus-low-rank approximations to blocks of the Jacobian. We construct the approximations from trajectory secants, avoiding automatic differentiation while retaining the favorable scaling of diagonal approximations and remaining compatible with parallel scan through rank compression. We prove finite-step recovery of the exact trajectory under the compressed scan, and show that convergence is tracked by the compression error and by the block approximation error along the current trajectory error, rather than in spectral norm. Empirically, we show that our method is advantageous when interactions are approximately low rank or when Jacobian evaluation is expensive: on low-rank RNNs, parallel MCMC algorithms, and denoising diffusion models, it converges in fewer iterations than other quasi-Newton methods and is the fastest parallel method when Jacobian evaluation is expensive or state dimensions are large.

AI 总结：该研究提出CP-MLPs张量框架，揭示绑定与非绑定MLP块的张量秩差异，实验表明非绑定块在匹配预算下具宽度效率优势，代码已公开。

链接：https://arxiv.org/abs/2611.13647

机构：Mila – Quebec AI Institute（米拉-魁北克人工智能研究所）； Université de Montréal（蒙特利尔大学）

作者：Md Rifat Arefin, Farzaneh Heidari, Irina Rish, Guillaume Rabusseau

英文摘要：Transformer feed-forward blocks combine several design choices: squared activations, gating, and tied versus untied parameter sharing, whose separate contributions to expressivity remain poorly understood. We introduce CP-MLPs, a tensor framework in which each multiplicative unit contributes a rank-1 interaction with three roles: detecting an input feature, reading a value, and writing to the output. Tied gated blocks such as ReLU$^2$ force the detector and value roles to share the same direction; untied gated blocks such as ReGLU and SwiGLU let them differ. We show that this distinction yields exact, degree-wise tensor-rank separations: untied detector-value units are strictly more width-efficient for indefinite quadratic interactions and role-asymmetric monomials, while tied units are structurally matched to symmetric pure-power targets. The width advantage of untying therefore depends on the target structure. Experiments show loss plateaus below the exact-representation width, but these plateaus alone do not establish a positive approximation gap: arbitrarily accurate approximation is governed by border rank. In trained gated MLPs, untied blocks retain an advantage in selected comparisons at matched input-projection budget. Code is available at this https URL.

AI 总结：本文提出开源运行时tritify，以精确整数算术实现模型前向传播，验证了其跨硬件可复现、决策保真、位置独立的特性，衍生出可验证分布式推理方案。

链接：https://arxiv.org/abs/2611.13730

机构：Pelloran Inc.（佩洛兰公司）

作者：Sayed Amir Karim (Pelloran Inc.)

英文摘要：Optimized floating-point inference is not reproducible across environments: the same model on different hardware, compilers or thread counts can produce different bits. We present tritify, an open-source runtime that executes the forward pass of real models -- decoder-only language models, Mixture-of-Experts decoders, and the Segment Anything 2.1 image/video stack -- with the entire run-time path in exact integer arithmetic; floating point is confined to offline quantization and the float reference used for verification. We make three separate claims and verify each. (1) Cross-hardware reproducibility: the integer path produces bit-identical output run to run, across CPU instruction-set architectures, and across CPU/GPU, carried by a reproducing SHA-256 of the output. (2) Decision fidelity: the integer path reproduces the float reference's decision on the models and inputs tested, with the numeric exceptions measured; per-layer fidelity through sixteen layers of OLMoE-1B-7B and of a 106-billion-parameter MoE shows no fidelity cliff. (3) Location independence: a model's computation can be split across instruction sets, CPU and GPU, and networked machines with zero effect on the output bits. GLM-4.5-Air (106,852,251,264 parameters, 427 GB of integer codes) generated its answer across three machines, none of which can hold its integer codes in RAM, bit-identically to a single-host reference. The consequence is verifiable distributed inference: an untrusted server or a fleet of devices runs the model while an open scalar oracle re-derives the exact bytes; a guarantee against a dishonest server is future work. We do not claim a per-device speed-up, bit-equality of floating-valued intermediates, or confidentiality. Code: this https URL

AI 总结：本文探究因果神经元回路的独特性与可迁移性，通过对Qwen、Llama、Gemma等模型的神经元干预实验，发现少量神经元即可改变模型输出，且效应仅在同一链重述时可迁移。

链接：https://arxiv.org/abs/2611.13885

作者：Ömer Tekin Yavuz (İletişim Yazılım)

英文摘要：Interpretability demonstrations often show that intervening on a group of neurons changes a model's answer. A recent example removes an 18-neuron group in Qwen3-32B and the model answers "Simba" instead of "Scar" to a two-hop Lion King riddle. Such a demonstration establishes that the group matters, but not how much of it is necessary, whether other groups could do the same, or whether the effect carries over to other inputs. We ask: when intervening on a group of neurons changes a model's answer, how much of that group is actually necessary? For every selected group we measure all 4,096 subsets of 12 MLP neurons. On a two-alternative capital task (Qwen3-0.6B), zero-ablating 12 selected neurons lowers the answer margin by 27% while random neurons barely change it, yet replacing the same neurons with their values from other questions also barely changes it: their activations are nearly constant across questions, and we find no evidence that they carry question-specific content. On two-hop relation chains, patching 12 neurons with activations from the first-hop question shifts the answer to the bridge entity (from Scar to Simba): the greedy answer starts with the bridge in 9/11 chains for Qwen3-1.7B, 9/11 for Llama-3.2-1B and 8/10 for Gemma-3-1B; random neurons never do. The full group is rarely necessary: a few neurons often suffice, and in many chains disjoint subsets each flip the answer on their own, in an exploratory analysis most often when the model is less confident. Group effects transfer to rephrasings of the same chain but are much weaker on other tested chains. We study neuron-level causal interventions and alternative subsets; we do not recover the full connectivity of circuits.

AI 总结：本文提出基于球形图网格Transformer架构的AI框架PARDIG，利用稀疏卫星观测生成高精度近实时全球降水估算，性能优于现有主流产品，相关数据集公开可用。

链接：https://arxiv.org/abs/2611.13941

机构：Google DeepMind（谷歌DeepMind）； Google Research（谷歌研究院）

作者：Guy Shalev, Stephan Rasp, Boris Babenko, Andrew El-Kadi, Mohammed Alewi Hassen, Samier Merchant, Fred Zyda, Dominic Masters, Ferran Alet, Salva Rühling Cachay, Ilan Price, Akib Uddin, Nofar Peled Levi, Ben Gaiarin, Aaron Bell, Peter Battaglia, Alvaro Sanchez-Gonzalez

英文摘要： Estimating precipitation at the global scale is vital for Earth system science and disaster response, yet observational coverage and measurement precision remain constrained. Existing solutions provide options that trade off desirable traits such as skill, latency, geographical coverage, temporal coverage, and spatial and temporal resolution. We present PARDIG, an artificial intelligence framework based on a spherical graph-mesh transformer architecture that produces near-real-time global precipitation estimates. PARDIG is trained on dense multi-modal inputs, including atmospheric state analyses and multi-spectral geostationary satellite imagery, to predict sparse, high-precision precipitation measurements from spaceborne radar. At inference, the trained model is applied densely across the globe at $0.1^\circ$ spatial and 15-minute temporal resolution, producing complete near-real-time precipitation fields from 2016 to the present, alongside a historical variant extending the record back to 1999. Evaluation against three independent radar networks (over the US, Europe, and Japan) and against over 36,000 surface stations worldwide shows that PARDIG consistently outperforms IMERG V07 (Final and Early), MSWEP V3, ERA5, and HRES across precipitation thresholds and regions, with the largest improvements at extreme intensities. Case studies of extreme precipitation events demonstrate that PARDIG successfully captures their spatial structure and timing. The improvements also extend to forecasting: WeatherNext 3, a global weather model, produces more skillful precipitation forecasts when trained to predict PARDIG. Both the historical and near-real-time PARDIG datasets are openly available under a CC BY 4.0 license.

AI 总结：该研究针对混合LLM服务中循环状态内存冗余、解码成本高的问题，提出SinkState方法，通过自适应低秩存储与局部SVD等技术，在保持精度的同时显著降低内存占用并提升解码速度。

链接：https://arxiv.org/abs/2611.14246

机构：Rutgers University（罗格斯大学）； University of Sydney（悉尼大学）； University of South Florida（南佛罗里达大学）

作者：Haolin Jiang, Weihang Li, Zhongzhu Zhou, Jinwen Pan, Santosh Pandey, Hang Liu

英文摘要：Hybrid models combine softmax attention with linear attention, which compresses history into fixed-size recurrent states to balance model quality and inference efficiency. However, linear attention has not pushed this compression to its limit: we find that the queries, keys, and values of linear layers concentrate in a stable, low-dimensional subspace, a phenomenon we call the linear attention sink, so each state is effectively low-rank, with a rank that varies across states. Yet existing serving systems keep these redundant states dense, reading and rewriting them in full at every decoding step, so their memory traffic grows with the batch size and dominates the decoding cost. Thus SinkState stores each recurrent state in an adaptive low-rank form throughout inference with buffered updates and a local SVD restricted to the top subspace. SinkState further consolidates the remaining components asynchronously on the CPU. On four hybrid models with Mamba-2 and Gated DeltaNet (GDN) layers, from 4B to 120B parameters, SinkState reduces the peak memory of recurrent states by up to 2.6x (1.9x on average) and speeds up decoding by up to 1.58x (1.25x on average) over Mamba-SSM and FLA, while maintaining accuracy.

AI 总结：本文通过压力测试MOMENT等4种TSFMs，量化其采样频率不变性，明确该属性与信号带宽差值的关系，为从业者提供TSFMs适用场景指导。

链接：https://arxiv.org/abs/2611.14425

机构：Bosch Research, USA - Bosch Center for Artificial Intelligence（博世美国研究中心 - 博世人工智能中心）； School of Interactive Computing, Georgia Institute of Technology, USA（佐治亚理工学院交互计算学院）

作者：Sourish Gunesh Dhekane, Abinaya Kumar, Torsten Woertwein, Yan-Bo Lin, Thomas Ploetz, Shabnam Ghaffarzadegan

英文摘要：Time-Series Foundation Models (TSFMs) are increasingly used in sensor applications, where sampling frequency (SF) varies across sensors, devices, and operating modes (e.g., power saving). Ideally, TSFMs must remain effective during such SF changes, provided the SF retains task-critical bandwidth. In this paper, we stress-test 4 prominent TSFMs: MOMENT, Chronos-2, Moirai-MoE, and TimesFM, to quantify their SF-invariance and investigate whether their behavior tracks the resulting loss of information bandwidth rather than the SF value itself. Our stress-test probes 3 dimensions: (i) the shift in embeddings, (ii) the degradation in forecasting performance, and (iii) the degradation in downstream performance, quantified by three novel metrics: FSI, FDI, and CDI, respectively. To our knowledge, this is the first study to place SF-invariance at the core of TSFM evaluations. Our results characterize: (i) how SF-invariance depends on the margin between signal's native-SF and the bandwidth its task requires, (ii) how much of that invariance the choice of TSFM explains, and (iii) how much of TSFM's performance degradation under a lower SF reflects information loss rather than a limitation of their abilities. These findings reframe SF-invariance as a measurable, signal-aware property of the (signal X TSFM) pair, and guide practitioners on when TSFMs display SF-invariance.

AI 总结：该研究针对语言模型训练后策略优化的覆盖度崩溃问题，提出近似策略镜像下降（APMD）方法，证明含参考重置的PPO、GRPO等算法为覆盖度保持算法，明确了正则化等因素对训练后成本的影响。

链接：https://arxiv.org/abs/2611.14467

机构：Kempner Institute at Harvard University（哈佛大学 Kempner 研究所）； UIUC（伊利诺伊大学厄巴纳-香槟分校）； University of California, San Diego（加利福尼亚大学圣迭戈分校）； Microsoft Research（微软研究院）

作者：Audrey Huang, Nan Jiang, Sham M. Kakade, Sadhika Malladi, Akshay Krishnamurthy, Dylan J. Foster

英文摘要： Policy optimization is central to language model post-training, yet principled guidance remains limited. We revisit its theoretical foundations to identify algorithms that learn near-optimal policies using total samples and gradient steps governed by the base model's response coverage, a measure of how readily it generates optimal responses. We call such algorithms coverage-preserving. This guarantee captures the goal of amplifying capabilities accessible through sampling from the base model. (1) Coverage collapse. REINFORCE, the basis for PPO and GRPO, can drive the probability of optimal responses to near zero, even as reward improves. This coverage collapse stalls optimization, and convergence can require exponentially many steps despite good initial coverage. The failure arises even with exact gradients and persists under KL regularization to the base model. (2) Reference resets. We show that repeatedly resetting the KL reference to the most recent policy mitigates coverage collapse. We formalize this intervention in Approximate Policy Mirror Descent (APMD), a meta-algorithm that approximately solves a sequence of KL-regularized RL subproblems. APMD enables stronger regularization, making each subproblem easier to solve, while reference resets allow convergence to global optimality. (3) Coverage-preserving algorithms. For softmax linear policies over pretrained features, we use APMD to establish the first global convergence guarantees showing that widely used families of post-training algorithms are coverage-preserving. These include stylized PPO and GRPO variants with explicit KL regularization and reference resets, and squared-loss algorithms that implement them implicitly, including those used to train the Kimi models. These results clarify how regularization, clipping, and objective choice affect the sampling and optimization costs of post-training.

AI 总结：本文针对Brugge油田历史拟合的歧义问题，提出基于图生成模型的集成方法，采用带专用生成器的联合潜在空间与协方差矩阵自适应进化策略，实现了与基准方法相当的拟合效果。

链接：https://arxiv.org/abs/2611.12927

机构：Tomsk Polytechnic University（托木斯克理工大学）； Skolkovo Institute of Science and Technology（斯科尔科沃科学技术研究院）； Heriot-Watt University（赫瑞瓦特大学）

作者：Gleb Shishaev, Anton Shirokov, Shadfar Davoodi, Mohammed Al-Shargabi, Vasily Demyanov, Daniel Arnold

英文摘要：Assisted history matching remains non-unique because many geological realisations can reproduce equivalent production responses with substantial differences between individual well matches. This ambiguity increases when a field comprises several depositional settings and modelling strategies. The paper applied a graph autoencoder approach to history matching of a benchmark Brugge field that comprises different depositional zones. Each stratigraphic zone is modelled with its dedicated generative model that comprises a joint latent space for optimisation history matching loops. Graph-based generators operate directly on the active cells of the corner-point grid, so its irregular geometry is handled without conversion to a regular lattice. The optimisation search in the 160-dimensional joint latent space is done with the Covariance Matrix Adaptation Evolution Strategy. The multi-component objective function includes static well data mismatch, dynamic production mismatch, and a geological realism penalty based on the Decoder induced latent metric. Latent space diagnostics with topological data analysis are applied prior to history matching to test whether one generator is sufficient or whether a multiple generator formulation is required. The results show that the zone-based formulation gives coherent latent spaces, preserves the main porosity and permeability relations that vary from one zone to another, and supports history matching while keeping the solutions within the learned geological manifold. The matched models reproduce field oil and water production within the prior uncertainty range and provide a comparable well-level match to previously published Brugge studies based on Ensemble Kalman Filter, Particle Swarm Optimisation with principal component analysis of parameterisation, and Multiple Kernel Learning with Particle Swarm Optimisation.

AI 总结：本研究提出MC-JEPA模型，在11年中国A股数据上，该模型在多类金融目标上优于基线模型，实现最佳RankIC与投资组合表现，且无需标签即可按市场贝塔排序股票。

链接：https://arxiv.org/abs/2611.13572

机构：The University of Hong Kong（香港大学）； Grace Investment Machine（格雷斯投资机器）

作者：Yuqi Wang, Fengyuan Liu, Mukai Li, Haochen Luo, Lei Li, Qi Liu

英文摘要：Much of the variation in financial time series is driven by market-wide shocks that cannot be predicted from an asset's own history. Per-asset self-supervised learning treats this variation as irreducible noise. Under a latent factor model, we show that conditioning prediction on the realized market factor makes the asset's market exposure necessary for optimal prediction, while reducing the noise against which the asset-specific component is learned. Building on this insight, we introduce Market-Conditioned JEPA (MC-JEPA), a joint-embedding predictive architecture for financial time series. During training, we construct a leave-one-out market condition for each asset from the future target embeddings of other assets and incorporate it through factor loadings derived from the asset's own representation. On 11 years of minute-level Chinese A-share data, MC-JEPA improves over per-asset JEPA across nine targets spanning returns, market exposure, risk, and liquidity, with the strongest and most statistically significant gains in exposure, risk, and liquidity. It achieves the best RankIC on all liquidity targets and on return horizons from two hours to five days, outperforming time-series foundation models and self-supervised baselines, and delivers the best portfolio performance in a historical backtest. Ablations show that the return and exposure gains require the condition to be aligned with the realized market movement. After restoring the volatility scale removed by normalization, the predictor's factor loadings rank stocks by market beta nearly as well as a supervised probe, without labels. A synthetic factor-model study with known parameters shows that conditioning improves the recovery of both market exposure and the asset-specific component.

AI 总结：本研究提出MouldIt模型，在注塑成型的多任务时间表征学习中，于2000个循环的多模态设置下即优于经典方法，零件分类损失达1.0±0.3%，为智能工业解决方案提供高效数据驱动模型。

链接：https://arxiv.org/abs/2611.13646

机构：University of Technology Nuremberg（纽伦堡工业大学）

作者：Abdullah Ayad

英文摘要：Smart industrial solutions that can address multiple tasks require models that understand the underlying processes. We examine representation learning for modeling the injection moulding process as an example of such solutions. We consider its data efficiency and multi-modal benefit compared to classical machine learning alternatives. To this end, we synthesize a process-informed dataset based on a real reference cycle of an Erlanger Träger part. We then introduce MouldIt, a temporal convolution and attention based variational autoencoder that reconstructs in-mould signals, while classifying parts as acceptable or as one of our synthetic failure modes. We evaluate MouldIt on two tasks, known-class classification and unknown-class detection, across nine dataset sizes in single-modal and multi-modal settings. We compare it with an accept-all baseline and the classical methods k-nearest neighbors, logistic regression, random forests, and principal component analysis, which we adapt to support both tasks. MouldIt outperforms classical methods across both tasks from 4000 cycles in the single-modal setting and from 2000 cycles in the multi-modal setting. It achieves part sorting losses of 6.0 $\pm$ 0.7% and 1.0 $\pm$ 0.3% in the single-modal and multi-modal settings respectively, ranking first across all baselines and a combination of principal component analysis and random forests.

AI 总结：ScaleBrain是一种尺度一致的EEG基础模型预训练框架，通过多尺度分词与尺度特定重建结合轴向时空注意力，在6个BCI基准上以500万参数实现优异性能。

链接：https://arxiv.org/abs/2611.14035

机构：College of Computer Science, Nankai University（南开大学计算机科学学院）； Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）； Beijing Tiantan Hospital, Capital Medical University（首都医科大学附属北京天坛医院）； School of Computer Science and Cyber Engineering, Guangzhou University（广州大学计算机科学与网络工程学院）

作者：Minyu Zhou, Yajing Wu, Yongqiang Tang, Wang Jia, Wensheng Zhang

英文摘要：Foundation models pre-trained on large unlabeled electroencephalography (EEG) corpora have emerged as a transformative paradigm for Brain-Computer Interface (BCI) applications. Recent EEG foundation models have shown that multi-scale representations derived from fixed-length patches can improve EEG representation learning. Existing frameworks commonly construct temporal scales through feature-level operations. However, these operations leave the correspondence between learned features and scale-specific characteristics of the original EEG signals implicit. Moreover, when the self-supervised objective is not aligned with the multi-scale input representations, information at some temporal scales may receive insufficient supervision during pre-training. In light of this, we propose ScaleBrain, a scale-consistent pre-training framework that explicitly coordinates multi-scale representations of EEG from both the model input and optimization objective sides. Specifically, ScaleBrain introduces multi-scale EEG tokenization to represent the same EEG segment through analysis windows with different temporal spans and shared temporal anchors, and pairs these representations with scale-specific reconstruction to align with the multi-scale input structure and provide supervision across temporal resolutions during self-supervised learning. We further develop an axial spatio-temporal attention mechanism to contextualize the fused multi-scale representation along spatial and temporal dimensions. Extensive evaluations on six public BCI benchmarks against 12 representative baselines show that ScaleBrain achieves the highest mean Balanced Accuracy on five datasets and ranks third on the remaining one, with only 5.0M parameters, demonstrating strong performance and favorable parameter efficiency.

AI 总结：本研究针对涡扇发动机健康表征学习中自编码器物理意义缺失的问题，提出物理约束自编码器，在 N-CMAPSS 数据集上验证其可生成高保真健康指数，支持组件级退化分析与可视化。

链接：https://arxiv.org/abs/2611.14266

机构：Zurich University of Applied Science（苏黎世应用科技大学）； Mondragon University（蒙德拉贡大学）； Delft University of Technology（代尔夫特理工大学）； Luleå University of Technology（吕勒奥理工大学）； Fragum Global（弗拉格姆全球公司）

作者：Leonardo Baldo, Jokin Alcibar Ibarzabal, Daniel Cisneros Acevedo, Kai Goebel, Manuel Arias Chao

英文摘要：Health Indices (HIs) summarize degradation from operational data and support diagnostic and prognostic tasks in Prognostics and Health Management (PHM). Their usefulness depends on capturing degradation and interpreting system condition. Autoencoders learn such representations from unlabelled monitoring data; however, when training is driven only by reconstruction, the physical meaning of latent variables remains undefined, limiting their interpretation and diagnostic value. This work proposes a physics-constrained autoencoder that embeds system knowledge as an inductive bias in the decoder architecture and as a learning bias in the loss function. The decoder contains differentiable physical equations that receive latent variables and operating conditions as inputs, structurally guiding each latent variable to represent a predefined physical quantity and enabling component-level analysis before aggregation into a system-level this http URL methodology is benchmarked on the N-CMAPSS dataset against alternative representations. Reconstruction reaches an $R^2$ of 0.9993 and an MAE of 0.0036, comparable to fully data-driven models. After operating-condition correction, the inferred physical quantities track degradation, and their aggregation produces an HI with a Spearman correlation of 0.9791 with the ground truth, an RMSE of 0.0286, and an MAE of 0.0233. The methodology is further evaluated using a single model trained across four datasets representing distinct degradation scenarios. This model preserves component-level signatures that distinguish compression-side from expansion-side degradation without diagnostic this http URL, the proposed framework enables degradation trends to be visualized on component maps, providing a health representation in a form familiar to engine operators and domain experts.

AI 总结：该研究针对个性化联邦RL中智能体目标差异阻碍协作的问题，提出PF-UCBVI与PF-EFVI算法，证明奖励多样性可作为探索的结构性资产，实现无需显式探索的高效集体学习。

链接：https://arxiv.org/abs/2611.12964

机构：Carnegie Mellon University（卡内基梅隆大学）； Yale University（耶鲁大学）

作者：Jiin Woo, Gauri Joshi, Yuejie Chi

英文摘要：While federated reinforcement learning (RL) has traditionally guaranteed high efficiency in homogeneous environments, collaboration is often hindered when agents possess distinct goals. In this paper, we flip this paradigm, framing reward heterogeneity not as a hurdle to consensus, but as a structural asset that catalyzes exploration. We first propose Personalized Federated Upper Confidence Bound Value Iteration (PF-UCBVI), which achieves optimal linear speedup by decoupling shared dynamics from personalized objectives. To eliminate the risks of explicit exploration, we then introduce Personalized Federated Exploration-Free Value Iteration (PF-EFVI), a purely greedy algorithm that leverages reward diversity to ensure state-action coverage. We prove that PF-EFVI attains logarithmic regret without explicit exploration bonuses under sufficient reward diversity. Our results reveal that agent disagreement shifts the exploration burden from temporal complexity to spatial diversity, enabling efficient collective learning without explicit exploration.

AI 总结：该研究提出适用于冻结强化学习智能体的分支与提交方法，发现其收益取决于离散度，在Craftax-Medium中可提升12.8%回报，用预测离散度控制分支可在减少调用量的同时保留大部分增益。

链接：https://arxiv.org/abs/2611.13003

机构：Technion, Israel Institute of Technology（以色列理工学院）； Nvidia Research（英伟达研究院）

作者：Sahar Admoni, Assaf Hallak, Yftah Ziser, Ofra Amir

英文摘要：Reinforcement learning agents are often deployed with policies that cannot be retrained or modified. In digital environments whose states can be copied and resumed, an operator has another option: sample several continuations from the current state, evaluate them, and continue from the endpoint of the selected branch. We call this procedure branch-and-commit. We study when it improves return, what produces the improvement, and where its computational budget should be spent. Across four environments, two frozen policies, five continuation verifiers, and a model-free selection rule, we find that the benefit of branching depends on dispersion: the difference between the best candidate return and the mean candidate return. Branching provides no benefit in low-dispersion settings, while increasing return by 12.8% in our main Craftax-Medium setting. Most of this gain comes not from selecting better actions, but from selecting among different stochastic realizations of the same action. Several competent verifiers achieve similar episode returns, suggesting that verifier quality is not the main constraint at our operating budget. Instead, computation is better allocated to states where useful dispersion is likely. On held-out seeds, gating branching using predicted dispersion retains 69% of the improvement from uniform branching while using 48% as many branching calls. These results characterize when branch-and-commit is useful and clarify what its gains represent.

AI 总结：本研究提出FaST转换因子合成方法，将转换分解为多阶段建模，在12个D4RL基准上实现与SynthER相当性能，生成速度快90倍以上，且在多数设置中优于真实数据训练。

链接：https://arxiv.org/abs/2611.13013

机构：School of Mathematical and Data Sciences（数学与数据科学学院）； West Virginia University（西弗吉尼亚大学）； Department of Computer Science and Engineering（计算机科学与工程系）； Jadavpur University（jadavpur大学）

作者：Daniel Campa, Spandan Das, Srinjoy Das

英文摘要：Synthetic data augmentation promises to scale offline reinforcement learning beyond fixed logged datasets, but leading diffusion-based generators model the transition tuple (state, action, next state, reward, terminal) as a monolith: no single factor can be natively redrawn in isolation. We present FaST (Factored Synthesis of Transitions), which factors transitions via the chain rule into state, action, dynamics, reward, and terminal stages, fitting each with an estimator matched to its statistical properties. Across twelve D4RL locomotion and maze benchmarks, TD3+BC and IQL agents trained on FaST transitions match the SynthER diffusion baseline while generating transitions over 90x faster. The explicit action-given-state conditional can also be re-queried for a fixed state, enabling targeted action synthesis: drawing N candidate actions per generated state and keeping the highest-advantage one under a value function fit to the same data, thereby synthesizing improved behavior during generation rather than filtering for it post-hoc. Best-of-10 synthesis beats training on real data on 11 of 12 medium and medium-replay settings (+8.9 D4RL points on average, up to +28.4), and is competitive with advantage-filtered SynthER at a fraction of its generation cost. On expert mixtures both mechanisms degrade identically, tracing the failure to the advantage estimate itself rather than to either targeting mechanism.

AI 总结：针对在线策略学习中反馈协变量偏移引发的风险控制问题，提出基于排列和Sinkhorn算法的加权风险评估器与高斯自举的策略选择方法，在多任务中验证其能维持目标风险并优于基线。

链接：https://arxiv.org/abs/2611.13163

机构：Johns Hopkins University（约翰斯·霍普金斯大学）

作者：Yihong Guo, Drew Prinster, Suchi Saria, Anqi Liu

英文摘要：We study risk control for online policy learning, where an agent adaptively collects data and aims to learn a new policy whose risk stays below a tolerance threshold at every round with high probability. Such policy updates and data collection induce feedback covariate shift (FCS): observed data depend on history through evolving policies and the data-collection process. FCS brings two substantive problems for risk control: 1) efficient risk evaluation for each candidate policy and 2) valid policy selection with high-probability risk control. To address these, we propose a permutation-based weighted risk evaluator with weights efficiently approximated by the Sinkhorn algorithm and select policies whose estimated risk upper confidence bounds (UCB) fall below the target threshold, with the UCB obtained through a proposed Gaussian bootstrap method. Theoretically, we prove that our risk evaluator has a no-larger variance upper bound than that of the standard importance-weighted evaluator. Our policy control method also yields an asymptotically valid simultaneous risk bound and provides asymptotic high-probability risk control for the selected policy. Empirically, we evaluate our method on the online contextual bandit problem, safe question answering benchmark, and Medical Claim Filtering tasks. Our method maintains the target risk level, achieves lower risk than baseline methods, has lower variance in risk estimation, and achieves reliable policy selection with high-probability risk control. Further ablation studies support our theoretical claims.

AI 总结：本研究提出符号模仿学习框架，结合在线ARS调优，在20个基准RL任务中逼近并超11个任务的专家策略，生成可解释策略。

链接：https://arxiv.org/abs/2611.13182

机构：Faculty of Computer Science, Dalhousie University（达尔豪斯大学计算机科学学院）

作者：Bryce MacInnis, Malcolm I. Heywood

英文摘要：Deep reinforcement learning (DRL) methods have been the dominant paradigm in continuous control tasks due to their efficacy in function approximation, while interpretable evolutionary methods have lagged largely behind for the opposite reason. In this work, we introduce Symbolic Imitation Learning, a framework for distilling an oracle policy through an offline dataset into a decision list with linear model motion primitives. We show that this, in conjunction, with online Augmented-Random Search (ARS) tuning of the constants and motion primitives approximates the performance of the oracle across 20 benchmark RL tasks and exceeds the oracle in 11 of those 20 tasks while simultaneously producing parsimonious, interpretable policies that domain experts can better understand.

AI 总结：该研究针对流控制中基于模型的强化学习，提出通过审计代理模型动作梯度与高保真求解器估计值，结合锚定策略减少高保真求解器交互，生成的控制器性能优于或匹配无模型基线且交互次数大幅降低。

链接：https://arxiv.org/abs/2611.13203

机构：University of Washington（华盛顿大学）； NSF AI Institute in Dynamic Systems（美国国家科学基金会动态系统人工智能研究所）

作者：Ruige Kong, Stefania Fresca

英文摘要：Feedback controllers for flows are expensive to design, since each candidate must be evaluated with a high-fidelity solver. Differentiable surrogates reduce that cost, but they also introduce a failure mode that prediction error cannot detect. Policy optimization uses the derivative of the objective with respect to the action, which we call the action gradient and which can be computed either through the surrogate or through the high-fidelity solver. Optimization through the surrogate uses only its action gradient, and a surrogate can follow the flow trajectory closely while misestimating this derivative. The surrogate action gradient is compared against a central finite-difference estimate from the high-fidelity solver on held-out states, under acceptance criteria fixed before the comparison, and we call that comparison an audit. A low-dimensional feedback law is then fitted to the surrogate action gradient when the surrogate passes the audit, and to the solver estimate when it does not. This law initializes a neural policy, and a quadratic penalty holds the policy near it during optimization. The proposed approach is validated on four test cases. The resulting controllers outperform or match a model-free baseline, while using 3 to 156 times fewer high-fidelity solver interactions across the four test cases. One surrogate predicts the objective at correlation 0.9996 but its action gradient at only 0.459, so prediction accuracy does not identify a usable surrogate. Another surrogate, whose action gradient agrees with the solver estimate at correlation 0.996, makes the objective 0.52% worse when the policy is optimized through it without the anchor penalty, so an accurate gradient alone does not provide a working controller. Finally, a surrogate that passes the audit at one fixed action fails it around the actions the deployed controller applies, so a pass is local.

AI 总结：该研究在受控环境中分析RLVR的奖励黑客行为，发现跨提示参数共享会同时抑制和放大该行为，验证器修复需针对模型跨提示重复的错误，解释了早期研究分歧的原因。

链接：https://arxiv.org/abs/2611.13218

机构：Carnegie Mellon University（卡内基梅隆大学）； University of Washington（华盛顿大学）

作者：Michael Li, Rohan Pandey, Ray Chen

英文摘要： Two accounts of how reinforcement learning with verifiable rewards (RLVR) interacts with an imperfect verifier coexist: in one, reward hacking is inherited from the base model; in the other, RL amplifies it. We study both in a controlled environment small enough to compute the entire response distribution exactly. We vary the universality of a leak, or the fraction of prompts that accept the same wrong response, and the verifier's static false-positive rate. At its KL-regularized optimum, RLVR acts as a filter that reweights the accepted set without altering its composition, so the hack rate is inherited exactly. Taking that as a null model, a known decomposition writes the drift from it as a sum over prompts of covariances between reward and the empirical tangent kernel. The terms from other prompts vanish without parameter sharing. With sharing, every prompt that accepts a wrong answer pushes it up and every prompt that rejects it pushes it down through the baseline, in proportion to the model's probability of producing it there. A shared-parameter policy thus drives a leak that few prompts accept below the base rate and amplifies one that most prompts accept. At every convergent learning rate under Adam, SGD, and sampled GRPO, hacking ends higher at high universality than at low, but whether either end crosses the base rate depends on the step size. Independent per-prompt copies remove the dependence. The direction of this push depends on neither the verifier's error rate nor universality alone, but on how much of the model's kernel-weighted probability of a wrong answer falls on prompts that accept it, which frozen base model controls support. Verifier fixes should target the mistakes the model repeats across prompts: two patches that remove equally many errors differ by up to 38 points of true accuracy. We propose this alignment as a hypothesis for why earlier studies disagree.

AI 总结：该研究针对离策略强化学习的偏差-方差权衡问题，提出偏移策略方法，推导其动态规划结果并开发对应时序差分算法，在连续控制基准中验证其效率-稳定性权衡效果。

链接：https://arxiv.org/abs/2611.13275

机构：University of Alberta（阿尔伯塔大学）； Amii（阿尔伯塔机器学习研究所）； Google DeepMind（谷歌DeepMind）

作者：Jiamin He, Mark Rowland, Zhaohan Daniel Guo, Hado van Hasselt, Diana Borsa

英文摘要：Off-policy reinforcement learning (RL) tries to learn the value of a target policy while data comes from a different behavior policy. This often requires trading off bias for variance or stability. One way to trade these off is to linearly interpolate between the behavior and target policies, but this represents only one point in a broader space of interpolation mechanisms. We introduce excursion policies--a structured alternative where the agent follows the target policy for a multi-step "excursion" before reverting to the behavior policy. We derive dynamic programming results for these policies, establishing the theoretical basis for their use in RL. We further develop provably sound off-policy temporal-difference (TD) algorithms for excursion value estimation. We find that excursion policies provide an efficiency-stability trade-off comparable to mixture policies in continuous control benchmarks, and our analysis illustrates conceptual differences in their behavior.

AI 总结：该研究针对计算受限的多阶段训练流水线构建分析模型，结合预训练、SFT与RL，揭示各阶段交互及计算分配规律，为基础模型多阶段训练提供理论支撑。

链接：https://arxiv.org/abs/2611.13276

机构：Gatsby Computational Neuroscience Unit（盖茨比计算神经科学单元）； University College London（伦敦大学学院）； Sainsbury Wellcome Centre（桑斯伯里惠康中心）

作者：Nishil Patel, Jin Hwa Lee, Basile Confavreux, Andrew Saxe

英文摘要：Modern foundation models are trained through multiple stages, yet how these stages interact and how compute should be allocated across them remain poorly understood. We introduce an analytically tractable high-dimensional model combining unsupervised pretraining with a linear autoencoder, supervised fine-tuning (SFT) with noisy stepwise labels, and reinforcement learning (RL) with sparse episodic rewards. Using random matrix theory and an operator-projection closure, we derive approximate low-dimensional learning dynamics that closely match simulations. We show that finite pretraining data imposes a representation-dependent performance limit, with Bayes classification error scaling as the inverse square root of the pretraining sample ratio, and observe the same exponent in asymptotic SFT and RL errors. Better pretraining also reduces sensitivity to noisy supervision. SFT initialization shortens the initial RL plateau induced by sparse rewards, an effect we also observe in OLMo-2 1B on GSM8K. Longer SFT warm starts additionally improve asymptotic RL performance. Optimizing compute across all three stages reveals that the optimal allocation depends on supervision quality and total budget, with a non-monotonic dependence of the RL compute share on budget. Finally, a complementary compositional model shows how RL selectively strengthens task-relevant pretrained skills while largely preserving unused skills. Together, these results identify mechanisms linking representation quality, feedback structure, and compute allocation in multistage training.

AI 总结：该研究利用认知心理学的视觉注意分布发现，将注视数据融入奖励模型训练，提升了偏好预测准确率，且结果符合人类阅读注意规律。

链接：https://arxiv.org/abs/2611.13477

机构：KAIST（韩国科学技术院）； School of Digital Humanities and Computational Social Sciences, KAIST（韩国科学技术院数字人文与计算社会科学学院）； Department of Brain & Cognitive Sciences, KAIST（韩国科学技术院脑与认知科学系）； Graduate School of Data Science, KAIST（韩国科学技术院数据科学研究生院）

作者：Chaeho Lee, Junyup Kim, Wansoo Kim, Young Min Jung, Sang Ho Lee

英文摘要：Integrating human feedback into language models has gained increasing attention as a means to align model outputs with human preferences, with Reinforcement Learning from Human Feedback (RLHF) serving as a prominent framework for preference-based alignment. Recent studies have incorporated eye-tracking (ET) data as an additional supervisory signal for reward model training, grounding reward learning in human reading behavior. Building on this approach, we extend prior fixation-centric models by leveraging established findings from cognitive psychology on human letter recognition. Specifically, we model visual attention as a spatially graded distribution centered on the current gaze location and incorporate this representation into reward model training. Our results show that the reward model trained with the proposed gaze distribution achieves higher preference prediction accuracy than a baseline model. Moreover, some results from our experiments are qualitatively consistent with established findings on human attention during reading, suggesting that they potentially capture cognitively meaningful aspects of human attention allocation.

AI 总结：本研究针对KL正则化强化学习，推导了折扣MDP下插件估计器的样本复杂度上界与下界，得到其渐近常数，验证了模拟结果，明确了其样本效率特性。

链接：https://arxiv.org/abs/2611.13480

机构：Columbia University（哥伦比亚大学）

作者：Andrew Liu

英文摘要：KL-regularized reinforcement learning, the objective behind maximum-entropy RL and RL from human feedback, is variational inference in the graphical model of "control as inference": the regularized optimal policy is a posterior over trajectories, and the suboptimality of a policy is $\beta$ times a KL divergence between path measures. However, it is not known how many samples it takes to learn this posterior, even with a generative model. Earlier $1/\epsilon$ guarantees scale with $H^5$ in the effective horizon $H$, and no lower bound for MDPs depends on the temperature $\beta$. In this work, we show that for a discounted MDP with $K$ state-action pairs, the plug-in estimator finds a policy that is $\epsilon$-optimal from every initial state with $\widetilde{O}(KH^3/(\epsilon\max\{\epsilon,\beta\}) + KH)$ samples, and that no algorithm, even one that chooses where to sample, can do with fewer when $\epsilon \lesssim \min\{H, H^2/\beta\}$. For $\epsilon < \beta$ this is a $1/\epsilon$ rate with the same horizon dependence as unregularized RL. We also find the exact asymptotic constant of the plug-in estimator, a Fisher-weighted variance of the estimated soft Q-function. The estimator is locally asymptotically minimax among methods that sample every pair equally often, and we give the allocation of samples that is optimal among adaptive designs. Simulations on small MDPs match the rates and the constants.

AI 总结：该研究针对强化学习后训练中BF16收敛慢于FP16的问题，提出用随机舍入量化FP32影子权重为BF16，既匹配FP16收敛速度，又保留BF16宽动态范围且无需损失缩放。

链接：https://arxiv.org/abs/2611.13545

机构：NVIDIA（英伟达）

作者：Mahdi Nazemi, Oleg Rybakov, Tijmen Blankevoort, Asit Mishra, Carlo del Mundo

英文摘要：BF16 is widely used to pretrain large language models because it has the same dynamic range as FP32 and converges similarly to FP32 training. However, in reinforcement learning with verifiable rewards, where learning rates are typically low, we find that standard mixed-precision BF16 training converges more slowly than FP16 training at the same learning rate. We attribute this gap to weight sticking: when FP32 shadow weights are quantized to BF16 using round-to-nearest, small updates often leave the BF16 weights used for rollout generation and training unchanged. Consequently, several optimizer steps may be needed before updates to the shadow weights are reflected in the BF16 weights. Raising the BF16 learning rate reduces weight sticking and matches FP16's convergence speed, but it also increases the KL divergence from the reference model. We instead apply stochastic rounding when quantizing the FP32 shadow weights to BF16. This approach matches FP16's convergence speed across different models and datasets while reducing the KL divergence relative to BF16 training with a larger learning rate. Unlike FP16, it keeps BF16's wider dynamic range and needs no loss scaling.

AI 总结：本研究针对孤立交叉口，提出基于强化学习的方法，联合优化交通信号与车道渠化，通过带人类反馈的深度Q网络过滤不可行动作，实验表明可提升交叉口通行能力。

链接：https://arxiv.org/abs/2611.13599

作者：Jiwon Kang, Jiyoung Kim, Hyunseok Jun, Keemin Sohn

英文摘要：Optimizing traffic signal control has been a long-lasting critical issue in traffic engineering, and a number of robust methodologies have emerged to solve the problem so far. On the other hand, the road design for lane channelization is a static problem that is difficult to solve jointly with the real-time traffic signal control. If traffic volume does not vary, it is possible that both problems can be solved in an integrated manner as a single static optimization problem. Recently, researchers have devised a mixed integer linear problem (MILP) to jointly determine the traffic signal and road width designs. However, this joint optimization problem must be solved repeatedly whenever the traffic flow pattern varies. Solving the MILP may also take impractical computing time as the problem size gets larger. We present a robust methodology for solving the joint optimization problem using reinforcement learning (RL) on a real-time basis. Abrupt changes in lane markings with an incompatible signal phase can jeopardize drivers. To sidestep such a dangerous action change, we train a model with human feedback to filter infeasible action transitions in a rule-based manner. We devised a deep Q-network that can accommodate joint actions for traffic signal and lane marking plans. To ensure the implementation of the proposed operation, a variable sign for lane markings is necessary at each stop line of an intersection. The experimental results confirm that the proposed methodology enhances the intersection performance by increasing throughput.

AI 总结：针对测试时强化学习中多数投票易强化错误解的问题，提出TRUER框架，利用有限监督训练错误感知奖励校正器修正奖励，性能优于朴素TTRL。

链接：https://arxiv.org/abs/2611.13714

机构：Shanghai Jiao Tong University（上海交通大学）

作者：Mingxiang Shen, Jinhao Li, Jiaming Xu, Yibo Han, Guohao Dai, Yijia Zhang, Ningyi Xu

英文摘要：Test-time reinforcement learning enables language models to adapt using rewards derived from their own answers. However, majority voting can reinforce incorrect solutions even when correct alternatives are available. Our training audit shows that false positives (incorrect responses receiving positive rewards) dominate reward errors. We focus on a practical regime where ground-truth rewards are difficult to obtain at scale, while a limited labeled source set is available. We introduce TRUER, a framework that uses limited supervision to train a reward rectifier that distinguishes correctness from consensus. To achieve this goal, we train the rectifier with an emphasis on incorrect majorities and correct minorities, targeting cases where consensus and correctness disagree. After that, we freeze the rectifier and use it to correct rewards during subsequent test-time adaptation. By improving reward accuracy and thereby model performance, TRUER consistently outperforms naive TTRL in both metrics across all evaluated configurations.

AI 总结：该研究提出Omni-JEV模型，通过对齐的语音到语音理解骨干网络结合双语弱监督与RLCD方法，实现文本和双流语音的交互动作决策，在音频和文本任务上分别达到84.08%和96.55%的测试准确率。

链接：https://arxiv.org/abs/2611.13777

机构：Merit Interactive Research Institute（merit互动研究院）

作者：Shengzhou Kong

英文摘要：A user utterance does not uniquely determine the next action in full-duplex dialogue: its consequences depend on the assistant's response state and interaction history. Speech activity, utterance completion, and interruption intent provide related evidence, whose implications for a response require an explicit action policy. We formulate this task at collected-segment endpoints and introduce Omni-JEV, which maps text or role-indexed user and assistant-reference speech to probabilities over five interaction actions. An aligned speech-to-speech understanding backbone encodes the acoustic roles separately and shares its candidate readout with text inputs. Bilingual weak supervision supports cross-entropy fine-tuning followed by reinforcement learning for probability-sensitive decisions (RLCD). We compare Gaussian-logit and categorical Choice RLCD from the same supervised initialization under matched training exposure, observing different trade-offs between classification and probability quality. Gaussian Omni-JEV achieves test accuracies of 84.08% for audio and 96.55% for text. This common action interface supports decisions about when a full-duplex voice assistant should respond, preserve ongoing speech, interrupt, or resume an interrupted response.

AI 总结：该研究提出TBTCL框架，以时序行为树为统一接口，将任务规范、奖励、课程与诊断功能整合，在6个离散及连续域的规范上验证了其有效性。

链接：https://arxiv.org/abs/2611.13993

机构：CISPA Helmholtz Center for Information Security（CISPA亥姆霍兹信息安全中心）； Technical University of Munich（慕尼黑工业大学）

作者：Vladimir Krsmanovic, Bernd Finkbeiner

英文摘要：Reinforcement learning from formal specifications often separates task definition, reward shaping, curriculum design, and policy diagnostics into different mathematical objects. This fragmentation requires additional conversion steps and discards structure already present in the specification, such as subtask boundaries, ordering constraints, and quantitative measures. We introduce TBTCL (Temporal Behavior Tree Centric Learning), a framework that uses Temporal Behavior Trees (TBTs) as a unified interface for specification-guided reinforcement learning. A single TBT specifies the task, induces dense robustness-based rewards, defines a curriculum, and supports better understanding of learned policies. By adapting TBT syntax and semantics for specification-guided reinforcement learning, we provide a formalism that is more expressive than popular specification languages such as LTL and inherently hierarchical. We present algorithms to automatically extract a curriculum DAG and derive dense, robustness-based rewards directly from TBT semantics. Furthermore, we utilize TBT trace segmentation to build understandable competence profiles and automatically adapt the reward toward bottlenecks. We evaluate TBTCL on six specifications across discrete and continuous domains and show that Temporal Behavior Trees can serve as a specification, a reward, a curriculum, and a diagnostic.

AI 总结：本研究提出 CAMP-II 方法，结合 CAMG-II 环境，通过异步 OPD 训练策略，使编码智能体的文件记忆能力在 CAMG-II 及两个外部基准上均实现显著性能提升。

链接：https://arxiv.org/abs/2611.14009

机构：Peking University（北京大学）； Sichuan University（四川大学）； Beijing Institute of Technology（北京理工大学）； Renmin University of China（中国人民大学）； The Hong Kong University of Science and Technology（香港科技大学）

作者：Lirui Luo, Kelong Mao, Xinwei Yang, Rongqing Li, Luyu Chen, Kieran Wong, Yudong Guo, Xinrui Wang, Jiayin Zhu, Gu Simiu, Sulong Xu, Cong Fang

英文摘要： Language-model agents increasingly tackle long-horizon tasks whose interaction histories exceed the model's active context. They must therefore decide what to write down and what to read back, and when to do either. Recent work has begun to use on-policy distillation (OPD) to make memory control part of the policy, relying on memory tools and compaction protocols that lie outside either model's pre-training. However, this weakens the distillation on both sides: the teacher cannot give useful guidance on an interface it has never used, and the student samples too few meaningful memory actions for that guidance to take effect. To study this problem, we introduce Coding Agent Memory Gym II (CAMG-II), a suite of long-horizon agentic-RL environments spanning Shop, Coding, AutoResearch, General, and Visual. Alongside each environment's native task interface, CAMG-II provides ordinary Bash access and an episode-persistent workspace, enabling the policy to create, revise, search, and reuse files as memory across context replacements. We also introduce Coding Agent Memory Post-Training II (CAMP-II), which trains a single policy jointly across all five environments with fully asynchronous OPD to learn this file-based memory behavior directly from a frozen teacher with a prior for file operations. On CAMG-II, CAMP-II obtains 2.97 times the gain of the strongest OPD baseline. The memory behavior CAMP-II learns generalizes to two external benchmarks, Terminal-Bench 2.1 and SWE-Bench Pro V2, with improved success rates.

AI 总结：针对可验证奖励强化学习中GRPO平均分配奖励的问题，提出MCPO方法，利用边际分配信用，在竞赛数学基准上较GRPO取得显著性能提升。

链接：https://arxiv.org/abs/2611.14015

作者：Chi Liu, Xin Chen, Yuchuan Weng, Srinivasan Manoharan

英文摘要：In reinforcement learning with verifiable rewards, a language model is rewarded for a correct final answer, and GRPO passes that reward to every token of the solution in equal measure. We instead place the credit where the model hesitated, in a method we call MCPO (Margin-Credit Policy Optimization). We measure hesitation by the top-2 margin $M=p_1-p_2$ of the policy's next-token distribution, map it to a score with two fixed thresholds, and spread each solution's advantage over its tokens so that the mean stays unchanged. The margin is read from a forward pass the trainer already performs, so the method needs no extra model evaluation and no batch statistics. We train MiniCPM5-2B-Base with LoRA for 200 steps on 2,048 DAPO-Math problems and evaluate on six competition-math benchmarks. MCPO beats an unmodified GRPO baseline for all five data seeds, by 6.55 points of macro avg@32 on average (95\% CI $[4.1, 9.0]$), and a Qwen3-4B-Base replication improves for all three seeds ($+4.75$). The gain is largest on problems that GRPO solves only some of the time and grows with $k$ in pass@$k$, reaching 13.8 points at pass@32. MCPO models also write two to three times longer answers, and runs that write longer score higher; our experiments do not separate the two effects.

AI 总结：本文提出$OC^2$框架，以机会成本定义自适应审议代价，在Pinball目标偏移任务中前10集成功率达52%，优于固定代价基线的18%，提升了非平稳环境下的HRL适应能力。

链接：https://arxiv.org/abs/2611.14099

机构：American University of Beirut（贝鲁特美国大学）

作者：Khaled Ammoura, Aya Kayal, Naseem Daher

英文摘要：Deliberation costs are commonly used in hierarchical reinforcement learning (HRL) to prevent options from collapsing to primitive actions. Existing approaches rely on fixed penalties or duration-based adaptations, both of which are indirect proxies for an option's utility. This paper argues that option persistence should instead depend on the opportunity cost of abandoning the current behavior. We propose the Opportunity-Cost Option-Critic ($OC^2$) framework, which defines a state-dependent deliberation cost using the positive value gap between the active option and its best alternative. This encourages persistence when the current option is clearly preferable and allows rapid switching when a better option emerges. We show that directly learning such costs leads to a degenerate zero-cost solution and introduces an auxiliary objective that aligns the learned cost with the value-gap target. The approach is evaluated in Four Rooms and Pinball domains under stationary, goal-shift, and geometry-shift conditions. In the Pinball goal-shift task, the $OC^2$ framework achieves a 52\% first-ten-episode success rate compared with 18\% for a fixed-cost baseline. The results suggest that opportunity-cost-based persistence improves adaptation in non-stationary environments.

AI 总结：该研究针对HOFLON的AAE惩罚效果展开分析，将其评判器与AAE替换为sGP评判器，通过调整κ值优化动作选择，在聚合反应器基准上大幅降低启动失败率并提升回报。

链接：https://arxiv.org/abs/2611.14374

机构：Imperial College London（帝国理工学院）

作者：Marco Chan, Christos Evripidou, Liqiu Dong, Benoit Chachuat, Mehmet Mercangöz (Sargent Centre for Process Systems Engineering, Department of Chemical Engineering, Imperial College London)

英文摘要：Start-ups and grade changes remain largely operator-driven, while developing and maintaining dynamic models for model-based control requires substantial engineering effort. Hybrid Offline Learning and Online Optimisation (HOFLON) uses one-step offline reinforcement learning to select actions from historical transition data. In its original implementation, the online objective combines the predicted return from a tree-ensemble critic with an adversarial-autoencoder (AAE) penalty for unfamiliar state-action pairs. We examine the effect of this penalty and replace the critic and AAE with a sparse variational Gaussian-process (sGP) critic. Actions maximise $\mu - \kappa\sigma$, an uncertainty-adjusted estimate of the return from one action followed by the historical operating practice. On the HOFLON polymerisation-reactor benchmark, removing the AAE from HOFLON led to 4 failed start-ups in 15 runs over five critics. At its published weight, the variation in the AAE penalty is four orders of magnitude smaller than the variation in the tree-ensemble critic. Nevertheless, it changes actions when the predicted returns of competing candidates are nearly equal, and paired runs with and without the penalty diverged from the first move. With a Matérn-5/2 sGP critic, $\kappa = 0$ led to failure in 13 of 15 runs, whereas with $\kappa = 3$ none of 15 runs (five critics, three noise realisations) failed. The median and lowest returns were 13,464 and 13,398 (HOFLON: 13,393 and 12,411). A squared-exponential sGP critic with learned inducing inputs needed $\kappa \geq 3$ to remain within the empirical data support in steady state, and then settled with a temperature offset of 2.0 K.

AI 总结：本文提出AutoScaffold框架，将智能体训练经验转化为自适应脚手架，在ALFWorld等四领域的RL任务中，该框架训练的智能体性能优于基线，且训练步数更少。

链接：https://arxiv.org/abs/2611.14515

机构：University of Minnesota, Twin Cities（明尼苏达大学双子城校区）； JPMorgan AI Research（摩根大通人工智能研究院）

作者：Zijian Zhang, Rizhen Hu, Jiangweizhi Peng, Shiyang Li, Sujay Bhatt, Sihan Zeng, Mingyi Hong

英文摘要：Reinforcement learning with verifiable rewards (RLVR) produces rich training experience, yet outcome rewards reduce each attempt to a single number and leave most of this information unused. Training-time scaffolds, such as strategies or hints added to the prompt during training, offer a natural way to turn this information into guidance, but their design and adaptation are largely left to humans and tailored to specific domains. We introduce AutoScaffold, a unified AI-for-AI framework that turns a student's training experience into adaptive scaffolds, with minimal domain-specific information and no human intervention during training. A frozen teacher periodically examines this experience, diagnoses bottlenecks, and designs scaffolds in two forms: (i) strategies shared across a class of tasks and (ii) hints for individual hard tasks, adjusting the content and amount of support as the student learns. The scaffolds are removed at evaluation. We apply the same framework to ALFWorld, search-based QA, mathematical reasoning, and GPU kernel generation. When evaluated without scaffolds, students trained with AutoScaffold outperform those trained with the corresponding RL baselines in all four domains, with gains of 7.2 points in ALFWorld success and 4.0 points in average QA exact match. On GPU kernel generation, AutoScaffold increases the fraction of tasks with a correct kernel that runs faster than the PyTorch reference by up to 5.0 points. On math, our student surpasses QuestA's reported accuracy using approximately one-third as many student training steps. Analyses show that the teacher identifies a different failure in each task class and shifts its scaffolds from step-by-step workflows to targeted fixes as the student improves, and that its revised scaffolds help the student learn more than keeping the previous scaffold or using no scaffold.

AI 总结：该研究针对流匹配机器人策略的欧拉采样，发现轨迹精度与局部稳定性可能要求不同步数，通过实验表明仅端点精度无法排除单独的稳定性要求。

链接：https://arxiv.org/abs/2611.13311

机构：Los Alamos National Laboratory（洛斯阿拉莫斯国家实验室）

作者：Mahindra Rautela

英文摘要：Flow matching robot policies commonly use a fixed Euler step count selected from offline accuracy sweeps. At inference time, however, each policy input (visual observation, robot state, and language instruction) together with a source-noise draw defines a different ODE trajectory and sequence of local Jacobians. We ask whether a step count that accurately follows the nominal noise-to-action trajectory can still be too coarse for strongly contracting directions around it. An exact two-coordinate system and a controlled augmentation of six trained flow models show that trajectory accuracy and local Euler stability can impose different step requirements even when the original learned transport is unchanged. We then study a frozen $\pi_{0.5}$-LIBERO checkpoint without modifying its weights or action dynamics. Across 30 policy inputs from 10 LIBERO-10 tasks, the Jacobian exhibits strong contracting directions near the action endpoint, yet no sampled mode violates the Euler non-expansion condition. On five of those inputs, accuracy sweeps require 4--9 Euler steps at 5% terminal error and 87--112 at 1%. These five paired cases are therefore consistent with accuracy-limited rather than stability-limited sampling under this local criterion, while the controlled experiments show why endpoint accuracy alone cannot rule out a separate stability requirement.

AI 总结：该研究针对流匹配在目标损坏与标签噪声下鲁棒性不足的问题，提出时间残差分数（TRS）和对比条件分数（CCS）映射样本权重，实验验证了方法的鲁棒性提升。

链接：https://arxiv.org/abs/2611.13324

机构：KTH Royal Institute of Technology（瑞典皇家理工学院）； Istanbul Technical University（伊斯坦布尔理工大学）； University of Cambridge（剑桥大学）； Science for Life Laboratory（生命科学实验室）

作者：Mert Can Kurucu, Erik Englesson, Hossein Azizpour

英文摘要：Flow Matching (FM) is a strong framework for generative modeling due to its stable training and efficient sampling. However, FM assumes clean target data and labels, an assumption often violated in practice. We formulate FM training as a time-dependent regression problem over conditional vector fields and analyze how the two corruption types affect the objective: i) target corruption adds a time-dependent residual term that grows with time, ii) label noise causes updates to the wrong class-conditional vector field. Based on this analysis, we propose to improve FM robustness by mapping corruption-specific residual scores to sample weights. For target corruption, we propose the Temporal Residual Score (TRS), which emphasizes residual magnitudes at intermediate-to-late FM times, where clean and corrupted samples are more separable. For label noise, we propose the Comparative Conditional Score (CCS), which compares the residual of the observed class-conditional vector field against alternative class-conditional vector fields for the same trajectory and velocity target. Experiments across controlled corruption settings and real-world noisy labels show that the proposed methods improve robustness over standard FM and robust-training baselines, supporting corruption-specific residual structure as a practical basis for robust FM.

AI 总结：AdjointFlow是基于随机插值的框架，将物理规律注入生成式传输，在反问题重建中比条件扩散模型更准确，推理速度比PnP方法快达五个数量级，可选优化可进一步提升精度。

链接：https://arxiv.org/abs/2611.13389

机构：University of Wisconsin–Madison（威斯康星大学麦迪逊分校）； Carnegie Mellon University（卡内基梅隆大学）； Google Research（谷歌研究）

作者：Ethan Hanold, Nicholas M. Boffi, Qin Li, Leonardo Zepeda-Núñez

英文摘要：Inverse problems are ubiquitous across scientific applications ranging from seismic imaging to astrophysics to nondestructive testing, yet classical solvers often struggle with nonconvexity and repeated evaluations of expensive physical models. Diffusion models can amortize this computation offline, but largely data-driven conditional models underutilize known physics and suffer from ill-conditioned transport maps when the target distribution is concentrated. Plug-and-play (PnP) methods explicitly leverage the physical model to guide a learned prior, improving data consistency but reintroducing costly iterative computations. We instead propose AdjointFlow, a framework built on stochastic interpolants that injects the known physics into the generative transport itself. We design a data-dependent, adjoint-based initialization to reduce transport distance to the target, modify the transport velocity to remain regular as the endpoint distributions concentrate, and parameterize the network to build the initialization directly into the learned velocity. Together, these enable accurate reconstruction at substantially lower cost than generative models starting from Gaussian noise. Experiments on inverse scattering, seismic imaging, and black hole imaging show that AdjointFlow consistently produces more accurate reconstructions than conditional diffusion models while being up to five orders of magnitude faster at inference than PnP methods. Optional inference-time refinement that queries the physical model can further improve data consistency, matching or exceeding the accuracy of PnP methods without retraining.

AI 总结：提出基于ADMM的即插即用框架结合流匹配去噪器，通过辅助变量、对偶更新及轨迹校正机制解决数据与先验一致性问题，实现低成本高质量逆问题重建。

链接：https://arxiv.org/abs/2611.13742

机构：TH Würzburg-Schweinfurt（维尔茨堡-施韦因富特应用科学大学）； University of Würzburg（维尔茨堡大学）

作者：Dibyanshu Kumar, Magda Gregorova

英文摘要：Plug-and-Play (PnP) methods with generative priors have achieved strong performance in inverse problems, with recent flow matching models enabling fast and scalable reconstruction. However, a key challenge in PnP methods is maintaining consistency between data fidelity updates and the learned generative prior, as data-consistency steps can move iterates away from the flow trajectory. We propose an ADMM-based PnP framework for flow matching priors that enforces stronger consistency between the measurement model and the prior through auxiliary variables and dual updates. In addition, we introduce a trajectory correction mechanism that restores consistency with the flow via a combination of deterministic and stochastic projections. Our method provides a more principled integration of flow priors within PnP frameworks, achieving high-quality reconstructions with reduced computational cost.

AI 总结：本研究将权重空间编程应用于EEG、ECG生物信号模型，发现前沿语言模型可调试故障、用更少受试者训练网络，其构建的网络性能优于经典手工流程，但存在无法改进成熟模型等局限。

链接：https://arxiv.org/abs/2611.13758

机构：École Polytechnique Fédérale de Lausanne (EPFL)（洛桑联邦理工学院）

作者：Dara Varam, Taraneh Aminosharieh Najafi, David Atienza

英文摘要：Weight-space programming allows language models to read a trained neural network's weights and edit them. This opens the door to ``training'' by reasoning rather than through gradient descent. Prior work has demonstrated its functionality and promise for game-playing networks and fixing specific failures in language models, even extending to diffusion models. In this work, we present its applicability to models trained on biosignal data, specifically EEG and ECG data. We show that frontier models are not only able to debug specific faults planted in biosignal-based models without being told what these faults are, but are even capable of writing all the weights of a network from scratch, given only the architecture, the task, and a minimal set of training subjects. In particular, in one set of experiments, a language model was able to match standard gradient-descent training using $6\times$ fewer training subjects, through reasoning alone and without any gradient descent step. Networks written by Claude Opus 5.5 also scored higher than classic hand-crafted EEG pipelines (alpha band power, and common spatial patterns with linear discriminant analysis) when built from the same 5 subjects. We also highlight some of the limitations, which include the inability of language models to improve already well-trained classification models, and their incapacity to reliably recognize, from the weights alone, the task a network was trained on. Overall, our results indicate that weight-space programming can be useful for small networks, such as those deployed on edge devices and wearables, but the concept still requires further investigation before it can be used in practice.

AI 总结：该研究提出RB-LDC方法，通过在编码器与扩散过程间引入线性层平衡隐变量编码冗余，在ImageNet实验中显著降低含噪隐变量重建的FID，提升了噪声鲁棒性。

链接：https://arxiv.org/abs/2611.13955

机构：DGIST（大邱科技研究院）

作者：Hyunseok Jeong, Jaeho Jeon, Young-Sik Kim

英文摘要：Latent diffusion models decode approximately correct latents, making decoder sensitivity to perturbations relevant to image quality. After whitening, forward diffusion corrupts every direction at the same signal-to-noise ratio (SNR), despite unequal decoder sensitivities. We study an invertible linear layer between the encoder and diffusion process that shifts each direction's log-SNR. Its linear minimum mean-square-error (LMMSE) recovery risk has an exact spectral expression for any whitened latent distribution. Under an energy budget and a floor ensuring invertibility, minimizing decoder-sensitivity-weighted LMMSE risk integrated over the diffusion schedule yields a unique optimal allocation. The layer aligns with the decoder's expected pullback metric and equalizes marginal risk reduction across directions, computed by a deterministic spectral solver. At large budgets, the allocation follows a power law in sensitivity whose exponent is determined by the low-SNR tail of the loss weighting: one half for noise prediction and two thirds for data prediction under the cosine schedule, with a crossover on truncated schedules. We instantiate this principle as RB-LDC using Jacobian-based sensitivity estimates. On ImageNet at 256 x 256 resolution with SD-VAE and VA-VAE, trained RB-LDC improves reconstruction from noisy latents over budget-matched isotropic coding in all four metrics and five coding configurations, averaged over 15 SNRs and three pipeline seeds. It reduces average perturbed FID by 17.1 for VA-VAE block-local coding while preserving clean reconstruction. We also report intermediate SNRs where perturbed FID worsens. With a frozen diffusion model, RB-LDC yields lower FID point estimates in all 12 tested combinations of sampling progress and channel SNR under corruption during sampling.

AI 总结：本文将结构因果多臂老虎机扩展至多类场景，开发相关准则与算法，通过实验验证了因果知识与实验能力对动作选择及学习的指导作用。

链接：https://arxiv.org/abs/2611.14066

机构：Seoul National University（首尔大学）

作者：Min Woo Park, Sanghack Lee

英文摘要：Structural causal bandits use causal knowledge to guide exploration when reward mechanisms are unknown. This paper extends the approach to partially known causal structure, information from other environments, and realizable counterfactual actions. First, we develop graphical criteria for minimal intervention sets under Markov equivalence and characterize possibly-optimal minimal intervention sets for maximal ancestral graphs. A search algorithm finds sets that are possibly optimal in at least one graph represented by a partial ancestral graph. Second, we derive target reward bounds from source distributions through partial transportability, accounting for non-manipulable variables. Dominance relations provide further bounds for action pruning and restricted exploration, with a regret guarantee under valid prior bounds. Third, we formulate counterfactual structural causal bandits over actions supported by specified experimental capabilities. Graphical characterizations of minimality and possible optimality yield algorithms that remove equivalent or dominated choices while preserving an optimal action. Experiments with specified structural causal models examine how these reductions and transferred bounds affect learning. Together, these results clarify how causal knowledge and experimental capabilities determine which actions merit exploration and guide learning within the retained action space.

AI 总结：本研究采用表格基础模型TabPFN，在无需超参数调优的情况下，精准预测含能材料的撞击感度、分解温度和晶体密度，其性能优于多种机器学习模型，还具备内置不确定性量化能力。

链接：https://arxiv.org/abs/2611.14462

机构：Norwegian Defence Research Establishment (FFI)（挪威国防研究机构（FFI））； Norwegian Computing Center (NR)（挪威计算中心（NR））

作者：Vagn Krister Nes Glomsrud, Mathias Bynke, Lars Henry Berge Olsen, Dennis Christensen

英文摘要：We use TabPFN, a recently developed tabular foundation model, to predict three properties of energetic materials: impact sensitivity, decomposition temperature and crystal density. Without any hyperparameter tuning, TabPFN outperforms the best published models for all three properties. It also substantially outperforms a wide range of established machine-learning models, including tuned gradient boosting (XGBoost), random forests and Gaussian processes. This holds both when the models are given hand-crafted molecular descriptors and when they are given learned representations from Chemprop's message-passing neural network. In particular, for all three properties, TabPFN applied to Chemprop's own embedding is more accurate than Chemprop's predictions. TabPFN is also among the fastest models we consider, more than an order of magnitude faster than XGBoost. As TabPFN predicts the entire distribution of the target rather than just a single value, it offers built-in uncertainty quantification, which we compare with the empirical distributions of compounds with repeated measurements. Finally, we interpret the predictions using Shapley values, which rank the features by their importance.

AI 总结：该研究针对多回合在线策略蒸馏的参考选择问题，提出以最长公共子序列（LCS）的动作序列对齐作为轨迹兼容性代理，实验显示其在ALFWorld及航空、零售领域均能提升性能。

链接：https://arxiv.org/abs/2611.13135

机构：Amazon Web Services（亚马逊网络服务）

作者：Mufei Li, Nikki Lijing Kuang, Zhuoer Wang, Samson Koelle, Jie Yuan, Manqing Mao, Hong Wang, Yanjun Lin, James Feng, Daniel Edmiston, Pan Li, Wei Niu

英文摘要： On-policy distillation (OPD) reduces train-inference distribution mismatch by supervising a student on its own trajectories. For multi-turn tool-use agents, conditioning the teacher on a validated same-task successful trajectory as a reference can provide useful hindsight, but different successes may reach incompatible interaction states and offer misleading supervision. We cast successful-reference selection as a free-target optimal transport problem, which reduces to assigning each failed trajectory its most compatible successful reference. Because latent state paths are hidden, we use action-sequence alignment as an observable proxy for compatibility and instantiate it with longest common subsequence (LCS), which identifies a shared ordered backbone of actions while allowing recoverable detours. This perspective makes trajectory compatibility the central criterion for reference selection in multi-turn OPD. Experiments on ALFWorld and the airline domain of $\tau^3$-bench show that LCS is the only evaluated reference selector to improve over reference-free OPD on both ALFWorld overall scores and all three airline metrics. Without further optimization, the airline-trained models also transfer to the retail domain, where LCS improves all three metrics.

AI 总结：该研究针对高维线性回归的尾平均SGD，建立了与初始化显式匹配的尖锐风险边界，填补了现有分析中有效方差的差距，还验证了谱反馈条件的适用场景并补充了分离构造。

链接：https://arxiv.org/abs/2611.13383

机构：University of California, Los Angeles（加州大学洛杉矶分校）

作者：Qingyue Zhao, Xuheng Li, Quanquan Gu

英文摘要：We study the excess risk of constant-stepsize online stochastic gradient descent (SGD) with tail-averaging for high-dimensional linear regression. Building on the instance-dependent spectral framework of Zou et al. (2021), we establish matching initialization-explicit upper and lower bounds on the total excess risk of tail-averaged SGD in terms of the population covariance spectrum $\mathbf{H}$ and initialization $\mathbf{w}_0-\mathbf{w}^*$ whenever a trajectory spectral feedback condition holds. Existing analyses exhibit a fundamental gap in the effective variance: the upper bound of Zou et al. (2021) retains an undamped contribution from the initial error in the head eigenspace $\mathbf{I}_{0:k^\dagger}$, whereas their lower bound omits the variance induced by the head initialization altogether. We show that both the existing upper and lower bounds are loose: the single-hop fast-mode variance contribution from the head initialization is exponentially damped during the burn-in phase, whereas fourth-moment energy scattered from the head initialization into slow spectral modes during the early burn-in steps survives undamped into the averaging window. We prove that our spectral feedback condition is satisfied in two concrete constant-stepsize regimes: (i) for arbitrary covariance spectra $\mathbf{H}$ and any finite horizon, under an explicit horizon-dependent stepsize chosen prior to execution; and (ii) for polynomially decaying spectra $\lambda_i\asymp i^{-a}$ ($a>1$), under a single horizon-independent positive stepsize. Finally, we complement these positive results with explicit one- and two-dimensional separation constructions proving an exponential-versus-polynomial gap in the lower bound of Zou et al. (2021) and showing why fourth-moment assumptions alone cannot determine the asymptotic exponential decay rate without additional structural conditions.

AI 总结：本研究探讨复合建模在多目标贝叶斯优化（MOBO）中的益处，通过理论分析和多组实验，发现当中间响应到最终目标的映射为非双射且非仿射时，复合建模可提升优化效率。

链接：https://arxiv.org/abs/2611.13565

机构：Cornell University（康奈尔大学）

作者：Ricky Ye, Sahas Goli, Muhammad Ali, Poompol Buathong, Peter I. Frazier

英文摘要：Scientific design often requires balancing competing objectives that are expensive to evaluate through experiments or simulations. Multi-objective Bayesian optimization (MOBO) addresses this challenge by using probabilistic models to guide the selection of promising designs. In many applications, tunable inputs map to observable intermediate responses, which are then combined through known functions to compute the final objectives. Prior work on Bayesian optimization of composite functions shows that learning from these intermediate responses, which we call Composite modeling, can substantially improve sampling efficiency in single-objective optimization, compared to methods that condition only on black-box observations of the objective function. The benefits of Composite modeling in MOBO, however, remain understudied. In this work, we study the benefits of Composite modeling in MOBO and investigate the factors driving variation in its gains. We first establish single-objective conditions under which intermediate observations provide no additional predictive information beyond objective values, where composite and black box methods use the same warped Gaussian process prior. Controlled numerical experiments with versions of four MOBO methods across eight synthetic and application-inspired scientific-design benchmarks examine how these conditions relate to optimization gains in the multi-objective setting, where the black box approach models the objective functions with a standard multi-output Gaussian process. Both investigations support the claim that composite modeling tends to help more when the function mapping intermediate responses to final objectives is non-bijective and non-affine.

AI 总结：该研究探讨有界欧氏域上浅层ReLU$^k$网络的参数几何对近似率、谱性质、稳定性等的影响，给出相关上下界估计及离散恢复等结果，明确参数几何对近似与稳定性的支配作用。

链接：https://arxiv.org/abs/2611.13952

作者：Yuqing Liu, Tong Mao, Jinchao Xu

英文摘要： On bounded Euclidean domains, we study how hidden-parameter geometry affects shallow ReLU$^k$ networks through results on approximation rates, saturation-type lower bounds, Gram and stiffness matrix spectra, Bernstein and Marcinkiewicz--Zygmund inequalities, discrete least-squares recovery, and inverse approximation. Our analysis uses smooth localization and dyadic Fourier estimates. We quantify parameter geometry by the fill distance $h$ and antipodal separation distance $\underline h$. For the prescribed network space $L_n^k$, representative approximation estimates take the form \begin{equation*} \begin{aligned} \inf_{g_n\in L_n^k}\|f-g_n\|_{\mathcal L^2(\Omega')}&\lesssim h^r\|f\|_{\mathcal H^r(\Omega)},\\ \inf_{g_n\in L_n^k}\|f-g_n\|_{\mathcal L^2(\Omega)}&\gtrsim \underline h^{\frac{d+2k+1}{2}}\|f\|_{\mathcal L^2(\Omega')}. \end{aligned} \end{equation*} The upper bound exploits target regularity up to the critical order $\frac{d+2k+1}{2}$, while the lower bound applies asymptotically to fixed targets smoother than this order. The upper estimate holds on an interior subdomain $\Omega'$, whereas the lower estimate concerns the larger domain $\Omega$. We also obtain geometry-dependent spectral estimates that quantify conditioning and coefficient stability. Further results establish norm discretization and regularized recovery from function values, and infer target regularity from approximation rates under coefficient or separation conditions. Altogether, these results describe how parameter geometry governs approximation accuracy, the stability of coefficient representations, discrete recovery, and the regularity information contained in convergence rates.

AI 总结：该研究针对分布式学习中Muon优化器的鲁棒性问题，在一阶信赖域框架下建立拜占庭鲁棒Muon的收敛保证，经多基准测试其受攻击准确率更高、困惑度更低，性能优于SGD等方法。

链接：https://arxiv.org/abs/2611.13137

机构：King Abdullah University of Science and Technology (KAUST)（阿卜杜拉国王科技大学）

作者：Grigory Malinovsky, Yury Demidovich, Peter Richtárik

英文摘要：Optimization methods, including the Muon optimizer based on the Linear Minimization Oracle (LMO) and momentum orthogonalization, have shown strong empirical performance in large-scale and distributed learning, yet their robustness to adversarial behavior and data heterogeneity is not well understood. We study these methods within a unified first-order trust-region framework and analyze Muon-based updates with robust aggregation for Byzantine-robust distributed and federated learning. Our analysis does not rely on the i.i.d. assumption and introduces a non-Euclidean notion of heterogeneity consistent with the trust-region geometry. Under this framework, we establish convergence guarantees for both general non-convex and star-convex objectives. Our results provide a theoretical characterization of Byzantine-robust Muon-based optimization, with an additive heterogeneity error floor consistent with known impossibility results, under explicit aggregation assumptions. Under an explicit additive oracle-error condition, the guarantees also cover inexact orthogonalization, with the non-convex Byzantine terms in the bound's numerator unchanged but a smaller progress denominator. In simulated federated training with ResNet-18 and ViT-Tiny on CIFAR-10, additional image benchmarks, and a 38M-parameter byte-level language model on TinyStories, the tested Muon configuration attains higher attacked accuracy (lower perplexity) than the tested SGD and normalized-SGD configurations, and than sign updates on CIFAR-10. Fixed-honest controls further show a substantial RFA advantage over averaging under a bounded RFA-query attack, higher attacked accuracy for Muon-NS than tensorwise Euclidean normalization from shared checkpoints, and lower held-out TinyStories perplexity than a tuned AdamW baseline.

AI 总结：针对传统联邦学习依赖共享模型骨干限制异构LLM协作的问题，提出FEDVP范式，通过汇集本地训练的验证器选择答案，在5个医学问答数据集上使平均准确率提升1.46-5.79个百分点。

链接：https://arxiv.org/abs/2611.13708

机构：University of Bath（巴斯大学）； Artificial Intelligence Innovation and Incubation Institute, Fudan University（复旦大学人工智能创新与孵化学院）

作者：Yifan Li, Xi Chen

英文摘要：Conventional federated learning relies on the shared model backbone to aggregate client updates, limiting collaboration among institutions that deploy large language models (LLMs) from different families and at different scales. We propose Federated Verifier Pool (FEDVP), a new federated learning paradigm that enables heterogeneous clients to collaborate through shared verification rather than parameter aggregation. Each client trains a verifier using local data and responses generated by its chosen LLM backbone. The server pools these independently trained verifiers and combines their judgments to select answers without merging model parameters. This allows both LLM backbones and verifiers to differ in architecture and size, while keeping the LLM backbones frozen and deriving collaborative gains from improved answer selection. We evaluate FEDVP across five public medical question-answering datasets treated as federated clients, using Qwen and LLaMA backbones ranging from 3B to 14B parameters. FEDVP consistently improves performance across all five clients and outperforms the strongest conventional federated baselines by 1.46-5.79 percentage points in mean accuracy.

AI 总结：该研究提出联邦学习（FL）框架，利用广东、澳门、葡萄牙的异步疫情数据训练LSTM网络，实现跨区域隐私保护的COVID-19短期预测，精度接近集中式模型且开销可接受。

链接：https://arxiv.org/abs/2611.14272

机构：Macau University of Science and Technology（澳门科技大学）； INESC-ID； Instituto Superior Técnico（里斯本高等技术学院）； University of Lisbon（里斯本大学）； National Clinical Research Center for Respiratory Disease（国家呼吸系统疾病临床研究中心）； State Key Laboratory of Respiratory Disease（呼吸系统疾病国家重点实验室）； Guangzhou Institute of Respiratory Health（广州呼吸健康研究院）； The First Affiliated Hospital of Guangzhou Medical University（广州医科大学附属第一医院）； Guangzhou Medical University（广州医科大学）

作者： Ruibin Liu, Francisco Fernandes, Xuanfeng Li, Zige Liu, André Salgado, Chitin Hon, Zifeng Yang, Arlindo Oliveira, Ana Teresa Freitas

英文摘要：Objective: Epidemic outbreaks are often asynchronous across regions, yet collaborative forecasting is hampered by privacy laws and heterogeneous public health interventions. We developed and evaluated, as a proof of concept, a federated learning (FL) framework for cross-regional COVID-19 forecasting under asynchronous outbreaks among Guangdong, Macau, and Portugal, while maintaining data sovereignty. Methods: Using daily case data from WHO and China CDC for the three regions, whose outbreaks were asynchronous and interventions heterogeneous, we trained Long Short-Term Memory networks via FL for short-term forecasting, benchmarked against localized and centralized training using Root Mean Square Error, Mean Absolute Error, and Symmetric Mean Absolute Percentage Error. All experiments used the Flower FL framework with consistent hyperparameters across sites. Publicly available, aggregated case-count data were used so that the centralized benchmark could be computed as a reference. Results: The dataset comprised 1,233 daily observations (411 days) of new confirmed COVID-19 cases from November 24, 2021, to January 8, 2023. The earlier Omicron wave in Portugal supplied transmission dynamics that improved forecasts in the later-onset regions of Macau and Guangdong. The FL model achieved a normalized Root Mean Square Error of 0.028, an 88.3% improvement over the local model (0.240) and comparable to the centralized model, with network overhead mainly determined by model size and aggregation rounds. Conclusion: FL achieved short-term forecasting accuracy close to centralized training while maintaining data residency, with acceptable overhead. Leveraging asynchronous outbreaks across regions may offer a practical path to privacy-preserving collaborative forecasting, though prospective validation is required.

AI 总结：针对并行分割学习（PSL）因客户端聚合缺失导致的梯度发散问题，提出带参考梯度半空间对齐的PSL-RGHA框架，通过RGPC与GHA组件缓解梯度冲突，性能优于现有基线。

链接：https://arxiv.org/abs/2611.14312

机构：Xiamen University of Technology（厦门理工学院）； SnT, University of Luxembourg（卢森堡大学SnT学院）； Universiti Malaya（马来亚大学）； City University of Hong Kong（香港城市大学）； Peking University（北京大学）

作者：Han Li, Zheng Lin, Miao Yang, Zehang Lin, Jing Yang, Zihan Fang, Qiyang Zhang, Zhaolu Kang, Jianmin Li, Lin Li, Mingwei Hong, Shunzhi Zhu

英文摘要：The escalating complexity of deep neural networks hinders the deployment of federated learning (FL) on resource-constrained devices, especially for multimedia applications such as on-device image classification and speech recognition. Parallel split learning (PSL) has emerged as a promising solution by offloading a substantial portion of the computational workload to the server through layer-wise model partitioning, making it particularly suitable for resource-limited visual and speech intelligence scenarios. However, the absence of client-side aggregation renders PSL sus ceptible to severe training divergence and potential convergence failure under heterogeneous data distributions. To address this challenge, we propose PSL-RGHA, a gradient-aligned PSL framework that employs reference gradient half-space alignment mechanism to mitigate training divergence among client devices. PSL-RGHA comprises two novel components: reference gradient progressive construction (RGPC) and gradient half-space alignment (GHA). Specifically, RGPC evaluates the directional consistency of client gradients and progressively synthesizes a robust reference gradient via Weiszfeld updates, effectively approximating the global descent direction. Then, GHA constrains client gradients within the reference half-space defined by RGPC to mitigate gradient conflicts and rectify local updates. Extensive experiments demonstrate that our PSL-RGHA framework outperforms existing baselines in various settings.

AI 总结：提出ConceptTime模型，以预测性惊喜为自 grounding 概念瓶颈，在17个基准上兼具可解释性、准确性与弃权能力，表现优于可解释基线且适配OOD检测等任务。

链接：https://arxiv.org/abs/2611.12955

机构：Nanyang Technological University（南洋理工大学）； Institute of Advanced Intelligence and Computing (IAIC), A*STAR（新加坡科研局高级智能与计算研究所）； Khalifa University（哈利法大学）； Singapore University of Technology and Design（新加坡科技设计大学）

作者：Sachith Abeywickrama, Emadeldeen Eldele, Min Wu, Xiaoli Li, Chau Yuen

英文摘要：Time-series classifiers deployed in risk-sensitive domains must be (1) accurate, (2) interpretable, and (3) able to abstain when evidence is weak. Existing methods do not deliver all 3 at once, and the obstacle is structural, rather than logistical. Meaningful temporal patterns cannot be defined without knowing segment boundaries, yet meaningful boundaries cannot be placed without knowing which patterns to expect. Existing approaches break this circularity by using fixed windows or expert-annotated vocabularies, thereby sacrificing at least one of the 3 properties. We introduce ConceptTime, which closes the loop through a single self-supervised signal. A frozen probabilistic forecaster predicts a distribution over the next observation at every timestep of the sequence. We call the mismatch between its prediction and the realized signal the predictive surprise. This signal both locates segment boundaries and characterizes each resulting segment through the forecaster. Segments are thus grounded by how they behaved relative to expectation, not by what they look like. These summaries cluster into a vocabulary of dynamical regimes that we call concepts. A lightweight head classifies from the concept sequence, and the distance to the nearest concept yields reliability by design. Across 17 public benchmarks, ConceptTime is competitive with state-of-the-art black-box accuracy, beats every interpretable baseline, and retains near-98% accuracy on Epilepsy with 1% of labels. It further remains competitive with softmax, prototype-distance, and SHAP/TimeX/TimeX++ on OOD detection and deletion faithfulness, without a single annotation, language model call, or domain expert. Code is available at: this https URL.

AI 总结： 该研究针对递归自我改进中模型再次更新后自我更新是否仍有用的问题，提出Third Eye方法分离即时性能与后续更新结果，通过多模型实验发现即时效用接近零且连续更新会降低效用，建议将持续自我更新作为适应过渡评估。

链接：https://arxiv.org/abs/2611.12988

机构：Mahindra University（马恒达大学）

作者：Sujato Dutta, Padala Indira Sneha, Kypu Mahitha, Shashank Vanga, Neeraj Choudhary

英文摘要：Does a useful self-update remain useful after the model updates again? Third Eye separates immediate performance from the outcome of a further native update. We evaluate five 3-8B backbones on math and code: 64 fresh runs, three candidates per run, and three continuations per candidate. Selection and final scoring use disjoint items. In the balanced Qwen3-4B/Llama-3.2-3B comparison, immediate-selected utility is near zero, but continuation reduces it by 1.78 percentage points (simultaneous 95% interval [0.68, 2.88]). Candidate-averaged immediate utility is already negative: selection partly accounts for the apparently neutral first step. Continuation also lowers utility for SmolLM3-3B and Llama-3.1-8B; Qwen3-8B remains inconclusive. Target and out-of-distribution losses dominate retention changes. Matched recipe and recipient-donor controls identify neither a reliable mitigation nor a single mechanism. The results support evaluating persistent self-updates as adaptation transitions, with independent scoring and a no-update baseline, rather than interpreting immediate performance as their full value.

AI 总结：提出LEX框架，通过高斯混合模型聚类训练输入空间嵌入并关联误差裕度，在34个基准应用上使OOD下的置信度包含率达94%，裕度比全局基线紧17%，提升了RM模型的OOD鲁棒性。

链接：https://arxiv.org/abs/2611.13199

机构：Karlsruhe Institute of Technology（卡尔斯鲁厄理工学院）； Huawei Heisenberg Research Center（华为海森堡研究中心）

作者：Mahmoud Sanad, Mohammed Bakr Sikal, Osama Abboud, Heba Khdr, Jörg Henkel

英文摘要：Machine learning-based resource management (RM) models degrade under out-of-distribution (OOD) inputs at runtime, yet existing approaches in the literature rely on a single global error margin that is overly conservative and fails to distinguish reliablepredictions from catastrophic failures. We propose Lightweight Runtime EXplainability (LEX), the first post-hoc XAI framework for input-specific runtime confidence estimation in ML-based RM. LEX clusters the embeddings of the training input space using a Gaussian Mixture Model and associates per-cluster prediction error bounds to input samples, enabling fine-grained runtime confidence signals without retraining. Evaluated on a state-of-the-art LSTM-based IPS forecasting model on Intel Alder Lake, across 34 benchmark applications, LEX achieves ~94% containment on unseen workloads and produces ~17% tighter per-input confidence bounds than global baselines. With its negligible runtime overhead, LEX advances the field by providing explainable and reliable input-specific confidence signals of ML predictions, thereby making ML-based resource management resilient to OOD at runtime.

AI 总结：本研究通过系统评估发现，Transformer组合泛化性能受组合等价性与模块覆盖率影响，直接模型的表观性能源于利用组合等价性，基准设计需避免此类等价性以准确评估泛化能力。

链接：https://arxiv.org/abs/2611.13498

机构：University of Massachusetts Amherst（马萨诸塞大学阿默斯特分校）； Google DeepMind（谷歌DeepMind）

作者：Purva Pruthi, Andrew Yuan, Alexander D'Amour, David Jensen

英文摘要：Compositional generalization is an important form of out-of-distribution generalization, defined as the ability to train on some combinations of modules (e.g., functions such as sort, filter) and then generalize to unseen combinations. Despite substantial empirical work evaluating compositional generalization in transformers, findings have been inconsistent because evaluations were often conducted on a limited set of train-test splits, defined in an ad hoc manner. Through a systematic evaluation, we identify two key aspects of the train-test distribution that strongly affect performance: composition equivalence and module coverage. We show that the apparent performance of direct models (trained only on final outputs) results primarily from exploiting composition equivalences, meaning different sequences of modules that reduce to identical end-to-end functions. When benchmarks eliminate these equivalences, performance drops significantly, showing their inability to generalize to compositions of known modules that produce novel end-to-end functions. We further demonstrate that performance degrades monotonically as module coverage shifts increase (when modules appear in different contexts between train and test), affecting both direct and step-by-step models (trained on intermediate and final outputs). These findings have concrete implications for benchmark design: benchmarks that include compositional equivalences can lead to significant overestimation of the generalization abilities of direct models, and benchmarks that include both module coverage violations and compositional equivalences can exacerbate shortcut learning. As a case study, we demonstrate the role of compositional equivalence on the generalization performance of the pre-trained model Llama3-8B on the code-output prediction task in the coding benchmark LiveCodeBench.

AI 总结：针对数据稀缺下非侵入式模型降阶的高方差问题，提出MF-OpInf方法，结合高低保真数据，经实验验证其在相同预算下预测PDE解场精度优于仅高保真的OpInf。

链接：https://arxiv.org/abs/2611.13624

机构：Georgia Institute of Technology（佐治亚理工学院）

作者：Vivian Zhang, Tomoki Koike, Atticus Rex, Elizabeth Qian

英文摘要：Many-query analyses, in which a computational model is repeatedly evaluated, play an important role in uncertainty quantification and optimization for engineering design. However, many-query analyses can be prohibitively expensive when considering high-fidelity numerical models, especially those governed by partial differential equations (PDEs). Surrogate modeling replaces costly simulations with cheap approximations, enabling practical uncertainty quantification and design workflows. Machine learning (ML) approaches to surrogate modeling have gained significant traction in the development of surrogate models due to their non-intrusive nature; they do not require access to the underlying simulation code. The Operator Inference (OpInf) approach, in particular, fits reduced operators to state and corresponding time derivative data. However, high-fidelity training data is scarce due to limited computational budgets, and OpInf models trained on such scarce data often exhibit high variance, a critical liability in high-risk engineering design. We propose Multifidelity Operator Inference (MF-OpInf), which reduces this variance by supplementing scarce high-fidelity snapshots with abundant low-fidelity data drawn from coarser discretizations of the same problem. MF-OpInf handles data across varying fidelity levels by utilizing a multifidelity linear regression framework to estimate the reduced operators. Numerical experiments on model PDE problems demonstrate that MF-OpInf predicts PDE solution fields with higher accuracy compared to the high-fidelity-only OpInf with the same computational budget.

AI 总结：针对对抗训练的“伪攻击”问题，本文提出基于Fisher几何的结构化置信度扰动对抗训练方法，通过混合目标与Fisher信息矩阵相关的高效算法，提升了对抗鲁棒性与鲁棒可迁移性。

链接：https://arxiv.org/abs/2611.14046

机构：Xi’an Jiaotong-Liverpool University（西交利物浦大学）； University of California, San Diego（加利福尼亚大学圣迭戈分校）； University of Exeter（埃克塞特大学）

作者：Chenlong Wei, Hanling Wang, Chong Zhang, Qinkai Yu, Fangyu Wu, Xiaobo Jin

英文摘要：Adversarial training is often plagued by "pseudo-attacks", perturbations that significantly increase the loss value but fail to truly alter the model's confidence structure. To address this, we propose two complementary approaches to directly attack and defend against confidence structure deformation. First, building upon the traditional loss maximization perspective, we introduce a hybrid objective that integrates cross-entropy and KL divergence, encouraging adversarial examples to simultaneously disrupt prediction accuracy and distribution consistency. Second, we geometrically reconstruct adversarial training by generating perturbations that maximize the distribution distance on the confidence manifold. Theoretical analysis shows that the optimal perturbation direction aligns with the principal eigenvectors of the Fisher information matrix. We also propose an efficient algorithm to avoid high-dimensional matrix computations. Experiments on multiple datasets and model architectures demonstrate that our method outperforms existing methods in adversarial robustness and robust transferability, particularly excelling in structural perturbations within high-confidence regions.

AI 总结：该研究针对神经过程（NPs）假设输入无噪声的缺陷，提出去噪神经过程（DnNPs），通过变分推断训练使其联合去噪与元学习，在PM2.5插值等任务中显著提升预测与不确定性量化效果，且有理论支撑。

链接：https://arxiv.org/abs/2611.14424

机构：Northeastern University（东北大学）

作者：Fatemeh Tohidian, Chengzhi Shi, Soomi Lee, Matthew Elia, Amy Mueller, Stratis Ioannidis, Jennifer Dy

英文摘要：Neural Processes (NPs) are flexible meta-learning models capable of uncertainty quantification across diverse applications. However, the standard NP framework assumes access to noise-free inputs. This is unrealistic in many practical settings, where sensor noise or measurement errors are unavoidable. This discrepancy between model assumptions and real-world data can significantly degrade prediction quality and uncertainty estimates. We introduce Denoising Neural Processes (DnNPs), a principled extension of the NP framework that explicitly models input uncertainty. We train DnNPs via variational inference. DnNPs jointly learn to denoise corrupted inputs and perform meta-learning, making them robust to noisy training data. We demonstrate DnNPs' effectiveness on both synthetic and real-world datasets, showing consistent improvements in prediction and uncertainty quantification compared to standard NPs. On PM2.5 spatial interpolation from real-world air quality sensors, DnNP reduces RMSE by up to 30% and NLL by up to 73% relative to standard NPs. We also provide theoretical grounding: we show that the Bayes-optimal predictor under noisy inputs belongs to the DnNP model class, and that plug-in models with fixed observation variance incur an irreducible KL divergence gap that DnNP provably avoids.

AI 总结：本研究提出仅基于蛋白-配体序列训练的基础模型Bonbon，其在6种规模缩放后涌现出三类零样本生物能力，结合与亲和力预测性能达SOTA且速度远超共折叠方法，湿实验验证对9个疾病靶点的全新化合物命中率达61.5%。

链接：https://arxiv.org/abs/2611.14504

机构：Output Labs（输出实验室）

作者：Adi Mashiach

英文摘要： Foundation models trained on large unlabeled corpora develop emergent capabilities, tasks that go beyond the training objective. We extend this paradigm to multimodal biology with Bonbon, a foundation model for molecular interactions trained on approximately 3.2 trillion tokens of paired protein-ligand sequences. From a single representation space, Bonbon exhibits three zero-shot emergent capabilities at hierarchical levels of resolution: functional (mechanism of action), residue (binding site localization, including orthosteric/allosteric discrimination), and bond (covalent versus non-covalent engagement). All three capabilities generalize to out-of-distribution complexes. Bonbon was trained on protein-ligand sequences alone, without any task-specific supervision and without structural coordinates. We conducted controlled scaling experiments across six model scales from 42M to 1.6B parameters; every capability evaluated shows a sharp jump at the largest scale on continuous metrics, even as pretraining loss decreases smoothly. This demonstrates that loss-based scaling laws alone are insufficient to predict when zero-shot emergent capabilities become available. The capabilities reported here emerge after training at a 1:2,000 parameter-to-token ratio to accommodate the sparsity of paired biological interaction data. The same frozen representations achieve state-of-the-art binding and affinity prediction at five orders of magnitude greater speed than co-folding methods. In prospective wet-lab validation across nine disease targets spanning seven protein families, Bonbon achieves a 61.5% dose-response hit rate on de novo compounds, yielding structurally novel molecules not previously reported to inhibit these proteins, including for out-of-distribution targets.

AI 总结：IC-GNN是一种以图谱为锚定的模型，可从fMRI数据推导个体化多尺度社区，在三类共超3万名被试的疾病预测中性能优于15种基线方法，且能识别疾病相关社区变化。

链接：https://arxiv.org/abs/2611.13070

机构：Yale University（耶鲁大学）； Brain Health Institute, Rutgers University（罗格斯大学脑健康研究所）； Pratt Institute（普拉特学院）； Department of Molecular Biophysics and Biochemistry, Yale University（耶鲁大学分子生物物理学与生物化学系）； Department of Computer Science, Yale University（耶鲁大学计算机科学系）

作者：Yang Zhang, Xiao Zhou, Jonathan Warrell, Avram Holmes, Xuan Zhang, Mark Gerstein, for the Alzheimer's Disease Neuroimaging Initiative

英文摘要：Brain networks exhibit modular organization, where community structures vary substantially across individuals and diseases. Understanding these individual differences is crucial for advancing both basic neuroscience research and precision medicine. However, widely used predefined group-level brain atlases often overlook this heterogeneity. While individualized functional parcellations capture subject-specific variability in community boundaries, they often require lengthy scan times to achieve fine-grained resolution and complicate cross-subject comparisons. To address this trade-off and investigate how individualized community organization captures disease-associated changes across individuals, we propose IC-GNN, an atlas-anchored model that forms distinct, individualized functional communities directly over atlas-defined brain regions. IC-GNN derives discrete, individualized multiscale communities from resting-state functional magnetic resonance imaging (fMRI) data, refines community-guided connections, and adapts shared regional weights to each subject's community context to enable both individualized modeling and group-level comparisons. Across more than 30,000 participants from ABIDE (autism spectrum disorder), ADNI (Alzheimer's disease), and the UK Biobank (depression), IC-GNN achieves the highest performance compared with 15 baseline methods. We further demonstrate that these individualized communities are reproducible within scans and across sessions, and carry predictive power on their own. Furthermore, IC-GNN identifies disease-related shifts in community organization that align with established literature, highlighting a promising approach for modeling individualized community structure for disease prediction and biomarker discovery.

AI 总结：针对现有图池化方法无法保留子结构及缺乏分层基准的问题，提出Pooled-Substructure-Aware GNN，引入带可控分层结构的合成数据集，在分层图任务上性能优于基线。

链接：https://arxiv.org/abs/2611.14126

机构：EPFL（洛桑联邦理工学院）； University of Oxford（牛津大学）

作者：Luca Sbicego, Xiaowen Dong, Dorina Thanou

英文摘要：Hierarchical graph-level tasks that require reasoning over substructures that extend beyond local neighborhoods are posing challenges for standard Graph Neural Networks (GNNs). While graph pooling aims to address this by coarsening graphs into hierarchical representations, existing methods often fail to explicitly preserve the substructures they induce, leading to inconsistent empirical performance. Moreover, evaluating such approaches remains challenging due to the lack of benchmarks with explicit hierarchical structure. We propose Pooled-Substructure-Aware GNNs, an approach to graph pooling that explicitly preserves and processes substructures uncovered by structure-based pooling schemes. We show that this design enhances the expressive power of pooling-based GNNs in the sense of the Weisfeiler-Lehman (WL) test. To enable principled evaluation, we introduce two new synthetic datasets with controllable hierarchical structure and propose a metric to characterize the prevalence of hierarchical patterns in real-world graph-level benchmarks. On tasks with relevant hierarchical structure, our approach outperforms matched pooling baselines, while remaining competitive on structurally simpler benchmarks. Code and datasets are publicly available at this https URL.

AI 总结：该研究针对医学图像分类，提出联合输出-归因正则化方法，验证了归因差异惩罚可补充输出层面公平性代理，在相关数据集的960个DINOv2-LoRA模型实验中，提升了效用-公平性权衡性能并降低了均等机会差异等指标。

链接：https://arxiv.org/abs/2611.13052

机构：Idiap Research Institute（伊迪亚普研究所）； École Polytechnique Fédérale de Lausanne (EPFL)（洛桑联邦理工学院）； Universidade Federal de São Paulo (UNIFESP)（圣保罗联邦大学）

作者：Gökhan Özbulak, Oscar Jimenez-del-Toro, Lilian Berton, André Anjos

英文摘要： Output-level fairness does not constrain whether demographic groups rely on similar predictive contributions in representation space. For a nonlinear encoder with a linear head, we show that outcome-conditioned output score disparity equals the signed sum of feature attribution differences across subgroups by allowing small prediction gaps to arise through cancellation effect. We test whether an attribution disparity penalty complements output-level equalized odds based fairness surrogates. Across 960 DINOv2-LoRA models on HAM10000 and Harvard-GF datasets for gender and race fairness, attribution augmentation increases utility-fairness trade-off performance for hypervolume in six different settings with paired 95% confidence intervals and relative gains of 1.49-10.57%. Under a predefined utility constraint, augmented model families also achieve 9.0-38.7% lower equalized odds difference, lower attribution disparity, and reduced cancellation ratio while preserving predictive utility. These results support attribution fairness as a complementary explanation-aware constraint on output fairness.

AI 总结：针对分布式在线策略后训练中LoRA适配器传输的带宽瓶颈，提出HiLo框架结合Spar梯度压缩器，仅传输1.6%适配器字节，加速更新最多2.2倍，同时保持学习质量。

链接：https://arxiv.org/abs/2611.13306

机构：Uppsala University（乌普萨拉大学）； Research Institutes of Sweden（瑞典研究院）

作者：Shenghui Li, Li Ju, Thiemo Voigt

英文摘要：Transmitting LoRA adapters to rollout workers can bottleneck distributed on-policy post-training over bandwidth-constrained networks. Lowering the training rank may limit the capacity a workload needs, while adaptive optimizer updates are difficult to compress. Rollout workers can instead reconstruct updates from gradients by maintaining replicated optimizer states. We introduce HiLo, a compressor-agnostic framework that replaces parameter transmission with compressed gradient dissemination and local optimizer replay. By applying the same decoded gradients, the learner and workers maintain consistent policy states at the same version under reproducible execution. We further develop Spar, a budget-aware gradient compressor that combines spectral truncation, adaptive rank allocation and error feedback to decouple training rank from communication cost. Experiments across on-policy distillation and reward-guided post-training show that HiLo with Spar maintains learning quality close to full synchronization while transmitting as little as 1.6% of the adapter bytes, and accelerates end-to-end updates by up to $2.2\times$ under bandwidth constraints.

AI 总结：本研究针对联邦LoRA的优化问题，提出SPRA-LoRA策略，通过冻结重采样$\boldsymbol{A}$矩阵与秩感知学习率，将开销减半且性能优于现有最优基线。

链接：https://arxiv.org/abs/2611.13331

机构：Stanford University（斯坦福大学）； Carnegie Mellon University（卡内基梅隆大学）； McGill University（麦吉尔大学）

作者：Saber Malekmohammadi, Kevin Kuo, Virginia Smith, Golnoosh Farnadi

英文摘要：Fine-tuning large-scale foundation models across resource-constrained, distributed clients presents significant computational and communication challenges. While merging Low-Rank Adaptation (LoRA) with Federated Learning (FL) enables rank-based flexibility, the interdependence of the adaptation matrices ($\mathbf{A}$ and $\mathbf{B}$) often complicates optimization. In this work, we theoretically demonstrate that LoRA fine-tuning with Gaussian initialization and a frozen $\mathbf{A}$ matrix acts as a stochastic approximation of full fine-tuning, where batch gradients are subject to a random, rank-dependent perturbation. Motivated by this insight, we propose freezing and periodically resampling $\mathbf{A}$ matrices at the client level. This strategy decouples the $\mathbf{A}$ and $\mathbf{B}$ gradients, diminishes LoRA low-rank constraints and reduces both computational and uplink communication overheads by 50\%, allowing for the reallocation of resources toward higher adaptation ranks. Furthermore, we show that in this frozen-$\mathbf{A}$ regime, the adaptation rank $r$ directly governs the signal-to-noise ratio (SNR) of the perturbed gradients. Specifically, we find that the SNR increases linearly with $r$, implying that clients' learning rates must be dynamically scaled with their adaptation ranks for higher utility. Building on these foundations, we introduce SPRA-LoRA (Stochastic Projection Rank-Aware LoRA), a federated rank-aware fine-tuning strategy that utilizes frozen, resampled $\mathbf{A}$ matrices and rank-dependent learning rates for clients. We also prove its convergence with heterogeneous client resources. Our experiments across benchmark datasets demonstrate that SPRA-LoRA consistently outperforms state-of-the-art baselines with half the overhead on clients.

AI 总结：该研究提出共形风险控制框架，在技术驱动域偏移下为感知QoT的光通路可行性评估提供不确定带误覆盖率的有限样本保证，提升决策准确率且无需重新训练。

链接：https://arxiv.org/abs/2611.13893

机构：Chalmers University of Technology（查尔姆斯理工大学）； University of Applied Sciences and Arts of Southern Switzerland（瑞士南部应用科学与艺术大学）

作者：Kiarash Rezaei, Omran Ayoub, Paolo Monti, Carlos Natalino

英文摘要： Artificial intelligence and machine learning (AI/ML)-based regression models support operational decision-making in optical networks but typically expose only point predictions, which carry no statistically controlled information about their reliability. This paper proposes a post-hoc conformal risk control (CRC) framework that keeps a pretrained AI/ML regressor fixed, provides a finite-sample guarantee on uncertainty band miscoverage, and maps the band to operational decisions through policy rules. The guarantee applies to interval miscoverage, while decision-error rates are evaluated empirically. We instantiate the framework for quality of transmission (QoT)-aware lightpath (LP) feasibility assessment, in which a regressor predicts optical signal-to-noise ratio (OSNR) under domain shifts caused by fiber, amplifier, and switching upgrades. Feature shift contribution scores identify the LP features whose relation to QoT changes the most across a technology transition, and two policies translate the bands into feasibility decisions, allowing operators to tune the trade-off between conservatively rejecting or aggressively accepting LP realizations. Experiments on open space-division multiplexing (SDM) QoT datasets over the CONUS topology show that, for a target coverage of 90%, the uncertainty bands meet the target in all 170 trials spanning 17 domain shifts and ten calibration-test partitions, whereas a source-only quantile regression (QR) baseline meets it in none. Under severe shifts, the efficiency-oriented (EO) policy empirically improves zero-shot (ZS) feasibility-decision accuracy from about 85% to above 99%, approaching a transfer learning (TL) policy while avoiding retraining. The framework provides finite-sample control of uncertainty band miscoverage for regression-based optical network decision support, with potential applications beyond QoT-aware LP feasibility assessment.

AI 总结：本文提出单极神经元网络及循环学习规则，结合任务代码与巩固机制减少顺序学习的遗忘，在MNIST和Iris序列任务上，该网络准确率高且遗忘程度远低于梯度训练的网络。

链接：https://arxiv.org/abs/2611.14153

作者：Florin Leon

英文摘要：Unipolar neurons are the predominant type of nerve cell in the octopus and in most other invertebrates. A single neurite (a fiber extending from the cell body) branches into a tree that carries both input and output synapses. The branches send graded local outputs to nearby cells, and one site on the main neurite sends a global output to distant targets. This paper proposes a network of such neurons and a learning rule for it. Each neuron is modeled as a tree of passive segments, whose activations are computed exactly as the steady state of a linear system. The network learns from reward alone, with a rule called recirculation. After each forward pass, a second pass sends requested changes of the outputs back through the same connections. Each synapse then adjusts its weight using only quantities available at its own location. Because the system of each neuron is symmetric, this update closely follows the gradient of the error in the predicted reward. Two mechanisms limit forgetting when tasks are learned in sequence: a task code inhibits a different subset of branches in each task, while consolidation lowers the learning rate of synapses that changed strongly in the recent past. The model is evaluated on a sequence of MNIST and Iris tasks. The network maintains high accuracy with low forgetting, whereas gradient-based training of the same network forgets earlier tasks almost completely.

AI 总结：本文探究循环TSFMs中未被利用的初始状态的作用，以TiRex为案例，通过训练小型头将上下文向量写入初始状态，发现其可提升预测精度，为TSFMs研究提供新方向。

链接：https://arxiv.org/abs/2611.14357

机构：NXAI GmbH（NXAI公司）； Institute for Machine Learning, Johannes Kepler University（约翰内斯·开普勒大学机器学习研究所）

作者：Bernhard Voggenberger, Christian Ganhör, Levente Zólyomi, Sebastian Böck, Sepp Hochreiter

英文摘要：Time series foundation models (TSFMs) forecast unseen series without task-specific training. Increasingly they are built on recurrent backbones, whose constant per-step cost makes long-context inference cheap, and which, unlike transformers, carry a state. A recurrent backbone computes its prediction from two inputs: the observed context window and the initial recurrent state. In practice only the first is utilised and the initial state is unused by convention. We ask whether the initial state can be utilised in recurrent TSFMs. Using the xLSTM-based zero-shot forecaster TiRex as a case study, we freeze the backbone and train only a small head that writes a context vector into the initial state, so that any change in accuracy is attributable to the injected state rather than to adapted weights. A spectral summary of the input window, learned once and applied unchanged to every series, lowers mean GiftEval MASE by 0.175%, with every seed improving, which neither a parameter-matched LoRA adapter nor a learnable static state reproduces. An embedding of text attached to each series, evaluated on six MoTime datasets, reveals two roles for the state. As a per-dataset adaptation surface, a static state lowers MASE by 0.04-6.96% of the zero-shot baseline (8-70% of the gain of full fine-tuning). As a channel for exogenous information, the text-conditioned state lowers it by a further 0.15-0.41% of the zero-shot baseline on four of six datasets and, combined with a LoRA adapter, exceeds full fine-tuning, which has no access to the text, on one of them. In TiRex, the effect concentrates in the mLSTM blocks and is strongest at short context. The results are preliminary and the improvements small, but they point to the initial state as a direction for further research.

AI 总结：本研究从间隔视角解释LoRA对灾难性遗忘的鲁棒性及适配器合并特性，提出MaLoRA合并规则，经实验验证其在多类分类任务上的性能优于或相当于现有基线。

链接：https://arxiv.org/abs/2611.14397

机构：University of Pennsylvania（宾夕法尼亚大学）； Shanghai Jiao Tong University（上海交通大学）； Johns Hopkins University（约翰斯·霍普金斯大学）

作者：Ziqing Xu, Hancheng Min, Lachlan Ewen MacDonald, Salma Tarmoun, Enrique Mallada, Weijie Su, Rene Vidal

英文摘要： Low-rank adaptation (LoRA) is the de facto method for parameter-efficient fine-tuning of large neural networks, achieving strong task adaptation with minimal trainable parameters and low optimization cost. Beyond this efficiency, practitioners have observed two striking properties: \textit{robustness to catastrophic forgetting} and the ability to \textit{merge independently trained adapters} into a single model that performs competitively across multiple tasks. These properties are central to continual and multi-task learning, yet remain poorly understood. In this work, we provide a theoretical explanation through a unified \textit{margin-based perspective}. We analyze LoRA in multiclass linear classification under a near-orthogonal task regime with $\ell_2$-regularization, and characterize the optimal LoRA adapter across regularization regimes. Our analysis shows that, in an intermediate regime of regularization parameters, the optimal adapter aligns with the \emph{max-margin solution} on the fine-tuning data. Building on this characterization, we derive two key consequences. First, we obtain closed-form expressions for the margins on pre-training and fine-tuning data, revealing a precise margin trade-off: the regularization parameter controls the balance between retention and adaptation. Second, we analyze adapter merging, proving that merged models achieve positive margin on each task and deriving optimal mixing coefficients that balance margins across tasks and maximize the margin over their union. These results lead to a simple, training-free merging rule, which we term \textit{margin-based LoRA merging} (MaLoRA). Experiments on modern architectures and real datasets validate our theoretical predictions, showing that MaLoRA matches or outperforms several adapter-merging baselines across a range of vision and language classification tasks.

AI 总结：本文提出Continual-ARC终身任务流基准，对比智能体工具与权重学习，发现智能体工具的持久性需专家指导才提升记忆，且未提升泛化，权重学习在未见过任务上增益显著。

链接：https://arxiv.org/abs/2611.14505

机构：unify（Unify公司）

作者：Daniel Lenton

英文摘要：Agent harnesses have emerged as a promising mechanism for continual learning, alongside test-time parametric learning. However, the current benchmarks have several limitations: they generally focus on a single applied domain, serve each instance only once, and assume a single family of learner, so memory-based and weight-based systems are rarely compared on the same scale. We propose Continual-ARC, a lifelong stream built from the three ARC-AGI editions, whose tasks need no domain knowledge. The same tasks return after dormant gaps with fresh instances, while new tasks keep arriving. On each instance, the learner may pay for demonstrations, submit answers, or act inside the harness, with retries permitted on incorrect submissions, which are paid for. Each instance is scored based on the share of this feedback budget consumed. Since a learner may keep anything between instances, memory-based and weight-based systems are scored on the same scale. After the stream, the learner is frozen and tested on fresh instances of the tasks it met, which measures memorization, and on unseen tasks, which measures generalization. We run an agent harness with frozen weights, with and without persistence, on Qwen3.5-9B and GLM-5.3 Flash, comparing them with two prompting floors, a program-induction method and Qwen3.5-9B's weights trained to imitate GLM-5.3 Flash. Persistence in the harness generally does not improve memorization unless provided with explicit expert guidance. Most harness-based methods still learn less than a floor that simply caches each task's demonstrations, across all models we tested. No harness-based method gains on unseen tasks, whereas weight-based learning shows substantial gains. This opens the door to future weight-based continual learning that could teach models to drive harnesses in new domains.

AI 总结：本综述聚焦多模态图学习领域，提出统一消息传递公式划分方法层级，综述基准并关联智能体与图工程，探讨开放问题与研究方向。

链接：https://arxiv.org/abs/2611.12990

机构：North Carolina State University（北卡罗来纳州立大学）

作者：Rui Xue, Tianfu Wu

英文摘要：Multimodal attributed graphs (MAGs), in which nodes carry text, images, and other modalities while edges encode relations among entities, underlie e-commerce, social media, content platforms, and creative applications. Learning on MAGs has developed along two complementary lines: graph-centric methods redesign how modalities are fused, propagated, and connected within message passing, while foundation-model-centric methods pretrain transferable multimodal graph encoders or bring graph context into multimodal large language models for prediction and generation. In parallel, several benchmarks have appeared with different datasets and evaluation protocols. This survey takes an algorithm-centric and evaluation-grounded view of the field and makes three contributions. Modeling: we introduce a unified message-passing formulation and organize methods into five integration levels (feature, propagation, topology, learning, and foundation model), covering discriminative and graph-conditioned generative models; within this formulation we derive representative methods as special cases, compare their assumptions, expressiveness, and cost, and identify cross-modal sensitivity as a property that separates otherwise similar architectures. Datasets and evaluation: we review recent benchmarks, summarize their datasets and metrics, define model-independent descriptors (label homophily, average degree, clustering, and edge provenance), and map each evaluation protocol to the baselines it uses. Frontiers: we connect multimodal graph learning to AI agents, covering the multimodal graphs that agents perceive and remember and graph engineering, the practice of organizing agent workflows as explicit graphs; we then discuss open problems, concrete efforts for the community, and research directions, including when graph structure helps in the presence of strong multimodal foundation models.

AI 总结：本研究推出One Table基准测试，对比通用表格模型在生成、插补、预测任务的表现，发现部分通用模型效用优于专家组合，冻结基础模型在三项任务具潜力，推动统一表格模型发展。

链接：https://arxiv.org/abs/2611.13001

机构：StableAI（稳定人工智能公司）； Tsinghua University（清华大学）； Beijing University of Posts and Telecommunications（北京邮电大学）； Southeast University（东南大学）

作者：Yuanrui Wang, Kaifang Li, Xuanyue Li, Ziyun Li, Yan Lu, Xingxuan Zhang, Peng Cui

英文摘要： Generation, imputation, and prediction query different aspects of the same tabular distribution, but are usually evaluated separately. This leaves unclear which capabilities they share, how their training objectives interact, and whether one model can perform all three well. We introduce the One Table Benchmark (\ours{}), spanning 56 datasets, 46 registered methods, and 20 capability axes under aligned data splits and explicit task contracts. We compare task ranks on aligned support, train five backbones under all seven non-empty objective combinations on five datasets with three seeds, and contrast fixed-model profiles with specialist portfolios. Rank comparisons connect imputation and prediction through numerical recovery and dependence preservation, and link generation to both through predictive utility; distributional fidelity follows different, sometimes opposing, rankings. Objective interventions show the largest pooled gains from including the recipient's own objective. Generation has the broadest coverage of positive pooled auxiliary effects, while prediction shows potential to improve imputation and synthetic-data utility, with responses that vary across backbones. Cross-addition effects reveal that objective benefits need not add and depend on the objectives already present. Fixed-model comparisons expose corresponding utility--fidelity trade-offs: some generalists surpass specialist portfolios on utility, while fidelity continues to favor specialization. Frozen foundation models show strong prediction and numerical point imputation, together with promising synthetic-data utility, highlighting their potential across all three tasks. These findings motivate unified tabular models that share useful conditional information while improving categorical reconstruction and faithful joint sampling.

AI 总结：本研究提出SLIC框架，训练4B规模可本地部署的语言模型用于多轮模拟电路优化，在多项基准测试中超越现有大模型，且可跨工艺节点迁移，支持本地部署无需传输专有数据。

链接：https://arxiv.org/abs/2611.13452

机构：MIT（麻省理工学院）； Analog Devices, Inc.（亚德诺半导体公司）； IBM Research（IBM研究院）

作者：Yan Xu, Zhengqi Gao, Ching-Yun Ko, Zhang-Wei Hong, Wenjie Lu, Tao Yu, Xin Zhang, Duane S. Boning, Ruonan Han

英文摘要：Analog integrated circuit (IC) design is a multi-turn, simulator-grounded engineering task that demands iterative reasoning over coupled performance tradeoffs, yet proprietary process design kits (PDKs) make hosted large language model workflows difficult to deploy in practical design settings. We present SLIC (Small LM Iterating on Circuits), a reinforcement fine-tuning framework that trains a 4B-scale, locally deployable language model for closed-loop analog circuit optimization. SLIC pairs structured circuit representations with discretized relative-update actions, domain- and success-gated multiobjective rewards, and Reward-Guided Rollout Steering (RGRS), which turns per-step simulator feedback into signals for both policy learning and online rollout construction in this sparse, highly constrained domain. We further release an open, goal-conditioned benchmark of 41 amplifier topologies with standardized protocols for in-distribution validation, held-out topology evaluation, and cross-technology-node transfer. Trained solely with open source 65 nm CMOS predictive technology models (PTM), SLIC-Random achieves 87.1% in-distribution four-specification Pass@1 within eight turns, surpassing the 763B DeepSeek-V4.1-Flash Think by 48.5 percentage points. Beyond the training distribution, SLIC-Softmax reaches 72.6% Pass@1 on held-out topologies, and the trained policies consistently transfer across five unseen technology nodes without additional fine-tuning. This transfer extends to a commercial TSMC 65 nm CMOS PDK, where SLIC-Random achieves 49.9% held-out-topology Pass@1, demonstrating practical deployment through local inference without transmitting proprietary data to external LLM services. Performance further scales with test-time simulator interactions, reaching up to 83.2% held-out-topology Pass@1 at a 32-turn budget.

AI 总结：本研究推出大规模基准Chem World及物理感知成分建模PACM，解决分子性质预测基准碎片化问题，实验证实PACM可提升预测性能，为AI辅助生物医学及药物研发提供基础。

链接：https://arxiv.org/abs/2611.13945

机构：Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）； School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences（中国科学院大学先进交叉科学学院）； Cleer Science（Cleer科学公司）； Imperial College London（伦敦帝国学院）； Tsinghua University（清华大学）

作者：Tianyou Bai, Huan Wang, Mingchen Gao, Tongrui Yang, Fangyue Lin, Pinze Ren, Zhenlin Zhao, Siming Dong

英文摘要：Molecular property prediction is fundamental to computational chemistry and drug formulation research, yet existing benchmarks remain fragmented across datasets, properties, and evaluation protocols, limiting systematic evaluation of model generalization. We introduce Chem World, a large-scale benchmark integrating 17 public datasets with over 800,000 molecular and chemical-system instances across 10 property tracks, covering single-molecule and multi-component systems, including drug solubility. Chem World provides a unified testbed for evaluating molecular representations and compositional modeling across diverse chemical properties. Building on the benchmark, we propose PACM (Physics-Aware Compositional Modeling), which combines molecular representations with composition-aware modeling and domain-specific physical constraints. Extensive experiments show that molecular language models provide strong and transferable representations, while PACM consistently improves prediction across diverse chemical systems. Chem World and our PACM-powered drug-property prediction Agent support generalizable chemical modeling and offer a foundation for AI-assisted biomedical research and drug formulation development.

AI 总结：本研究提出BioXArena-Pro基准测试集，包含64项生物医学建模任务，测试22种智能体配置，发现脚手架与模型同等重要，预训练模型可提升性能，任务经多轮审核验证。

链接：https://arxiv.org/abs/2611.13994

机构： Mohamed bin Zayed University of Artificial Intelligence（穆罕默德·本·扎耶德人工智能大学）； The Westminster Schools（威斯敏斯特学校）； Carnegie Mellon University（卡内基梅隆大学）

作者：Loka Li, Duzhen Zhang, Xingbo Du, Leonard Song, Kun Zhang, Le Song

英文摘要：Coding agents are increasingly proposed as autonomous builders of biomedical models, but existing benchmarks are small, near saturation or open to leaked information. We present BioXArena-Pro, 64 research-level modelling tasks from eight biomedical domains, each rebuilt from a published study; most hold out whole biological units such as species or cohorts. Across 22 agent configurations, the best scores 55.1 of 100, and the scaffold matters almost as much as the model. We re-ran a published method for every task: where the published recipe applies unchanged, the best agent and the methods tie; on rebuilt tasks the methods trail by 8.2 points, suggesting that agents gain by adapting to new data. Pretrained models (including biological foundation models) matter where the strongest agent uses them: forbidding them decreases its score on those tasks by 9 points. Every task was audited in several AI-assisted rounds and a final human review.

AI 总结：该研究针对解释引导学习（EGL），通过构建含不同类型捷径的数据集，发现EGL效果具目标特异性，不同EGL目标在鲁棒性、合理性及行为修正上表现各异，且EGL成功需结合多维度评估。

链接：https://arxiv.org/abs/2611.14134

机构：German Research Center for Artificial Intelligence (DFKI) GmbH（德国人工智能研究中心）； RPTU University Kaiserslautern-Landau（凯撒斯劳滕-兰道理工大学）

作者：David Dembinsky, Adriano Lucieri, Andreas Dengel

英文摘要：Explanation-Guided Learning (EGL) promises models that are not only accurate but also rely on task-relevant evidence. However, whether improved agreement between model explanations and human-provided rationales also implies a changed decision strategy remains unclear. We study this tension under controlled synthetic shortcut introduction using a constructed shortcut dataset. Inspired by real-world applications, our benchmark covers shortcuts of different kinds, including localized, background-based, and global cues. This allows us to evaluate whether different EGL objectives produce behavioral changesand explanatory changes. Across these axes, we find that the effects of EGL are strongly objective-specific. Rationale-based data augmentation provides the clearest robustness gains when shortcut evidence can be separated from task-relevant foreground information, reaching average accuracy improvements of up to +4.8 percentage points (pp) together with increased foreground reliance. Attribution-constraining objectives primarily improve explanation plausibility, increasing Importance Mass by up to +14 pp on average, but provide weaker and less consistent behavioral correction, with average accuracy gains of at most +1.6 pp. Moreover, neither improved robustness nor improved plausibility reliably implies higher perturbation-based fidelity. Overall, our results show that EGL can be useful, but its success cannot be inferred from rationale agreement alone and should instead be evaluated along complementary behavioral and explanation-level axes. We publicly release all experimental results to enable further analysis ( this https URL ).

AI 总结：该研究提出MMDM基准，含5448个16类任务问题，评估28个决策模型，发现其在常识任务表现好但在逻辑、约束等决策任务上不足，为决策模型发展提供测试平台。

链接：https://arxiv.org/abs/2611.14309

机构：The Chinese University of Hong Kong（香港中文大学）； Hong Kong University of Science and Technology（香港科技大学）

作者：Junhao Shen, Qing Zong, Yue Guo, Zhanqiu Zhang, Yiwen Guo, Hong Cheng

英文摘要：System-1 Decision Models directly make structured probabilistic decisions over query-specified options given a state in a single forward pass, without autoregressively generating reasoning trajectories or textual responses. Recent evaluations cover diverse tasks but often rely on existing benchmarks, where prior exposure to test material can obscure generalization. Moreover, real-world decision making often involves long contexts, large option spaces, and complex logical dependencies, challenges that existing benchmarks do not adequately capture together. Thus, we introduce MMDM (Massive Multitask Decision Making), a benchmark of 5,448 high-quality questions across 16 task categories, combining newly constructed tasks with source-grounded recompositions and rewrites for single-select, multi-select, and soft-label decision making in fresh contexts. It covers realistic scenarios such as agent state tracking, tool and skill selection, preference judgments, and rubric scoring, with option spaces ranging from 2 to 1,024 and inputs extending to 64,817 tokens. With MMDM, we conduct a comprehensive quantitative evaluation of 28 decision models, showing that many perform well on common knowledge yet struggle with interdependent options, constrained decisions, and multi-select tasks. For example, Perplexity Decider-V1.2 reaches 78.75% Hard Single-select Accuracy on world knowledge, but only 53.95% on option logic and 47.95% on constraint planning, while the best Hard Multi-select Accuracy across all models is 36.81%. These results highlight joint reasoning, constraint-aware selection, and complete multi-select recovery as key priorities for decision model development, with MMDM providing a testbed for measuring this progress.

AI 总结：该研究针对基础式时间序列预测模型在观测数据有限时难以暴露时间结构的问题，提出StatPrompt-LLM，通过结合Time-LLM提示机制与数据集统计信息、引入零初始化自回归旁路，在10%缩减前缀配置下降低了预测误差。

链接：https://arxiv.org/abs/2611.12974

机构：The University of Sydney（悉尼大学）

作者：Aniket Sinha, Veronica W. Tsou

英文摘要： Time-series forecasting has recently seen a paradigm shift with the introduction of foundation-style models that adapt large pretrained language architectures to sequential data. However, effectively exposing fundamental temporal structures, such as trend, scale, and temporal dependence, to these models remains an open challenge, particularly when the available observations are limited. In this paper, we conduct an exploratory empirical study across different generations of forecasting architectures and propose StatPrompt-LLM, a lightweight statistical refinement of Time-LLM. StatPrompt-LLM augments the existing Time-LLM prompting mechanism with dataset-level statistical information and introduces a zero-initialised autoregressive bypass as a complementary linear forecasting pathway. We evaluate our framework on electricity and meteorological time-series benchmarks, including ETTh1, ETTh2, and Weather, under full-prefix and reduced-prefix configurations. Under the 10% reduced-prefix configuration, StatPrompt-LLM achieved lower in-sample prediction errors than Time-LLM, with unweighted average relative reductions across datasets of 40.2% in MSE and 21.4% in MAE. Full-prefix results were more dataset-dependent. Ablation analyses revealed relatively modest and dataset-dependent changes when selected additional statistical descriptor groups were masked within the combined architecture. Computational profiling indicated additional model parameters and runtime overhead relative to Time-LLM.

AI 总结：该研究以IGRA探空数据为对象，对比表格学习器与解析参数化方法重建边界层风廓线，发现预测因子决定精度，傅里叶神经算子可提升近地面预测精度，建议通过新预测因子优化模型。

链接：https://arxiv.org/abs/2611.13359

机构：HSE University（高等经济大学）； Generalizable AI Systems（通用人工智能系统公司）； MIRAI（未来智能研究院）

作者：Fedor Buzaev (1), Dmitry Efremenko (1), Daria Pugacheva (1 and 2), Fedor Ratnikov (1), Egor Bugaev (1), Vladimir Gordin (1 and 3), Denis Derkach (1) ((1) HSE University, (2) Generalizable AI Systems, (3) MIRAI)

英文摘要：Wind profiles in the atmospheric boundary layer (ABL) drive short-range forecasting and wind-energy siting, yet sparse radiosonde coverage forces practitioners to interpolate them from a handful of physical predictors. We benchmark modern tabular learners against analytical parameterizations on 158,097 IGRA soundings. CatBoost, TabM, and FT-Transformer reduce RMSE to 1.3-1.6 m/s, against approximately 3 m/s for power-law and Monin-Obukhov fits; once the predictor set is fixed the three architectures differ by less than 0.05 m/s, so accuracy is driven by predictors rather than architecture. A cohort analysis of within-bin profile dispersion places an empirical floor of approximately 1.77 m/s on any point-wise predictor that uses standard radiosonde features, consistent with prior eddy-viscosity inversions. A one-dimensional Fourier Neural Operator applied as a post-hoc corrector to frozen tabular predictions reduces near-surface RMSE on U.S. stations from 1.5 to 1.2 m/s, while leaving upper-layer error unchanged. Further gains should come from new predictors rather than larger models.

AI 总结：本文提出AnchorMTS，通过从观测窗口导出自对齐的视觉与文本视图完善数值预测，在6个基准数据集上优于多数基线，验证了多视图证据对长期时间序列预测的提升作用。

链接：https://arxiv.org/abs/2611.13794

机构：College of Control Science and Engineering, Zhejiang University（浙江大学控制科学与工程学院）； Zhejiang Provincial Security Science and Technology Research Institute（浙江省安全科学技术研究院）

作者：Gao Wu, Tingting Xu, Changning Wu, Tengfei Long, Yong Liu, Jiabin Jin, Kexin Zhang

英文摘要：Long-term time-series forecasters typically optimize a single numerical representation, even though the same history exposes complementary geometry, recurrence, and semantic structure. We propose AnchorMTS, which derives self-aligned visual and textual views from each observed window and uses them only to refine a complete numerical forecast. A trainable evidence path produces a candidate forecast, while a parameter-free output projection enforces an exact per-variable bound on its displacement from the numerical anchor. Across six benchmarks and four horizons, AnchorMTS achieves the best macro MSE/MAE (0.336/0.367) among seven supervised baselines and wins 18 of 24 MSE settings against the cellwise baseline envelope; removing the auxiliary path raises macro MSE to 0.340. The projection leaves 95.8% of candidate refinements unchanged while yielding zero bound violations, showing that complementary multi-view evidence can improve forecasting while its final influence remains explicitly controlled.

AI 总结：该研究证明了元辛神经算子在非空有界Lipschitz域上的通用逼近性，仅需一个含非局部分支的隐藏算子层即可实现指定空间间连续算子的紧一致逼近，结果涵盖傅里叶等乘子构造。

链接：https://arxiv.org/abs/2611.13991

机构：RICAM, Austrian Academy of Sciences（奥地利科学院RICAM研究所）

作者：Ahmed Abdeljawad, Malek Yacine Rezoug

英文摘要：We prove universal approximation for metaplectic neural operators on nonempty bounded Lipschitz domains. The nonlocal branches are whole-space metaplectic multipliers with zero extension and restriction, associated with any fixed free symplectic matrix. Assuming a smooth, globally Lipschitz, nonpolynomial activation, we obtain compact-uniform approximation of continuous operators from $C^s$ to $C^r$ and from $H^s$ to $H^r$ in the respective output norms, for arbitrary nonnegative integers $s,r$ and finite input and output channel dimensions. In the Sobolev setting, compact input sets must additionally be uniformly bounded in $L^\infty$. One hidden operator layer with one nonlocal branch suffices. A single scalar symbol atom and finitely many positional features are prescribed by the transform and domain independently of the target operator, compact set and accuracy. The proof combines averaging neural operator theory with an exact chirp-conjugated averaging identity and derivative estimates. The result includes Fourier and fixed nondegenerate fractional Fourier multiplier constructions.

AI 总结：HaiRuo模型基于卫星测地基础模型，同时预测海洋涡旋的场与对象，在15天涡旋中心误差等指标上优于14种方法，大幅降低了涡旋预测误差

链接：https://arxiv.org/abs/2611.14136

机构：Ocean University of China（中国海洋大学）； Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）； University of Science and Technology of China（中国科学技术大学）

作者：Xukun Qin, Haochen Sun, Yuyang Tian, Ge Chen

英文摘要：Mesoscale eddies carry most of the ocean's kinetic energy and move heat, salt and nutrients across basins, yet where an eddy will be two weeks ahead is still read off ocean models with post-hoc tracking heuristics. We introduce HaiRuo, which forecasts eddies as fields and as objects at once: from 30 days of satellite sea-level maps, one forward pass returns 15 days of sea level and geostrophic velocity, daily eddy masks and the trajectory of every eddy alive today. HaiRuo pairs two components with a clear division of labour: (1) HaiRuo-1B, an open-weight 1.05-billion-parameter foundation model pretrained without labels on 24 years of global satellite-altimetry observations, which supplies the 15-day physical fields and masks (+0.14 field skill and +1.1 mIoU points under the same decoder) and further lowers 15-day centre error by 1.2 km frozen (1.35 km over training from scratch); and (2) XuanJi, a Lagrangian object decoder that samples the evolving forecast in each eddy's co-moving polar frame and refines the 15-day path across three look-ahead rounds, removing 2.5 km on HaiRuo-1B (2.9 km with joint fine-tuning) and 2.5-3.3 km via its object attention across six published backbones. Under one pre-registered global evaluation over the whole ocean, HaiRuo improves on fourteen methods on every endpoint (25.2 vs. 30.6 km 15-day centre error, -18%), holding its margin outside the labelled regions and through both phases of the El Nino-Southern Oscillation (ENSO), while reproducing westward drift, 50-60 deg S eastward reversal and poleward cyclone-anticyclone deflection. Over ten days through one fixed tracker, HaiRuo's forecast field cuts the eddy-centre error of WenHai and XiHe by 37-43% (-2.8 km vs. day 0 while they degrade +4.3-8.7 km), and its object trajectories cut pipeline error by 77-78% globally. Open-weight checkpoints, code and project page: this https URL

AI 总结：针对现有时间序列基础模型的两大缺陷，本文提出TimeSeek模型，通过多领域特征提取器与多任务协同训练实现多领域自适应建模与多任务统一预训练，在三类任务中达最优性能且推理效率良好。

链接：https://arxiv.org/abs/2611.14491

机构：School of Data Science and Engineering, East China Normal University（华东师范大学数据科学与工程学院）

作者：Peng Chen, Yihang Wang, Zhongwen Rao, Yang Shu, Yuying Qiu, Yuxuan Chen, Lujia Pan, Bin Yang, Chenjuan Guo

英文摘要：Time series analytics has been widely applied across various domains. While existing time series foundation models (TSFMs) have demonstrated promising results, they still face two major challenges: 1) Most TSFMs mainly rely on stacking Transformer layers to increase parameter size, lacking the ability to adaptively model complex time series patterns from multi-domain data. 2) Current TSFMs are primarily pre-trained for specific tasks, e.g., forecasting or anomaly detection, and a few models support multiple tasks, but they fail to fully leverage task specificity and commonality, resulting in suboptimal performance. To address these challenges, we propose TimeSeek, a novel generalized time series analytics model that adaptively models time series patterns from multi-domain data and unifies multiple tasks in pre-training. Conceptually, TimeSeek is guided by selective information sharing under domain and task heterogeneity. From the perspective of multi-domain data, we propose a Multi-Domain Feature Extractor that adaptively captures temporal dependencies and channel correlations based on domain-specific characteristics, including progressive multi-scale modeling and adaptive slot channel modeling. From the perspective of the diversity of tasks, we propose multi-task collaborative training, which leverages task specificity and task commonality through task customization and cross-task training, effectively unifying multiple tasks in pre-training and better adapting to diverse downstream scenarios. After pre-training on large-scale time series data, TimeSeek achieves state-of-the-art performance in forecasting, anomaly detection, and classification tasks across both few-shot learning and zero-shot situations, together with favorable inference efficiency.

AI 总结：本研究构建了蛋白质-配体姿态排序基准Pyrite，开发了PyriteReasoner和PyriteFeaturizer两种LLM智能体，其排序准确率超专家水平，证明LLM引导的特征发现是有前景的计算化学工具输出组合方法。

链接：https://arxiv.org/abs/2611.14511

机构：Stanford University（斯坦福大学）； Genesis Molecular AI（创世纪分子人工智能公司）

作者：Megha Srivastava, Michael H. Lim, Allison M. Keys, Pranav Murugan, Nate Gruver

英文摘要： Predicting how a ligand binds to its target is central to small-molecule drug discovery. Modern cofolding models can generate diverse candidate poses, but reliably ranking these candidates remains challenging. We introduce Pyrite, a pairwise pose-ranking benchmark comprising 1,500 pose pairs across diverse protein-ligand complexes. Expert chemists equipped with 3D visualization tools and physics-based scores perform near chance, highlighting the difficulty of interpreting available chemical signals. To investigate automated ranking, we introduce PyriteRep, a structured representation of molecular geometry, protein-ligand interactions, and physicochemical scores, and develop two LLM-based agents: PyriteReasoner, which directly compares poses through text-based reasoning, and PyriteFeaturizer, which generates executable code to construct features for logistic regression, optionally incorporating features derived from LLM subcalls. These agents achieve up to 67% and 74% accuracy, respectively, exceeding observed expert performance. LLM-guided feature discovery generally outperforms direct reasoning, while Group Relative Policy Optimization (GRPO) improves open-weight models, enabling them to match or surpass closed-source frontier models in our evaluations. These results establish Pyrite as a challenging test of pose-ranking ability and highlight LLM-guided feature discovery as a promising approach to interpret and combine the outputs of computational chemistry tools.

AI 总结：针对DSA索引延迟过高的问题，提出ADIndex方法，通过层校准、组级蒸馏和低延迟推理三个组件优化，在保持任务性能的同时，大幅提升长上下文推理效率。

链接：https://arxiv.org/abs/2611.12956

机构：Nanjing University（南京大学）； Alibaba Group（阿里巴巴集团）

作者：Jin-Fei Qi, Yiduo Li, Jiang-Xin Shi, Kan Liu, Tao Lan, Hanlin Tang, Yu-Feng Li

英文摘要：DeepSeek Sparse Attention (DSA) reduces the computational cost of attention through token-level selection. However, its indexer still scores the entire causal prefix for each query, making indexing a dominant source of latency in long-context inference. Notably, the indexer alone can account for more than 50% of total prefill GPU time at long context lengths. To address this bottleneck, we propose ADIndex, an Adaptive and Distillation-driven Indexing method for token-level sparse attention. ADIndex integrates three key components: (i) a layer-wise calibration strategy that adaptively accommodates cross-query retrieval overlap, motivated by our observation of pronounced inter-layer heterogeneity; (ii) a group-level distillation scheme that leverages this heterogeneity to efficiently train adaptive layer-wise group indexers, capturing salient retrieval signals from individual queries; and (iii) a low-latency inference strategy that reduces repeated full-prefix scans through token bank sharing within query groups while retaining query-specific refinement. Experiments on DeepSeek-V3.2 and GLM-5.1 demonstrate that ADIndex preserves near-lossless task performance on LongBench, LongBench v2, and RULER while delivering substantial efficiency gains over standard DSA, including prefill and decoding kernel speedups of up to 6.22x and 3.40x, respectively, at a context length of 256K across the layer-wise reuse configurations, as well as an end-to-end serving speedup of 1.54x at a context length of 128K.

AI 总结：该研究针对脉冲相机图像重建中SNN方法保真度不足的问题，提出MADSNet架构，经实验验证其在SREDS数据集上的重建性能更优，且延迟更低、能耗更少。

链接：https://arxiv.org/abs/2611.12980

机构：School of Data Science, The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）数据科学学院）； School of Computer Science, Peking University（北京大学计算机学院）； College of Computing and Data Science, Nanyang Technological University（南洋理工大学计算与数据科学学院）； School of Intelligence Science and Engineering, Harbin Institute of Technology, Shenzhen（哈尔滨工业大学（深圳）智能科学与工程学院）

作者：Sirui Li, Rui Zhao, Siqi Cai, Haizhou Li, Tiejun Huang, Yajing Zheng

英文摘要：Spike cameras encode dynamic scenes into high-temporal-resolution binary spike streams for efficient high-speed vision. Although spiking neural networks (SNNs) can process such signals with sparse and accumulate-only computation, existing SNN-based reconstruction methods are typically shallow and still lag behind ANN-based models in reconstruction fidelity. This paper proposes MADSNet, a Multi-scale Axial-attentive Dense Spiking Network for spike camera image reconstruction. MADSNet adopts a hierarchical spiking encoder--decoder, where a multi-scale spike stem extracts early features, axial spike attention captures long-range dependencies, a dense spiking bottleneck enhances feature reuse, and a spike-to-continuous decoder reconstructs continuous intensity images from spike-domain representations. Experiments on gray-scale and color spike reconstruction show that MADSNet outperforms SSIR by 4.3 dB PSNR and Spk2ImgNet by 1.94 dB on SREDS. Compared with Spk2ImgNet, it achieves about 15$\times$ lower latency and uses only about 1/40 of the estimated energy, demonstrating that carefully designed SNN architectures can improve reconstruction quality while preserving computational efficiency.

AI 总结：本研究提出四阶段贝叶斯管道，测试线性探针揭示的LLM算术几何的可解码性、形状、所有权、因果使用，发现多数形状继承自相关概念，仅用5维即可保留61%探针干预效果，建议报告四主张判定而非单一准确率。

链接：https://arxiv.org/abs/2611.12998

机构：Carnegie Mellon University（卡内基梅隆大学）； Indian Institute of Technology Jodhpur（焦特布尔印度理工学院）； Indian Institute of Technology Roorkee（鲁尔基印度理工学院）

作者：Anshul Kumar, Neha Prasad, Manoj Kumar, Barnabás Póczos, Deeksha Varshney

英文摘要： Linear probes are a standard method to study geometric structures, such as helices and circles, from the activations of large language models (LLMs), and such structure is often taken as evidence that the model computes with it. Probe accuracy alone cannot establish this, because it conflates four separate claims. These are (1) that the concept is decodable from the activations, (2) that they form a specific shape, (3) that the shape belongs to the probed concept rather than to a related concept the model also represents, and (4) that the output of the model depends on the shape. We call these claims decodability, shape, ownership, and causal use, and we test them separately with a four-stage Bayesian pipeline. Stage 2 compares candidate shapes and selects the best one only when its evidence passes a predefined threshold. Stage 3 removes a set of related concepts, specified in advance, and tests whether the shape is owned or inherited. Stage 4 intervenes on the shape to test whether the output of the model depends on it. We apply the pipeline to two-digit addition and multiplication in GPT-J 6B, Llama 3.1 8B, and Pythia 6.9B. Nearly every concept is decodable, but the decoded directions overlap substantially across concepts. Every strongly supported comparison selects a ribbon, which is a helix whose radius changes with the value, or a torus. Of 1,330 cells with an ownership verdict, 19% of shapes are owned and 81% are inherited. Owned and inherited shapes move the output about equally often. Intervening on the shape retains approximately 61% of the effect of intervening on the full probe while using at most five dimensions, against a median of nine for the probe. Our results indicate that a probe verdict alone overstates what a model computes with, and we propose that probe results be reported as verdicts on all four claims rather than as a single accuracy.

AI 总结：本研究将执行轨迹视为高维粗糙路径，用对数签名变换编码轨迹，在Karel等三个领域的程序归纳任务中，该编码器优于多种基线，还提出无外部神谕的自举流程，且诱导出更具可解释性的潜在几何结构。

链接：https://arxiv.org/abs/2611.13005

机构：CISPA Helmholtz Center for Information Security（CISPA亥姆霍兹信息安全中心）； Technical University of Munich（慕尼黑工业大学）

作者：Mohamed Ghanem, Bernd Finkbeiner

英文摘要：Synthesizing programs from execution traces is a fundamental challenge in algorithmic interpretability and the reverse engineering of complex systems. We bring a rough-path-theoretic perspective to program induction, treating execution traces as high-dimensional rough paths and encoding them with the path signature transform, a tool from stochastic analysis whose expectation provably characterizes the distribution of traces a program induces. On three program-induction domains, Karel, List, and RASP, our log-signature trace encoder outperforms LSTM, Transformer, and Fourier baselines in almost every setting, with its largest margins when training data is scarce. We further observe that the trace-to-program pipeline is naturally contractive: it often produces valid programs that are shorter than the ground truth. We turn this into a bootstrapping procedure that exposes redundancy in the ground-truth supervision without any external oracle, which we demonstrate on redundancy-injected versions of our datasets. Finally, the log-signature encoder induces a markedly more interpretable latent geometry than the baselines, in which the nearest neighbors of every state lie on its own execution trace.

AI 总结：提出连续时间集变换器，将异步WiFi-IMU观测表示为带时间戳标记的无序集合，通过单注意力块跨模态时间融合，在跨会话漂移等场景下提升室内定位性能。

链接：https://arxiv.org/abs/2611.13057

作者：Mohamed Bachar, Ilyass Abouelaziz (LINEACT), Yuehua Ding (LINEACT)

英文摘要：Indoor localization commonly combines WiFi received signal strength indicator (RSSI) fingerprints with inertial measurements. However, these sensor streams are sampled asynchronously at different rates; therefore, observations rarely align in time and may be missing or stale. Most existing methods resample the data onto a shared clock and process each modality using a separate branch. We instead present a continuous-time transformer that represents observations as an unordered set of timestamped tokens. Each token encodes its measurement, modality, and elapsed time since acquisition. A single attention block jointly fuses information across modalities and time, avoiding both resampling and modality-specific branches while remaining invariant to token order. During training, we randomly mask entire modalities and individual observations, enabling the model to operate under missing or outdated inputs. We validate the modality-specific encoders on their respective benchmarks and evaluate the fused model on both simulated and real datasets, including a crosssession dataset collected on unseen days. The proposed approach improves upon an open WiFi fingerprinting baseline under cross-session drift, remains competitive with learned fusion and dead-reckoning baselines, and degrades gracefully as observations become missing or stale. These results suggest that a single continuoustime transformer provides a practical framework for indoor localization from asynchronous sensors.

AI 总结：本研究提出公开基准INTENT，通过用户、时间段、目标物品是否在训练集中定义七个单元，评估模型在样本外、时间外及冷启动场景的表现，经59个数据集审核后保留五个，各单元含至少111,769用户，明确不同模型在各类场景的性能差异。

链接：https://arxiv.org/abs/2611.13069

机构：KTH AI Society（KTH人工智能协会）； Unbox AI（Unbox人工智能公司）； MIT（麻省理工学院）

作者：Mattias Kvist, Erik Lidman Hillbom, Edoardo de Cal, Davis Siemens, Adam Fredriksson, Rickard Brüel-Gabrielsson

英文摘要：Banks now train foundation models on ordered sequences of customer events such as transactions, app sessions, and authorizations. Deployed, such a model predicts forward in time, often for customers and merchants absent from training, yet each is evaluated on private data, so no two can be compared. Public benchmarks for such sequences rarely test a period the model has not seen; recommendation studies usually hold out each user's last events instead. We present INTENT (INTEraction, Novelty, and Transfer evaluation for next-action prediction), a public benchmark that scores a model in seven cells defined by whether the user, the period, and the target item appeared in training. An audit of 59 public dataset families retains five datasets; the most common defect is the lack of a calendar timestamp that orders events. No date lies within the time span of all five datasets, so each is split at its own cutoff, chosen by one rule fixed before any model is evaluated. Summed over the datasets, every cell holds at least 111,769 users. In reference results the cell determines the best model: a transformer leads on seen items after the cutoff but scores zero on new items, a variant that embeds item metadata helps mainly on MicroLens, and repeating the user's last item leads on new items in four of five datasets. This parameter-free control also shows that on three datasets most of the score lost after the cutoff reflects changed behavior rather than the model. The pipeline, the cutoff rule, and the per-cell user counts are publicly available.

AI 总结：该研究在pass@K设置中分析多模型多数据集，推导多温度分配方法，发现其增益有限，明确了多温度采样的适用场景与温度选择方式。

链接：https://arxiv.org/abs/2611.13077

作者：Aviv Orly (1), Ori Shem-Ur (1), Yaron Oz (1) ((1) School of Physics and Astronomy, Tel Aviv University, Israel)

英文摘要：Inference-time sampling is a central mechanism for improving large language models, with temperature controlling the tradeoff between reliability and diversity. While recent work suggests that varying temperature across samples can improve multi-sample inference, simple fixed-temperature schemes remain the standard in practice. We investigate this discrepancy in the canonical pass@K setting, asking when temperature diversity is useful, and when a single tuned temperature suffices. To this end, we analyze per-question success probabilities across multiple models and datasets, tracking how they vary with sampling temperature. We identify several consistent properties of these profiles, including that harder questions are less likely to share efficient sampling temperatures, although this trend is only logarithmic in hardness. We then derive an efficient method for evaluating the theoretically optimal multi-temperature allocation scheme, and show that although such a scheme improves performance, its gains over single-temperature baselines are typically modest when temperature is tuned as a function of the sampling budget K. We explain this phenomenon through mathematical mechanisms that we call dulling effects, and show how they ensure the robustness of temperature choices across different question distributions. Our results clarify when multi-temperature sampling is useful, and how temperatures should ideally be chosen.

AI 总结：本研究提出无参考指标RadRef，基于图像衍生信念与临床指征评估放射科报告，在ReXVal及MIMIC-CXR数据集上表现优异，可用于无参考时的报告质量评估与监测。

链接：https://arxiv.org/abs/2611.13168

机构：University of Virginia（弗吉尼亚大学）

作者：Shelly Schwartz, Alan Ropp, Michael B. Sneider, Nur Yildirim, Cree M. Gaskin, Chirag Agarwal

英文摘要：AI-generated radiology reports are usually evaluated by comparison with a radiologist-authored reference, leaving no way to score them when no reference exists. We introduce RadRef, a reference-free metric that scores report claims against image-derived beliefs over a clinical finding vocabulary, weighted by relevance to the clinical indication. On ReXVal, RadRef achieves 73.1% agreement on non-tied same-radiograph report pairs with radiologist rankings, comparable to leading reference-based metrics overall. It is also competitive on false findings and has its clearest relative result on omissions (missed findings). We further evaluate 100 model-generated and 100 radiologist-authored reports from 100 MIMIC-CXR studies, all independently annotated by two board-certified radiologists. On omissions, RadRef attains Kendall's tau-b = 0.295, against 0.386 between the two readers themselves, and no compared metric differs significantly from RadRef on this target. The same image-grounded representation provides a false-finding readout, RadRef-FF, with tau-b = 0.32 on ReXVal and 0.29 in the reader study, without requiring a reference. Controlled perturbations confirm that RadRef depends on the radiograph, report, and clinical indication. RadRef enables image-grounded evaluation before a radiologist-authored reference exists, supporting quality monitoring of report-generation systems in clinical settings, and complements reference-based metrics when one is available.

AI 总结：本文提出PP-Rank框架，结合LLM评判器预测与自适应收集的人类反馈，在自适应采样下提升LLM排序精度，实验显示其可降低评估成本且性能优于仅人类基线。

链接：https://arxiv.org/abs/2611.13186

机构：University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）； Purdue University（普渡大学）； University of Michigan（密歇根大学）

作者：Gabriel Sargent, Will Wei Sun, Zhengwu Zhang, Yufeng Liu

英文摘要：Ranking large language models (LLMs) through human evaluations is essential for downstream decision-making. Because human feedback is expensive, evaluation platforms often collect it adaptively, prioritizing informative model comparisons using previously observed feedback. To further improve precision with limited human feedback, evaluators can draw on abundant unrated responses and relatively inexpensive LLM judges; however, judge bias and adaptive dependence challenge standard i.i.d. inference frameworks. We propose Prediction-Powered Ranking (PP-Rank), a framework that combines LLM-judge predictions with human feedback to improve precision while providing valid inference under adaptive sampling. Data-driven prediction weights minimize the asymptotic covariance trace within the proposed class. In theory, we establish asymptotic normality for general evaluation targets and simultaneous ranking inference. A joint martingale analysis handles dependence between adaptive human feedback and auxiliary data. We further derive a rule for reducing exploration that prioritizes lower-cost comparisons while matching a conservative variance bound, and characterize allocation limits for regular $\sqrt{T}$-rate estimation of pairwise means. Semi-synthetic experiments on UltraFeedback-Extended show improvements over human-only baselines in score estimation, confidence interval width, and ranking certification. Synthetic experiments demonstrate robustness to judge quality and adaptive allocation, while showing that PP-Rank can substantially reduce evaluation costs without sacrificing precision.

AI 总结：该研究针对跨样本多模态归因问题，提出结构化核代理模型SKS，在多模态基准上较现有方法提升了预测相关性与下游数据选择准确率，可识别卫星图像中的事故相关道路路段。

链接：https://arxiv.org/abs/2611.13198

机构：Northeastern University（东北大学）； Emory University（埃默里大学）

作者：Ziniu Zhang, Zhenshuo Zhang, Jianglin Lu, Ruoxuan Xiong, Yun Fu, Hongyang R. Zhang

英文摘要：We study the problem of cross-sample modality attribution, which involves measuring how one modality in one sample affects the learning outcome in other multimodal samples. This problem is crucial in a multimodal system because a data modality (such as a text description) may be redundant for one sample (a self-explanatory image), but the same data modality would otherwise be uniquely informative for another sample, or informative when combined with other modalities. Prior work focuses on influence estimation at the level of one sample, and it remains an open question how one should capture the interactions across both samples and modalities. To capture scenarios like the above, we propose structured kernel surrogate models, which are surrogate mappings from both samples and modalities to the prediction outcome when they are trained together. We design an algorithm, SKS, which applies kernel ridge regression with a structured, compositional kernel on the selection states of individual samples and modalities. Specifically, the kernel function multiplies kernel-based similarity scores between every pair of data modalities for one sample and a pair of two samples. Additionally, we design a gradient-based estimation algorithm to efficiently estimate the structured kernel surrogate model. We evaluate our approach on two multimodal benchmarks, spanning a range of data modalities including satellite images, road networks, text, audio, and tabular data. We find that SKS achieves $57.8$% higher correlation with the true prediction outcomes than existing methods. In downstream data selection, SKS improves the average accuracy by $1.93$% over baselines. Models with data selected by SKS have lower Hessian traces, indicating a flatter loss surface. Using SKS to select satellite images identifies road segments, where accident scenes commonly occur, as opposed to background images.

AI 总结：本研究提出GigaEdge Spench开源SDK，支持跨平台流式语音增强，含多类模型与核心，经实验验证其处理速度优于实时，可配置且高效，为相关应用提供实用工具。

链接：https://arxiv.org/abs/2611.13243

机构：SberDevices； Sberbank（俄罗斯联邦储蓄银行）

作者：Artem Sokolov, Artem Khaidarov, Yana Lysova, Alexander Gladkikh, Vitaliy Yulis

英文摘要：We present GigaEdge Spench, an open-source SDK for streaming speech enhancement on CPU servers, mobile devices, and in browsers. Applications can build enhancement channels from neural denoising, normalization and dynamics blocks, sample-rate conversion, and acoustic echo cancellation where supported. These stages can be combined, while normalization can also run without a neural model. The release includes an 8 kHz INT8 denoiser and discriminative and hybrid 48 kHz FP32 models. A shared C++ core supports ONNX Runtime and TensorFlow Lite through native interfaces, Android and iOS SDKs, and WebAssembly integration. Independent stream contexts share model resources, and consecutive frames from one stream can be processed in a single inference call. We evaluate the released models on EARS-WHAM v2 and compare their quality and processing costs with RNNoise and DPDFNet2. All tested native and Web configurations process files faster than real time, although larger chunks require more time to collect input. We also measure how many streams a shared model can process on a CPU. Our contribution is a configurable deployment system and an experimental study of how model choice, chunk size, and execution environment affect its use. All reported SDK timing uses ONNX Runtime. Index Terms: speech enhancement, streaming inference, cross-platform deployment, temporal frame grouping, computational efficiency

AI 总结：提出DRD-MOPD方法，通过域路由更新参数分量，在多教师在线策略蒸馏中保留专家几何结构，在六个基准测试中取得32.72%宏观分数，优于初始任务算术合并。

链接：https://arxiv.org/abs/2611.13249

机构：School of Computer Science, Shanghai Jiao Tong University（上海交通大学计算机科学学院）； Zhongguancun Academy（中关村学院）

作者：Lin Yao

英文摘要：Multi-teacher on-policy distillation assigns each sample a domain teacher but typically updates shared student parameters. We introduce Domain-Routed Delta Optimization (DRD-MOPD), which specifies which parameter components each domain's samples update. The student is the sum of a frozen base and trainable components initialized from expert-minus-base weights. All components participate in every forward pass; four update modes define the correspondence between sample domains and components, each with independent optimizer states. We analyze these update rules and measure retention of the original expert geometry through angles, projection coefficients, and energy. Prior work such as Open-MOPD distills directly from the base model; 97.02% of its released final checkpoint's increment energy lies outside the span of the original expert directions. In comparison, our Own checkpoint after 20 updates retains coefficients near one along each component's original direction and pairwise angles ranging from 88.04 to 92.41 degrees; its exported BF16 student's increment has 4.90% outside energy, versus 10.52% for the sum of saved FP32 components. The same checkpoint achieves a 32.72% macro score across six benchmarks, compared with 32.44% for its initial task-arithmetic merge. These measurements describe geometry and capability at the evaluated training stages and settings.

AI 总结：本文提出NN-CLEAN算法，用多头分类器替代多维搜索实现高效多径参数估计，在部分场景下加速与内存优化显著，可降低信道估计误差但精度略逊于GS-CLEAN。

链接：https://arxiv.org/abs/2611.13298

机构：Georgia Institute of Technology（佐治亚理工学院）； Columbia University（哥伦比亚大学）

作者：Chaofan Deng, Linyu Sun, Jaeho Lee, Zishen Wan, Arijit Raychowdhury

英文摘要：High-resolution multipath estimation in wideband multiple-input multiple-output (MIMO) channels requires repeated searches over angle of arrival, angle of departure, and propagation distance. This paper presents neural network-assisted CLEAN (NN-CLEAN), which replaces the multidimensional search at each extraction step with a multi-head classifier while retaining least-squares (LS) estimation of the complex path gain, parametric reconstruction, and residual cancellation. The same classifier is applied to successive residuals, allowing the extraction depth to vary without changing its output structure. With the number of iterations prescribed by the reference multipath-component (MPC) count, off-grid synthetic tests show that NN-CLEAN slightly exceeds grid-search CLEAN (GS-CLEAN) in path accuracy for one and two MPCs, although its accuracy declines more rapidly at larger path counts. For complete three-MPC extraction at a batch size of eight, NN-CLEAN achieves an approximately $27\times$ speedup and a $52$-fold reduction in peak GPU memory relative to GS-CLEAN. Applied without retraining to two-MPC line-of-sight (LOS) and non-line-of-sight (NLOS) Sionna ray-tracing channels within a coded MIMO link, NN-CLEAN substantially reduces channel-reconstruction mean-squared error (MSE) and bit error rate (BER) relative to interpolated LS channel estimation. GS-CLEAN retains an accuracy advantage on these channels, indicating the estimation cost associated with the computational savings.

AI 总结：本文提出名为MicroLens的交互式查看器，结合logit透镜与J-lens，在Gemma-3-4B-it模型上验证其可用于分析残差流信息，发现模型在不同问题类型中读取与使用信息的规律。

链接：https://arxiv.org/abs/2611.13322

机构：Stanford University（斯坦福大学）

作者：Nika Zahedi, Fangyi Zhu, Anna Wu

英文摘要：Lenses that decode hidden states into words are a popular way to look inside language models, but a heatmap of decoded words only shows what a vector reads as and cannot distinguish whether, or how much, the model uses the information. We present an interactive viewer that puts reading, tracing and editing on the same layers-by-tokens grid of the residual stream. In the viewer, every cell can be decoded with two lenses, the logit lens and the Jacobian lens (J-lens), traced back to the tokens it drew from, and edited, with the rest of the model rerun on the spot. We demonstrate three usage scenarios on Gemma-3-4B-it, backed by sweeps over a few hundred prompts, to show what this tool can discover. First, the two lenses give different directions to write with, each suiting a different edit: the logit lens's word direction sets what a token outputs, while the J-lens's direction changes what later tokens read. Second, in two-hop trivia questions (the capital of the state or country containing a given city), the state or country is readable at the city token from as early as layer 7, and swapping it there can redirect the answer 92-98% of the time; however, at the answer position, the state or country information is not used. Third, when a chat model answers questions with a step-by-step reasoning prompt, the correct answer is often already the top word in the residual stream before any reasoning is written: for 89-92% of retrieval questions and 39-63% of multi-hop and serial ones, far above their respective distractors (at most 5%). However, ordering puzzles are an exception, as the model works them out only once it writes its final reasoning steps. Furthermore, for answers that are visible in the residual stream, the J-lens reads them a layer or two earlier than the logit lens across the different types of question.

AI 总结：本研究提出基于卷积LSTM的质量守恒框架PECAN，在雪主导流域验证其径流模拟精度，同时诊断出其错误表征积雪过程的结构局限，为物理感知水文神经网络开发提供指导。

链接：https://arxiv.org/abs/2611.13346

机构：School of Civil Engineering and Environmental Science, University of Oklahoma（俄克拉荷马大学土木与环境工程学院）； Lawrence Berkeley National Laboratory（劳伦斯伯克利国家实验室）； International Computer Science Institute（国际计算机科学研究所）； School for Environment and Sustainability, University of Michigan（密歇根大学环境与可持续发展学院）

作者：Yihan Wang, Lujun Zhang, N. Benjamin Erichson, Yang Hong, Tiantian Yang

英文摘要：Recent progress in physics-aware deep learning (DL) has embedded mass conservation within Long Short-Term Memory (LSTM) architecture for hydrologic modeling, improving physical consistency and enabling diagnostics of internal states. However, existing mass-conserving neural networks are primarily developed in lumped formulations. Consequently, previous studies provide only fragmented and often indirect evidence of when and why imposed physical constraints may lead to unrealistic hydrologic behavior. To address this limitation, we propose Physics-Enhanced Conservation-Augmented Net (PECAN), a mass-conserving convolutional LSTM-based framework for distributed hydrologic modeling and diagnosing structural limitations of physics-aware DL models. PECAN is implemented in two variants: (1) PECAN-MC with a strict precipitation-based conservation constraint and (2) PECAN-MCR with a relaxed conservation constraint allowing additional water inputs beyond precipitation. Both variants are trained solely on gauge streamflow and evaluated in a snow-dominated watershed in Colorado. We benchmark PECAN against lumped DL baselines and the distributed process-based Variable Infiltration Capacity (VIC) model. Results show that PECAN achieves competitive outlet streamflow simulation accuracy while producing spatially explicit latent states for interpretation. Comparisons with reference soil moisture indicate that PECAN cell states resemble meaningful temporal soil moisture dynamics. However, analyses also reveal that PECAN misrepresents snow processes by prematurely removing winter snow water storage from the system rather than retaining it for spring release. PECAN demonstrates accurate streamflow simulation at gauges and shows diagnostic potential to expose structural limitations in hydrologic process representation, thereby informing future physics-aware neural network development for hydrology.

AI 总结：本文提出一种冻结专家组合的评估协议，通过对21个深度伪造检测器的实验发现，特征组合在多数场景下优于预测组合，且需控制训练/测试重叠以避免结论偏差。

链接：https://arxiv.org/abs/2611.13356

作者：Alessandro Pereira, Keith J Ransom, Lewis Mitchell

英文摘要：Practitioners often combine independently trained models while keeping their weights frozen. These experts are commonly combined through their final predictions, although their internal features may retain useful information that those predictions omit. We ask when combining features improves on combining predictions, and how to make that comparison fair. We introduce a reusable evaluation protocol, apply it to 21 frozen deepfake detectors from DF40 (Yan et al., 2024), and audit how evaluation choices change the conclusions. The protocol uses strong prediction baselines, compares simple and larger feature combiners, varies architecture and training data separately, and reports results by held-out generator with train/test overlap controls. Under leave-one-generator-out evaluation, logistic regression on the features of six architecture-diverse experts improves macro AUROC by 0.081 over the best tested prediction combiner in each fold, with gains on seven of eight generators. The larger network shows no clear improvement over the linear model (approximately 3.5M versus 10K parameters), suggesting that access to richer information matters more than additional combiner capacity on this pool. Features may help when they retain information that the predictions omit and that transfers to new generators. Predictions may be enough when the strongest prediction baseline already performs well, while neither approach reliably compensates for poor expert transfer. Weakening prediction baselines on the overlap-affected standard split inflates the apparent feature advantage by up to 0.172 AUROC. Overlap-aware evaluation also changes the estimated gain, though the contrast does not isolate overlap. Results by generator and expert pool show where the advantage holds, weakens, or reverses. Without these controls, the same benchmark can support different conclusions.

AI 总结：该研究提出循环因果编解码器（CED），通过编码器循环实现零轮次KV增长，在与非循环CED内存相当的情况下，以1/6全局KV恢复循环Transformer 86%-93%增益，在5520亿参数模型上实现36.8倍KV压缩与3.6倍吞吐量提升。

链接：https://arxiv.org/abs/2611.13378

机构：University of California, San Diego（加利福尼亚大学圣迭戈分校）； Tsinghua University（清华大学）； Nanjing University（南京大学）

作者：Jiaqi Leng, Minshen Zhang, Zhaofeng Sun, Yu-Yang Qian, Peng Zhao, Hao Zhang

英文摘要：Looped language models deepen computation without adding parameters, yet caching full-context key-value (KV) states across layers and rounds severely bounds serving batch size. We design a looped causal encoder-decoder (CED) that achieves recurrent depth with zero round-wise KV growth, addressing where and how to loop an asymmetric stack in which only sparse encoder layers generate global KV. We find that recurrence must be placed in the encoder to iteratively refine global KV representations rather than re-reading static decoder projections, which both drives language-modeling gains and preserves long-context scaling. Furthermore, training the encoder loop under stochastic recurrence depth induces a fixed-point equilibrium where every round losslessly shares the last round's KV cache. Compounding cross-layer and cross-round KV sharing, the resulting Looped CED matches an unlooped CED's minimal memory footprint and recovers 86-93% of a looped Transformer's gain with 1/6 of its global KV. To test serving at production scale, we loop the 552B-parameter DeepSeek-V4.1-Flash for 8 rounds on one 4xGB200 node: with per-layer and per-round KV the loop cannot fit a single 128K-token request, whereas Looped CED holds 36.8x less KV per token, fits 38 such requests (as many as the unlooped model), and serves 3.6x higher throughput on 8K-token prompts.

AI 总结：本文提出一种零成本的序列级Rényi熵，基于序列蒙特卡洛解码估计，在多数场景中优于现有零成本信号，可优化自适应LLM推理的粒子使用效率。

链接：https://arxiv.org/abs/2611.13414

机构：University of Southern California（南加州大学）； Purdue University（普渡大学）； Intel（英特尔）

作者：Erfan Baghaei Potraghloo, Milad Heidari Sorkhehdizaj, Shahab Rahimirad, Aliasghar Pourghani, Seyedarmin Azizi, Souvik Kundu, Abolfazl Hashemi, Joseph Campbell, Vijay Gupta

英文摘要：Quantifying uncertainty over an entire LLM completion is critical for selective prediction and adaptive compute. Zero-cost signals such as mean token entropy and log-probability are local to tokens or to one completion, and self-consistency needs many extra completions. We show that Sequential Monte Carlo (SMC) decoding for power sampling already estimates the normalizing constant of the power distribution, which gives the sequence-level R'enyi entropy (SRE) of the model's completion distribution at zero marginal cost, with a guarantee at every particle budget. A R'enyi chain rule extends this readout to a family of stepwise entropies of any order. We prove that orders above one are tied to the most likely token of each step, which explains why they track log-probability, while the order-1/2 member accumulates the log of per-step search effort. This member is the best zero-cost selective-prediction signal in 10 of 13 settings. On MATH500 its AUROC is 0.022 to 0.044 above log-probability with all four models and within length quartiles, and the same order is selected when each setting is held out. An 8-particle probe identifies the problems that benefit from more particles, and a two-pass policy matches 64-particle accuracy on MATH500 with 34.2% fewer particles over three seeds.

AI 总结：该研究通过旋转不变核机分析ICL任务，发现ICL的泛化本质是对记忆解的欠拟合，证明特征学习对ICL非必要，且Transformer的ICL记忆机制符合核理论预测。

链接：https://arxiv.org/abs/2611.13431

机构：UC Berkeley（加州大学伯克利分校）； Imbue

作者：Joseph Turnbull, Jamie Simon

英文摘要： How should we understand memorization vs. generalization in in-context learning (ICL) tasks, and when does one happen vs. the other? To investigate, we study how rotation-invariant kernel machines learn the ICL linear regression task proposed by Raventos et al. (2023). By decomposing the task in the kernel eigenbasis, we obtain a spectral picture of learning in this ICL setting. This picture reveals that as sample size grows, kernel machines learn successively-higher-degree components of the memorizing solution, which depend upon successively-higher-degree statistics of the training task set. The generalizing solution is approximately a low-degree truncation of the memorizing solution, and so the model first generalizes, then degrades into memorization as training proceeds. The takeaways are threefold. First, if this case is indicative, learning a "generalizing" solution on an ICL task is really just underfitting the memorizing solution! Second, the fact that kernel machines can generalize on this task means that feature learning is inessential to ICL. Third, transformers empirically trained to do ICL are shown to memorize tasks through the higher-order statistics of the training tasks that the kernel-based theory predicts, suggesting it applies even to models capable of feature learning.

AI 总结：本文提出PC-Fermat，一种基于费马距离图的覆盖式主动学习方法，在多数分类数据集的不同标注预算下，尤其在极低预算场景中，性能优于现有方法。

链接：https://arxiv.org/abs/2611.13448

机构：University of California, Los Angeles（加州大学洛杉矶分校）； University of Pittsburgh（匹兹堡大学）； Brigham Young University（杨百翰大学）； Michigan State University（密歇根州立大学）

作者：Gokul Bhusal, Delanna Do, Kevin Miller, Ekaterina Merkurjev, Stephan Wojtowytsch

英文摘要：Active learning aims to reduce annotation cost by selecting the most informative examples for labeling. In the low-budget regime, where uncertainty-based methods often fail, coverage-based active learning methods such as ProbCover have been shown to be effective. However, these methods typically rely on Euclidean distance, which may not adequately capture the underlying geometry of data. In this article, we propose PC-Fermat, a coverage-based active learning method that replaces Euclidean coverage neighborhoods with neighborhoods defined by a Fermat distance, i.e., a data-adaptive distance that reflects the underlying distribution of the data points. To facilitate computation, we construct a weighted Euclidean neighborhood graph and use a truncated shortest-path algorithm to compute the Fermat coverage graph. We also establish sufficient conditions under which this construction exactly preserves the Fermat coverage computed on the complete graph. Finally, we evaluate PC-Fermat on eight classification datasets across varying labeling budgets, with particular emphasis on the very low-budget regime. Our experimental results demonstrate that PC-Fermat outperforms several existing active learning methods on the majority of the datasets.

AI 总结：该研究提出深度个性化方法，在本地主动式智能体Coco中通过观测智能体实现设备端闭环适应，在OSWorld-Persona基准和用户研究中均显著提升干预决策的宏F1值，增强主动式智能体的实用性。

链接：https://arxiv.org/abs/2611.13457

机构：NVIDIA（英伟达）； Stanford University（斯坦福大学）； Massachusetts Institute of Technology（麻省理工学院）； OpenAI（开放人工智能公司）

作者：Yijia Shao, Song Bian, Junxian Guo, Jindong Jiang, Zhen Dong, Huizi Mao, Diyi Yang, Song Han

英文摘要：As AI agents become increasingly capable, users must still recognize when and how to ask for help, leaving many opportunities for useful support unrealized. Proactive agents hold the promise to close this gap, but effective proactivity requires understanding users' real-time activities and personalizing intervention timing so that assistance is offered when it is likely to be helpful rather than disruptive. We introduce deep personalization, a closed-loop process that connects continuous context observation with continual agent adaptation: an agent observes user activity, selectively retains long-term context, and adapts intervention decisions from implicit feedback. We initiate deep personalization in Coco, a proactive desktop agent, by having a separate Observer Agent that monitors desktop activity, maintains compact long-term user modeling context, and learns when assistance is desirable from implicit signals such as engagement, dismissal, explicit requests, and repeated work patterns. We further develop efficient on-device strategies to enable the Observer Agent to run on local AI compute platforms without streaming raw activity data off-device. For evaluation at scale without collecting sensitive real-user activity, we introduce a data-synthesis pipeline that transforms existing computer-use trajectories into OSWorld-Persona, a benchmark with persona-conditioned intervention labels. On OSWorld-Persona, deep personalization improves macro F1 across three model backbones from 0.12-0.19 under direct prompting to 0.35-0.48 after adaptation, a 2.4-3.0x relative gain. In a week-long user study, we observe that Coco improves its macro-averaged F1 score from 0.199 to 0.515. These results suggest that on-device deep personalization can make proactive agents more timely and increasingly aligned with each user's everyday work.

AI 总结：该研究针对递归自我改进（RSI）的成本问题，提出RSI税概念，通过实验发现带边界基线的准确率优于部分RSI方法，验证成本占RSI税的主要部分，为优化RSI目标提供了依据。

链接：https://arxiv.org/abs/2611.13524

机构：School of Physics, Peking University（北京大学物理学院）

作者：Yitong Zhang, Ang Li, Yisen Wang

英文摘要： Recursive self-improvement (RSI) seeks to improve an agent using its own task experience. Its practical value depends on whether performance gains justify the full cost of self-improvement and task execution. We evaluate this trade-off on continual task streams, scoring each task before its feedback is used for subsequent updates. We define the \emph{RSI Tax} as the token cost of the self-improvement process and include it alongside task-execution costs in end-to-end evaluation. Across six methods and two language models, a bounded-note baseline matches or exceeds the observed accuracy of Meta-Harness, EvoSkill, and RRSI, with mean total-token savings of 65.6-71.0\%. We decompose the RSI Tax into verification, exploration, and writing costs, and track subsequent use of the associated state or artifact. For these three RSI methods, verification accounts for 74.8--93.6% of the aggregate RSI Tax. A substantial share of the RSI Tax is also associated with states or artifacts that are not used later in the evaluated streams. These results expose a practical limitation of costly RSI procedures: improvements in the deployed agent need not justify the expense of the improvement process. They motivate RSI objectives that jointly optimize the RSI Tax and expected task-execution cost while ensuring continued growth in agent capabilities across the task stream.

AI 总结：本文针对有限预算下神经缩放律偏离对数线性前沿的问题，通过Mellin–Barnes分析、Aitken差分等方法，提出估计器并在多数据集上验证，降低了分配遗憾与混合误差。

链接：https://arxiv.org/abs/2611.13534

机构：LingQiong AI Lab（凌琼人工智能实验室）； Shanghai Smart Logic Technology（上海智逻科技）

作者：Tianyu Zhao, Zhixiang Ren

英文摘要：Classical neural scaling laws model excess risk as a separable power law in model capacity $N$ and dataset size $D$. At finite budgets, however, empirical learning curves and compute-optimal allocations depart from this log-linear frontier. This paper analyzes how spectral and domain heterogeneity govern these finite-scale departures. Along the capacity axis, Mellin--Barnes analysis of covariance resolvents yields a three-term expansion where competing spectral and regularization modes produce finite-scale curvature and logarithmic transitions. Because fitting multiple power laws on narrow pilot grids is ill-conditioned, we eliminate the loss floor via first-order Aitken differencing to obtain a well-conditioned closed-form initializer. Along the data axis, KKT duality reduces $K$-domain mixing to scalar water-filling, yielding finite-budget bottleneck bounds, an online sampling curriculum, and a simplex Hessian metric for cross-domain transfer. Along the iso-compute frontier $C=6ND$, we prove that the optimal allocation exponent drifts in proportion to capacity and data exponent variances, while Schur-complement reduction separates intrinsic drift from frozen-hyperparameter bias. Across 45 operator spectra, Pythia (70M--12B), Chinchilla (245 runs), and RegMix (17 domains), our estimators achieve $5.89\%$ holdout MAPE, cut high-compute allocation regret by up to $76.4\%$, and reduce cross-scale mixture RMSE by $59.4\%$ with zero selection regret.

AI 总结：本研究提出可组合催化智能（CatAI）框架，学习180万种结构-性质关系，在基准任务中预测误差比次优模型低7.8%，实验中发现新型催化剂并优化光催化剂性能，实现跨系统知识重用。

链接：https://arxiv.org/abs/2611.13536

作者：Wentao Li, Qilong Cai, Qi Lei, Jinxing Chen, Ruixuan Chen, Jiangjie Qiu, Botian Wang, Kai Zhao, Honghao Chen, Leyi Zhao, Zemeng Wang, Yijun Li, Jun Yin, Fangfang Li, Hongxuan Liu, Yingdong Lu, Yawen Ouyang, Manu Suvarna, Wei-Ying Ma, Ya-Qin Zhang, Kostya S. Novoselov, Hao Zhou, Jiong Lu, Xiaonan Wang

英文摘要：Catalysis underpins chemical manufacturing, yet knowledge for catalyst design remains dispersed across experimental observations, computational data and catalysis subdisciplines. To overcome this challenge, we present Composable Catalytic Intelligence (CatAI), a physics-grounded modular AI framework that learns 1.8 million structure-property relations and organizes them into 30 physicochemical encoders. This reusable intermediate layer enables task-specific models to guide catalyst discovery for targeted reactions and objectives. Across benchmark tasks, CatAI achieves a 7.8% lower task-averaged prediction error than the next-best models. In experimental validation, CatAI identified a new confined-passable pocket catalyst applicable in industrial homogeneous polymer synthesis with 9 times the activity of the commercial benchmark. In parallel, it guided autonomous closed-loop optimization of a highly active Cu-based geminal-atom photocatalyst for C-O coupling reactions surpassing the traditional Ni-based systems. This approach connects property-level data to catalytic objectives, supporting physicochemical knowledge reuse across mechanistically distinct systems through computational screening and iterative model-experiment feedback.

AI 总结：该研究针对Transformer架构修改的合理性问题，提出三项可复用检验，通过9个预训练模型的实验验证自注意力移除的效果，发现训练增益依赖于注意力头自身值方向而非任意方向。

链接：https://arxiv.org/abs/2611.13543

作者：Basavaraj R M, Aryan Bharat Kumar

英文摘要： Transformer architectural modifications frequently measure a statistic on internal representations, observe that it is large, remove it architecturally, and report an improvement. Interpreting that sequence requires evaluating whether the statistic exceeds its noise floor, whether it survives a geometric null, and whether the training gain requires the specific direction removed. We formulate these requirements as three reusable checks and apply them to self-attention removal, which deletes the component of an attention head's output aligned with its own value vector. Across nine multi-head pretrained models (5,408 heads, 124M to 6.9B parameters), an earlier-position causal null accounts for 48% to 63% of the head-value cosine alignment, with no monotone scale trend. Under grouped-query attention, this alignment is specific to each head's own key-value group and reverses across groups (+0.24 to +0.27 within-group, -0.11 to -0.19 across-group). Split-half reliability demonstrates that the per-head intervention effect is resolvable in GPT-2 ($r_\Delta = 0.80$) but attenuated in Pythia-160M (0.42) and Pythia-410M (0.53), whereas structural self-cosine reliability is near-perfect in all three ($r_{\text{stat}} > 0.99$); which statistic predicts its own intervention is model-dependent. In a pre-registered 3-arm by 8-seed paired factorial at $4.0 \times 10^8$ tokens per run (399,900,672 tokens), a fixed arbitrary rank-one direction does not differ from baseline (+0.0011 nats, $p = 0.197$), while self-directed removal lowers validation loss (-0.0029 nats, Holm-corrected $p = 0.0023$) and outperforms the fixed arbitrary control directly (-0.0040 nats, $p = 0.0018$). At 50.9M scale the training benefit requires the own-value direction rather than a fixed arbitrary one, but a large cosine alignment alone does not establish redundancy in frozen models.

AI 总结：该研究发现大量激活值会混淆稀疏自编码器的评估，提出去除第一主成分后训练SAE再添加该成分的修正方法，提升了特征共享率并降低了模型性能下降幅度与重建误差。

链接：https://arxiv.org/abs/2611.13548

机构：University of Chicago（芝加哥大学）

作者：Nathan Delisle, Chenhao Tan

英文摘要：Sparse autoencoders (SAEs) are commonly evaluated by variance explained and ablation damage. We show how massive activations undermine both evaluations, and propose a correction. Massive activations cause up to 99.9% of activation variance to lie along the first principal component (PC1). Because the PC1 component has no unique sparse decomposition, each SAE run rebuilds it from a different set of features. These features dominate importance rankings: an average of 47 of the 50 most important features are unmatched in a separate run, compared with about 27 of a random 50. A dictionary can also report 99.96% of variance explained while leaving 33% of the variance outside PC1 unexplained. We bypass PC1 by training the SAE on activations with PC1 removed, then adding PC1 to the reconstruction. In affected models, the top-50 feature sharing rate rises from 4-39.5% to 73-99.5%. At the same TopK sparsity, the resulting SAE degrades the model up to 44% less and reduces content reconstruction error by up to 33.9%.

AI 总结：本研究将表格基础模型（TFMs）引入会计研究，通过与大语言模型、树基模型对比，发现Google Research的TabFM在企业下一年度收益变化预测中准确性更优、速度更快，为会计金融预测提供了新工具。

链接：https://arxiv.org/abs/2611.13550

作者：Ed deHaan, Allen H. Huang

英文摘要：This technical note introduces tabular foundation models (TFMs) to accounting research and evaluates their performance in one common benchmark task: predicting firms' next-year earnings changes. TFMs produce numeric predictions and learn in context, without post-release training or tuning, making them natural and easy to use for forecasting. The TFMs we study are pretrained on synthetic tabular data, so they avoid look-ahead bias from memorizing firms' outcomes. We compare Google Research's TabFM with a fine-tuned open-weight large language model developed in recent accounting research and with several tree-based models. TabFM is more accurate than all comparison models. It is also faster and less computationally intensive than the large language model. Several tree-based specifications produce predictions more quickly than TabFM, so choosing between them involves a trade-off between accuracy and speed. We conclude that TFMs have promise as a new tool for forecasting in accounting and finance. However, our evaluation is based on a single benchmark task, and no class of models is likely to dominate in all settings.

AI 总结：研究人员推出LLM引导的DisCo框架，通过含验证器和清理器的进化搜索，在三个行为任务中发现结构合理且可解释的认知模型，性能与任务特定基线相当，验证器对恢复心理学意义结构至关重要。

链接：https://arxiv.org/abs/2611.13561

机构：Graduate School of Data Science, KAIST（韩国科学技术院数据科学研究生院）； School of Digital Humanities and Computational Social Sciences, KAIST（韩国科学技术院数字人文与计算社会科学学院）； Department of Brain & Cognitive Sciences, KAIST（韩国科学技术院脑与认知科学系）

作者：Boyoung Kwon, Jiyoon Seo, Sang Ho Lee

英文摘要：Computational cognitive modeling has traditionally relied on manual model construction, which is a time-intensive process that requires substantial domain expertise. Although recent LLM-based approaches have begun to automate model generation, they often prioritize predictive accuracy at the expense of structural plausibility and parameter identifiability. We introduce DisCo, an LLM-guided framework that discovers interpretable symbolic cognitive models through evolutionary search with two structural components: a Validator that filters structurally invalid models, and a Cleaner that promotes parameter identifiability by removing redundant parameters. Across three behavioral tasks, DisCo consistently discovers models that achieve performance competitive with task-specific baselines. Ablation analyses show that the Validator is important for recovering psychologically meaningful structure. Without this component, models tend to violate behavioral constraints and show less consistent associations with external cognitive measures. These results suggest that LLM-driven model discovery, when coupled with appropriate structural constraints, can yield cognitively plausible models that go beyond predictive fit.

AI 总结：该研究通过13种变形关系测试17种AC-OPF代理配置的表示依赖性，发现多数配置违反变形关系，指出GridSFM-Open v1.1的预处理问题，表明数据增强的益处依赖训练协议，建议报告表示一致性。

链接：https://arxiv.org/abs/2611.13566

机构：AI Science Institute, Boston Consulting Group（波士顿咨询集团人工智能科学研究所）

作者：Stephen Peng, Vineet Jagadeesan Nair

英文摘要：Evaluating a learned AC optimal power flow (AC-OPF) surrogate on one grid serialization cannot show whether its accuracy depends on that encoding. We test this dependence with thirteen metamorphic relations, each specifying when a transformation is valid and how to compare its outputs. Our audit covers 17 model and pipeline configurations: two released trained checkpoints and 15 random-weight architecture or mechanism controls. Every configuration violates at least six relations, even at reporting tolerances ten times tighter or looser. For GridSFM-Open v1.1, source inspection and intermediate-tensor interventions point to an indexing bug and three noncanonical or numerically unstable preprocessing choices; a repaired checkpoint remains untested. In a separate case500 training experiment, augmentation lowers transformed-view error in three of four models and median discrepancy and constraint-violation P95 in all four. These benefits depend on the training protocol: some relations worsen, and every evaluated prediction remains AC infeasible. A GPU/CPU eigensolver mismatch accounts for most of the large raw-Laplacian-baseline gain. We recommend reporting representation consistency alongside accuracy, feasibility, and cost. These results establish neither deployment readiness nor universal invariance.

AI 总结：SepsisFlow是用于ICU脓毒症连续风险预警与轨迹建模的事件流基础模型，在发病前预警、发病背景区分及病程风险分层上均优于现有方法，为重症监护提供了新范式。

链接：https://arxiv.org/abs/2611.13661

作者：Yawen Zhang, Wanzhe Xu, Zhenhua Zhang, Jinyu Wu, Yizhou Zhuang, Yitao Yang, Kenta Nakai

英文摘要：Sepsis is a leading cause of ICU death with heterogeneous onset and course, yet most warning models predict only whether sepsis will occur, leaving clinicians uncertain of its onset context and early trajectory once an alarm sounds. To answer these questions on one hourly clock from ICU admission, we built SepsisFlow: before onset, it combines a Mamba model pretrained on 2.57 million EHRs with about 1,700 hourly features in gradient-boosted trees to predict 6/12/24-h risk; at onset, it fuses Mamba and tabular encoders via a gated network into continuous patient states defining onset contexts; after onset, it encodes 72-h organ-support and physiological trajectories via an attention bidirectional recurrent network into early courses. For pre-onset warning in MIMIC-IV, SepsisFlow detected more cases than NEWS2, qSOFA, and SIRS at equal alarms, gaining about 8 cases per 100 from hourly updating, while outperforming 16 retrained algorithms and maintaining risk ranking in eICU. At onset, its contexts separated 2% cardiovascular-ICU from 15%-17% other 28-day mortality, and within each context after onset, persistent support carried fivefold higher mortality than de-escalation. SepsisFlow establishes a foundation-model paradigm for continuous risk and trajectory stratification in critical care.

AI 总结：本研究针对在线策略自蒸馏（OPSD）训练后期性能退化问题，提出两阶段框架自举式在线策略蒸馏（BOPD），在数学推理和代码生成任务上优于初始模型与OPSD教师。

链接：https://arxiv.org/abs/2611.13680

机构：The Hong Kong Polytechnic University（香港理工大学）

作者：Yue Wu, Xueyu Zhou, Yangrong Hu, Jian Huang

英文摘要：On-policy distillation (OPD) improves language models by distilling teacher predictions on student-generated trajectories, while on-policy self-distillation (OPSD) removes the need for an external teacher by using a reference-conditioned copy of the model as the teacher. Despite its great early improvements, we find that OPSD does not remain consistently effective under continued training and can eventually suffer substantial performance degradation. We study this training behavior and link the degradation to a growing mismatch between the evolving reference-free student and the frozen privileged teacher. In the early stage of OPSD, the privileged teacher remains closely aligned with the student and provides useful guidance for improving the student's generation trajectories. However, as training progresses, the growing distributional mismatch between the privileged teacher and reference-free student can cause the teacher to provide misleading supervision, ultimately degrading student performance. This phenomenon motivates us to first leverage privileged reference supervision to acquire useful knowledge and then transfer it to a fresh student. We introduce Bootstrapped On-Policy Distillation (BOPD), a two-stage framework that first uses OPSD to internalize task-relevant information from reference-conditioned supervision into an intermediate model, and then distills this reference-free intermediate model into a fresh copy of the initial model. Experiments across models show that BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers. These results suggest that privileged reference supervision can be more effectively exploited by first transferring its task-relevant signal into a task-adapted model and then distilling from that model without reference conditioning.

AI 总结：针对信道快速去相关导致现有CKM构建失效的问题，提出W-AoI驱动的无人机轨迹与报告方案，结合二维U-Net实现动态CKM跟踪，仿真显示其稳态误差显著低于基线方法。

链接：https://arxiv.org/abs/2611.13726

机构：College of Computer Science and Software Engineering, Shenzhen University（深圳大学计算机与软件学院）

作者：Zeyu Zhao, Yueling Che, Sheng Luo

英文摘要： Existing channel knowledge map (CKM) construction assumes quasi-static environments, failing when channels rapidly decorrelate. This paper investigates continuous online dynamic CKM tracking over an airborne 2D plane via a UAV fleet. We formulate a long-term error minimization problem under UAV mobility and reporting constraints. This problem, however, is intractable with unobservable errors for unmeasured cells. To tackle this, we propose a metric-driven scheduling approach featuring the Weighted-Age of Information (W-AoI). W-AoI evaluates cells by combining exact temporal staleness with spatial uncertainty predicted by a 2D U-Net. We prove that top-$K$ W-AoI selection maximizes the one-slot error reduction bound. Guided by this guarantee, we develop a lightweight greedy trajectory policy incorporating continuous along-path sensing. Simulations demonstrate our method reduces steady-state error by 3.8 dB$^2$ and 4.8 dB$^2$ over AoI-only and FIFO baselines, closely approaching an idealized oracle.

AI 总结：本研究针对单细胞反事实编辑中线性协方差惩罚无法分离非线性信息的问题，提出加入归一化希尔伯特-施密特独立准则的非线性概念瓶颈模型，提升了反事实表达谱生成能力。

链接：https://arxiv.org/abs/2611.13748

机构：Institute of Computational Biology, German Research Center for Environmental Health, Helmholtz Zentrum München（德国环境健康研究中心计算生物学研究所（亥姆霍兹慕尼黑中心））； Department of Computer Science, TUM School of Computation, Information and Technology, Technical University of Munich（慕尼黑工业大学计算、信息与技术学院计算机科学系）； Computational Sciences Center of Excellence, Genentech（基因泰克计算科学卓越中心）

作者：Korbinian Traeuble, Edward de Brouwer, Archit Verma, Doron Haviv, Kipper Fletez-Brant

英文摘要：Counterfactual editing of single cells, predicting how an observed cell would look had one factor been different, enables in-silico screening where experiments are infeasible, but is useful only if the edit is faithful to the individual and not merely to the condition. Concept bottleneck generative models make the effects of interventions interpretable by decomposing a cell's representation into known and unknown factors. Previous work uses a linear cross-covariance penalty to disentangle known and unknown dimensions. We demonstrate that this first-order criterion does not fully separate information when non-linear interactions are present. We first demonstrate the inadequacy of linear cross-factor penalization with simulation experiments with controlled edits and known non-linear covariate effects. We then demonstrate that adding a nonlinear kernel criterion, the normalized Hilbert-Schmidt Independence Criterion, is able to recover the true intervention effects on simulation data. Finally, we apply the non-linear concept bottleneck model to a complex single-cell RNA-seq dataset and find that it improves the ability to generate counterfactual expression profiles of cells under different tissue environments.

AI 总结：研究针对特征交互选择中分数易拟合噪声的问题，提出模型无关框架LENS，通过双模型独立复制筛选交互，在合成数据、保险与信用数据集上验证其性能优于相关方法。

链接：https://arxiv.org/abs/2611.13765

机构：HKUST(GZ)（香港科技大学（广州））

作者：Jincheng Gao, Shuo Sun

英文摘要：Learning feature interactions is essential for accurate prediction and interpretable modeling, yet a high interaction score may reflect fitting noise rather than a reproducible pattern. Ranking methods score feature pairs but leave the cutoff to the user; testing methods set the cutoff but rely on repeated model refits or a model of the joint feature distribution. We present LENS, a model-agnostic framework for selecting feature interactions through independent replication. LENS trains two models with the same learning algorithm on disjoint data and accesses them only through predictions. For each feature pair, it removes main effects from the joint-effect surface to isolate an interaction pattern, ranks pairs by its strength, and tests whether the second model reproduces it at a meaningful strength. A single multiple-testing correction determines which pairs to retain, using only two model fits regardless of the number of features. We prove that a replication ratio alone can retain arbitrarily weak patterns shared by both fits, whereas a strength floor restores selection consistency, and we characterize when normalization in Friedman's $H^2$ reverses interaction rankings. On 16 synthetic designs, LENS matches leading model-specific ranking methods (e.g., TreeSHAP) and recovers 91% of material interactions, against 45% for the knockoff-based DIAMOND, without selecting false pairs; with 1,225 candidate pairs and no interactions, it makes substantially fewer false discoveries. On large-scale French motor insurance data, it identifies interactions documented in the actuarial literature and recovers 90% of the deviance gain obtained using all pairs. On two credit datasets, LENS selects few pairs while maintaining or slightly improving predictive performance over the additive baseline.

AI 总结：JAX-WAVE是用于非均匀介质逆波设计的可微物理监督框架，通过共享神经网络实现摊销式逆波设计，在5ms在线预算下较加速数值逆降低平均误差45%，为逆波设计提供可复用流程。

链接：https://arxiv.org/abs/2611.13826

机构：University of Minnesota（明尼苏达大学）； Columbia University（哥伦比亚大学）

作者：Honghui Du, Binyao Guo, Yushu Pan

英文摘要： Designing a prescribed wave field in a heterogeneous medium requires an initial pressure distribution that compensates for the medium's propagation response. Repeated numerical inversion solves this task afresh for each target and environment. We introduce JAX-WAVE for amortized inverse wave design: a shared neural network maps desired terminal fields and known wave-speed maps directly to bounded initial pressures. Training through a differentiable finite-difference time-domain solver uses the physical design objective without requiring optimized control labels. Linear propagation and symmetric pressure bounds imply an exact sign-reversal relation for the regularized optimum. A shared-weight projection enforces this relation, preserves feasibility, and satisfies an exact identity for the paired-sign objective. We characterize the discrete propagator and its weighted adjoint to connect learning, numerical inversion, and optimality assessment. On an independent confirmation set, three trained models achieve unseen-material errors of 25.24% for geometric targets and 4.77% for reachable targets. Matched-capacity comparisons show improved reconstruction with the sign parameterization. Within a 5 ms median online budget, the learned map reduces mean error by 45% relative to accelerated numerical inversion on the heterogeneous benchmark; longer numerical solves recover greater accuracy. The results establish a reusable inverse-design procedure and show how its accuracy and computational value depend on target structure and query workload.

AI 总结：该研究针对概率神经符号推理中加权模型计数（WMC）的近似问题，提出基于张量链分解的新编译流程，大幅提升WMC估计精度，增强下游学习性能，开辟了近似编译的新设计空间。

链接：https://arxiv.org/abs/2611.13862

机构：Sony AI（索尼人工智能公司）； Fondazione Bruno Kessler（布鲁诺·凯塞勒基金会）； University of Trento（特伦托大学）； Vrije Universiteit Amsterdam（阿姆斯特丹自由大学）； University of Edinburgh（爱丁堡大学）

作者：Samy Badreddine, Emile van Krieken, Luciano Serafini, Antonio Vergari

英文摘要：Probabilistic neurosymbolic methods rely on weighted model counting (WMC) to combine neural predictors with symbolic constraints. Exact WMC is #P-hard and typically scales poorly, so many methods resort to approximations. By unifying existing approaches under a three-step bottom-up approximate compilation pipeline, we study where approximation budget is best allocated. Inspired by tensor networks, we instantiate two steps with tensor train decompositions and derive new pipelines that exactly multiply approximated factors and recompress their products via SVD- or interpolation-based schemes. Across six tasks in Sudoku, arithmetic reasoning, and scene understanding, and across increasing problem sizes, our pipelines produce WMC estimates orders-of-magnitude more accurate than prior methods. This translates into stronger downstream learning on tasks where reasoning relies on tight interactions between symbolic factors, where cruder approximations ignore these dependencies. Our pipelines open novel design space for effective approximate compilation of challenging probabilistic reasoning tasks.

AI 总结：本文概述深度神经网络机制可解释性的研究进展，旨在揭示其内部机制，为相关分析提供方法、框架与视角。

链接：https://arxiv.org/abs/2611.13868

作者：Lorenzo Cirillo, Irene Amerini, Andrea Cavallaro

英文摘要：Deep neural networks underpin fundamental tasks in a wide range of signal domains, such as image, audio, and language processing. However, their internal mechanisms are notably difficult to interpret. Internal activations are structured signals transformed across layers that may encode a multitude of interdependent relationships, as individual neurons frequently represent multiple underlying concepts. Mechanistic interpretability aims to uncover the internal mechanisms of neural networks by identifying components whose activations are causally responsible for specific behaviors. In such a context, this article provides an overview of advances in mechanistic interpretability for the analysis of causal inference in deep neural networks.

AI 总结：本研究验证消失刺激MR P300界面的目标选择可行性，发现其准确率显著高于随机水平，且对象标识提示与均匀提示的性能无显著差异。

链接：https://arxiv.org/abs/2611.13906

机构：Imperial College London（帝国理工学院）； Graduate School of Frontier Sciences, The University of Tokyo（东京大学前沿科学研究生院）

作者：Runqi Pan, Takuto Niinuma, Yunshan Huang, Masato Sugino, Kenta Shimba, Kiyoshi Kotani

英文摘要：Mixed-reality (MR) P300 interfaces can register selectable destinations, such as wheelchair goals, directly in the user's surroundings. When the cues disappear between flashes, however, the user must retain an intended target that is no longer visible. We tested whether such an interface supports target selection and whether giving each cue a stable colour-shape identity improves it compared with visually uniform spheres. Six healthy adults completed four trials in each cue condition of a 5x5 row/column task on a head-mounted MR display, with EEG from Fz, Cz, and Pz. Shrinkage linear discriminant analysis was evaluated with nested validation in which complete trials were held out from class balancing, hyperparameter selection, and fitting. Both conditions evoked a parietal target-related positivity. The decoded cell was correct in 18 of 48 held-out trials, against 1.9 expected by chance (binomial p < 1e-12), and in three participants individually. Cue format made no detectable difference: complete-target accuracy was 9/24 in both conditions (paired difference 0.0 percentage points, 95% CI [-37, 37]), and held-out AUC was 0.609 for spheres and 0.585 for object cues. Disappearing-stimulus MR P300 selection is therefore feasible under strict trial-level evaluation, but the benefit of object identity, if any, is too small to resolve in a pilot of this size.

AI 总结：该研究提出一种带联合重置与先导保证的学习方法，通过分析首次错误决策的差异信息，在有限类场景下给出学习所需的运行次数与先导数量的上界，并证明了下界的必要性。

链接：https://arxiv.org/abs/2611.13913

机构：The World Bank Group（世界银行集团）

作者：Michal Čertík

英文摘要： A run that stops at its first wrong decision shows the queries and their correct labels up to and including that decision, and nothing after it. We study how many such runs are needed for learning. Each run tests a hypothesis, the pilot, chosen beforehand from a fixed binary class of VC dimension at most $d$, on a fresh record of at most $D$ queries drawn from one unknown distribution that does not depend on the pilot. The terminal information of a record is shown only if the pilot makes no discrepancy, and a bounded loss, fixed in advance, is attached to the position where a hypothesis stops. If the discrepancy probabilities of members of the class have infimum at most $f$ and their risks have infimum at most $r$, one learner uses $\widetilde O((d+1)[D(f+\varepsilon)+r]/\varepsilon^2)$ runs and $\widetilde O(D(1+f/\varepsilon))$ distinct pilots on every history. It returns a member of the class and an interval of width at most $\varepsilon$ containing, with probability at least $1-\delta$, both the output's risk and the optimal risk. We also treat unknown ceilings and expected costs. Lower bounds for finite classes show that each order is necessary in the worst case, possibly on different instances, at positive dimension, small accuracy and fixed confidence. The learner enumerates label patterns of the class on displayed queries, possibly in exponential time; the results, under standard measurability conditions, concern statistical resources only. The proof combines a gain computable from stopped displays, exponential weights, a sparse depth-robust graph and a harmonic-mean bound.

AI 总结：针对张量轮分解的计算与内存局限，提出双核张量轮分解（DCTW）及结构缓存加速策略，在彩色视频补全等任务中实现更优的重建质量、效率与可扩展性。

链接：https://arxiv.org/abs/2611.13938

机构：School of Mathematics and Computational Science, Xiangtan University（湘潭大学数学与计算科学学院）； National Center for Applied Mathematics in Hunan（湖南省应用数学国家中心）

作者：Wenchao Xie, Qingsong Wang

英文摘要：Tensor wheel decomposition captures complex correlations in high-dimensional data through interactions between outer-ring factors and a high-order central core. However, updating these components requires the explicit construction and manipulation of large-scale matrices, which limits computational efficiency and incurs particularly high computational and memory costs for the central-core update. To address these limitations, we propose dual-core tensor wheel decomposition (DCTW), which factorizes the central core into two lower-order cores linked by a shared latent bond. Each core is coupled to a different group of outer-ring factors. We further propose a structure-cached acceleration strategy that avoids explicit large-scale matrix multi- plications and caches intermediate tensors generated during tensor network contractions for reuse across block updates. This reuse reduces redundant computation and avoids the explicit construction of large matrices. Complexity analysis and experiments show that the strategy based on structural caching accelerates optimization compared with explicit and implicit up- date schemes while limiting the increase in peak memory as the ranks grow. Experiments on color video completion and robust denoising show that DCTW matches or outperforms representative tensor network methods in reconstruction quality, computational efficiency, and scalability.

AI 总结：本文提出自监督ECG特征提取框架，含DLE与ARE模块，在多公开数据集上较现有方法获最高6.5% AUROC、12.8% F1提升，可减少标注数据需求，助力心血管疾病早期检测。

链接：https://arxiv.org/abs/2611.14141

作者：Siyuan Yan, Haibo Zheng, Feng Lu, Wei Li

英文摘要：Electrocardiogram (ECG) signals are critical for cardiovascular disease (CVD) detection. Most current deep learning models rely on supervised, task-specific training, which limits their ability to generalize across diseases, patient populations, and acquisition settings. This paper presents a self-supervised ECG feature extraction framework that learns from nearly one million unlabeled 12-lead ECG recordings. The framework includes two modules: Deviation Localization Embedding (DLE), which detects waveform-level deviations from normal patterns, and Association Reconstruction Embedding (ARE), which models cross-lead associations through deviation-guided reconstruction to capture abnormal patterns expressed across multiple leads. These modules enable the framework to capture subtle deviation-related patterns and their cross-lead context. The framework achieves up to 6.5% AUROC and 12.8% F1 improvements over prior methods on multiple public datasets, even without additional fine-tuning the pretrained feature extractor. Embedding visualizations show that the learned features support meaningful disease grouping and anomaly identification. The approach reduces the need for labeled data and can support clinicians in diverse healthcare settings. The framework has potential to improve early detection and monitoring of cardiovascular diseases, contributing to more accessible and equitable heart care.

AI 总结：提出Tree-Ising方法将语言模型修复转化为稀疏离散优化问题，在数学基准上优于现有智能体编辑方案，减少权重修改量与比特翻转次数，还在IBM Heron处理器上完成了概念验证。

链接：https://arxiv.org/abs/2611.14210

机构：QDeep； ITMO University（ITMO大学）； LEAD (Lab for Efficient AI and Data)（高效人工智能与数据实验室（LEAD））

作者：Hadi Salloum, Ammar Ali

英文摘要： Weight-space programming, introduced in recent MIT work by Takehi et al., enables a frontier LLM agent to read a network's weights and write a repair. Their GPT-6 Astra agent corrected the reasoning loops of Qwen3.5-0.8B without gradients, using 158 agent turns, 89.7 million tokens, and 50 forward-pass requests. Our method, Tree-Ising, casts this repair as a sparse discrete optimization problem over the model's own BF16 weights. Forward statistics define an objective that weakens a loop direction, strengthens a concluding direction, and preserves the layer's output on productive tokens. Within a state-tree representation, where lossless BF16 storage acts as a special case, the admissible edits are sign flips and exponent shifts that preserve all shared group exponents and mantissas. We decompose this objective exactly into independent QUBO (Ising) problems, one per output row, which we solve on a GPU using simulated annealing or quantum-inspired simulated bifurcation. On the agent's two layers, Tree-Ising achieves 54.6% on MATH-500, outperforming the agent's edit (51.7%) under the same evaluation protocol, while the base model scores 7.9% as reported. The simulated bifurcation solver yields a statistically indistinguishable repair at 55.1%. Both solvers outperform the agent edit across all five math benchmarks, with statistically significant improvements on MATH-500 and GSM8K after correcting for multiple comparisons. Our approach modifies 13% to 17% fewer weights and reduces BF16 bit flips by a factor of 2.2 to 2.3. It requires only four forward-pass requests and zero LLM tokens, with the annealing recipe costing about $0.90 in GPU time. On Qwen3.5-2B, with layers and dose selected using only the base model, the same recipe increases greedy GSM8K accuracy from 21.6% to 69.4%. As a proof of concept, we execute an eight-variable repair subproblem using QAOA on an IBM Heron processor.

AI 总结：本文提出可验证智能体流量类（VATC），结合DSCP值、路由器分类器与速率受限队列，经仿真和真实内核测试，可降低智能体排队延迟、抵御伪造攻击，优于DiffServ与fq_codel。

链接：https://arxiv.org/abs/2611.14211

机构：Comcast Corporation（康卡斯特公司）

作者：Naveen Babu Narayanan Nair, Sajay Kurup

英文摘要：Software agents built on large language models generate traffic unlike human web traffic, yet routers have no class for it and cannot verify a sender's claim to be an agent. This paper proposes the Verifiable Agentic Traffic Class (VATC), which combines a reserved Differentiated Services Code Point (DSCP) value declaring agent traffic, a router-side classifier that demotes flows contradicting the declaration to a class served after best-effort traffic, and a rate-limited priority queue. VATC was compared with static Differentiated Services (DiffServ), a dedicated class that trusts the marking, and fq_codel, a per-flow fair-queuing scheduler, in two simulators and on a real Linux kernel. VATC matched DiffServ under steady load. In a constructed burst it cut mean agent queuing delay by 98% relative to DiffServ in simulation (assuming an immediate verdict), at a 12% increase in human delay; most of the gain came from the separate class. Under a spoofed flood in the kernel, the rate limit held human latency 11 times below DiffServ's. Without verification, DiffServ and a tightly configured fq_codel gave legitimate agents lower latency than VATC; a live four-feature classifier then demoted all 20 bulk-shaped spoofers and gave agents 0.30 ms, but missed six spoofers that copied the agent pattern. With emulated network delay, a timing feature of the classifier stopped separating agent from human flows. In real agent harnesses on a congested link, router queueing took 13% to 63% of task time, depending on harness and assumed model latency. The traffic was synthetic or sandbox-scale.

AI 总结：本研究提出基于LLM的故障诊断框架CoP，结合CKG与图路径检索，在FailureSensorIQ基准上较直接提示和RAG提升了诊断准确率。

链接：https://arxiv.org/abs/2611.14263

机构：Zurich University of Applied Science（苏黎世应用科技大学）； Fragum Global（弗拉格姆全球公司）； Luleå University of Technology（吕勒奥理工大学）； Delft University of Technology（代尔夫特理工大学）

作者：Wenqi Zhou, Leonardo Baldo, Kai Goebel, Manuel Arias Chao

英文摘要：Fault diagnosis in complex systems requires linking observed changes in physical quantities to the failure mechanisms that produce them. Experienced engineers perform this task by combining sensor information with knowledge of system behavior and fault propagation. Large Language Models (LLMs) can support this process, but their diagnostic reasoning depends on the availability of asset-specific engineering knowledge and on how this knowledge is represented and used during inference. This work introduces Chain of Physics (CoP), an LLM-based fault-diagnosis framework built around a Constrained Knowledge Graph (CKG) generated automatically from asset-specific reference documentation. CoP addresses failure-sensor association from textual engineering knowledge. The CKG stores diagnostic chains connecting failure modes, changes in physical quantities, and monitoring sensors. During inference, semantically similar graph paths are retrieved for each question and used as both information and in-context examples for the reasoning process. Candidate failure modes are extracted from the question and analysed independently through dedicated LLM calls, which generate plausible failure-to-observation chains. A final decision module compares these chains with the available answer choices. CoP is evaluated on the FailureSensorIQ benchmark across four industrial asset classes and three LLMs. Compared with direct prompting and conventional Retrieval-Augmented Generation (RAG), CoP improves diagnostic accuracy across all evaluated models, with gains ranging from 7.8 to 19.6 percentage points and from 11.7 to 18.2 percentage points, respectively. The results show that structured engineering knowledge can improve LLM-based diagnostic reasoning when graph retrieval is combined with in-context examples of fault propagation.

AI 总结：本研究针对大语言模型推理保证问题，提出基于注意力头激活的监督正确性探针，可在固定预算下提升错误识别效果，且仅适用于域内场景。

链接：https://arxiv.org/abs/2611.14323

机构：Freie Universität Berlin（柏林自由大学）

作者：Mayank Gulati, Gerhard Wunder

英文摘要： Language-model systems use self-judgment to decide which answers to release, review, or recompute. Because intervention is costly, an assurance score must rank errors well at the available budget, not merely over the full ranking. We compare the judge's continuous verdict-token score with a supervised correctness probe trained on attention-head activations from the same prompt. Exact identities link fixed-budget error recall to residual risk and show that AUROC averages recall over all budgets. Across three open 7-12B models on GSM8K and MATH, the probe finds 1.3-22.5 additional percentage points of errors at 5% review in all six model-dataset settings (median 7.4 points). At 10% routing, probe-guided allocation improves accuracy over equal-cost random allocation in both GSM8K experiments, but beats judge-guided routing only for the weaker judge. Cross-dataset transfer lowers probe AUROC for every model in both directions relative to target-domain training. Because dataset, weight precision, and generation budget change together, this measures a joint protocol shift rather than a dataset-only effect. Internal signals can improve supervised, in-domain budget allocation, but are not domain-general confidence estimators.

AI 总结：该研究针对随机扩散采样器NFEs过多的问题，提出前向值离散化框架，开发F-DPMSolver和F-DDPM变体，在多数据集上降低FID，缩小随机与确定性采样的效率差距。

链接：https://arxiv.org/abs/2611.14328

机构：Chinese University of Hong Kong（香港中文大学）； University of Michigan（密歇根大学）； University of Pennsylvania（宾夕法尼亚大学）

作者：Yuchen Jiao, Na Li, Changxiao Cai, Yuting Wei, Gen Li

英文摘要：Stochastic diffusion samplers often require substantially more neural function evaluations (NFEs) than deterministic samplers to achieve comparable sample quality. Although higher-order discretization accelerates deterministic sampling, its benefits for stochastic sampling remains unclear. We establish matching upper and worst-case lower bounds showing that stochastic samplers with higher-order approximations remain first-order convergent, even with exact score functions. Thus, increasing the approximation order does not generally improve the convergence order. Motivated by this limitation, we introduce forward-value discretization, a training-free framework that evaluates the data/noise prediction model at a lookahead estimate constructed from previous predictions. For deterministic sampling, the resulting Forward DPMSolver (F-DPMSolver), improves sample quality at low NFE budgets through its DDIM-based variant, despite remaining first-order convergent. For stochastic sampling, we develop Forward DDPM (F-DDPM), which evaluates the model at a cleaner lookahead point adjusted for subsequent noise injection. This estimate reuses both the previous model prediction and the previously injected noise. Both variants admit first-order convergence guarantees with one model evaluation per iteration. Across CIFAR-10, ImageNet, FFHQ, and LSUN, F-DDPM consistently achieves lower FID than the evaluated stochastic baselines at matched NFE budgets. With 1.5 times as many NFEs as deterministic baselines, it achieves competitive sample quality, narrowing the practical efficiency gap between stochastic and deterministic sampling. Code is available at this https URL.

AI 总结：本研究针对推测解码中现有训练目标非EAL精确形式的问题，提出REAL损失，通过路径权重无偏估计与Rao-Blackwellization降方差，在多任务子集上优于现有基线，且无需修改模型架构与推理流程。

链接：https://arxiv.org/abs/2611.14399

机构：The University of Hong Kong（香港大学）

作者：Yu Tian

英文摘要：Speculative decoding accelerates LLM inference by having a draft model propose multiple tokens for the target model to verify in one pass. The speedup is governed by the expected acceptance length (EAL), yet standard training objectives such as cross-entropy, KL divergence, and total variation focus only on distribution matching or single-step overlap. Recent studies instead propose block-level training objectives, aiming to optimize EAL directly. However, these objectives remain surrogates rather than exact formulations of EAL. For example, some average later acceptance over target-distributed continuations rather than the verification branches selected by accepted tokens, introducing a covariance gap that can reverse the gradient direction. We analyze this gap and derive REAL, a real EAL training loss that uses path weights for unbiased estimation and Rao-Blackwellization for variance reduction. In a controlled loss-level comparison with a Qwen3.6-35B-A3B target and a Domino drafter, REAL outperforms WTV, VSD, DBLast, and the static LK baseline on probabilistic serving acceptance length across 9 task subsets and 5 evaluation seeds. These gains require only a change to the training loss, with no modification to the drafter architecture or inference procedure, demonstrating the practicality of real EAL training. A reference implementation of the REAL loss is available at this https URL.

AI 总结：本研究结合神经网络与符号回归的符号蒸馏方法，从21年再分析数据中发现南亚季风水汽的可解释降水方程，其解释方差优于传统浮力诊断量，参数更少且保留了神经网络的主要性能，可提升降水模型的物理一致性。

链接：https://arxiv.org/abs/2611.14401

机构：University of California, Irvine（加利福尼亚大学欧文分校）； NVIDIA Corporation（英伟达公司）； University of Lausanne（洛桑大学）

作者：Savannah L. Ferretti, Jane W. Baldwin, Michael S. Pritchard, Tom Beucler

英文摘要： South Asian monsoon precipitation depends on the thermodynamic structure of the lower troposphere, but how its ingredients combine to shape precipitation remains unclear, partly because this dependence is nonlocal in the vertical. We use symbolic distillation, which combines neural networks and symbolic regression, to discover interpretable precipitation equations from 21 years of June-August reanalysis data. Neural networks first learn the vertical thermodynamic structure most relevant to precipitation, condensing each variable into a single interpretable feature not tied to a particular vertical grid. Then, symbolic regression derives closed-form equations from these features and the surface state. Our best equation explains 1.5 times as much precipitation variance as a widely used buoyancy-based diagnostic and recovers about 85% of neural network skill while using about a thousand times fewer parameters. Unlike the buoyancy diagnostic, which adds moisture and instability, the equations treat them as competing controls, with the heaviest precipitation occurring when the lower free troposphere is humid. Over ocean, sensible heating is associated with additional precipitation mainly when near-surface air is warm and moist, improving predicted precipitation and its diurnal cycle. Although the learned features were optimized only for precipitation prediction, the equations largely respond to each feature as expected from convective physics, consistent with the reanalysis. These responses reduce the dependence of monsoon precipitation on the entire vertical column to a few simple, testable constraints that could guide data-driven precipitation models to make them more physically consistent.

AI 总结：本文提出CCW-Net，将干预中介分析作为训练目标与评估工具，使神经模型的概念解释兼具必要性、充分性与独立性，实现可量化因果性的概念解释，提升自主系统的安全性与可信度。

链接：https://arxiv.org/abs/2611.14407

机构：Computer Science and Artificial Intelligence Laboratory, MIT（麻省理工学院计算机科学与人工智能实验室）； Indian Institute of Technology Madras（马德拉斯印度理工学院）； Laboratory for Information and Decision Systems, MIT（麻省理工学院信息与决策系统实验室）

作者：Josh Rountree, Pulkit Verma, Oswin So, Chuchu Fan, Julie A. Shah

英文摘要：Concept-based explanations in neural models are important for the development, certification, and adoption of autonomous systems. Yet, interpretability remains correlational. Without methods to guide and quantify causal relationships between explanations, a concept explanation can be provided without being necessary or sufficient for the observed output thereby appearing correct while misleading. We introduce Causal Concept Wrapper Network (CCW-Net), the first use of interventional mediation analysis from fields such as economics, medicine, and epidemiology as a training objective to align causal effects of concept-based explanations with a model's output. We further apply mediation analysis as an evaluation tool to quantify, per concept label combination, how much of any model's output is causally attributable to each concept explanation. As a training objective, CCW-Net shapes a model's concept representations such that they are both necessary and sufficient for the outputs they are credited with and independent of other concepts' contributions. Counterfactual concept representations anchor a set of causal alignment criteria: necessity, sufficiency, and independence. Necessity drives irrelevant concept effects toward zero, sufficiency ensures relevant concept representations fully reconstruct a model's output, and independence enforces the total effect decomposes linearly across concepts. We evaluate CCW-Net across aircraft control, driving, and image classification. CCW-Net provides a step toward concept-based explanations that are not merely interpretable, but quantifiably causal, supporting safer evaluation, certification, and deployment of transparent and trustworthy neural models.

AI 总结：该研究通过区分三个估计量，利用果蝇回路模型等检验连接组结构-功能主张，发现部分主张成立，部分不成立，强调可信结构主张需匹配优化、数据约束生理特性及有效模型。

链接：https://arxiv.org/abs/2611.14508

机构：University of Virginia（弗吉尼亚大学）

作者：Yeyuan Liu

英文摘要：Connectome-constrained models are often used to argue that a measured wiring diagram outperforms shuffled or rewired wiring, and hence that structure shapes function. Such comparisons can confound wiring with how the models are tuned: a connectome fixes who connects to whom, not the excitability, gains or plasticity rates the model also needs, and modellers often tune these until the real-wiring model works. We separate three estimands: physiology tuned on the real wiring and transferred to nulls (P1), each wiring tuned on its own (P2), and P2 restricted to physiology consistent with independent data (P3). The biases involved are known in machine learning (under-tuned baselines, inherited weights, winner's curse); we measure what they do to biological conclusions, using synthetic ground truth, placebo nulls and three Drosophila circuit models. Some claims survive every protocol we test: segregated reward and punishment dopamine improves learning, and head-direction topography is required for a persistent bump. Others do not. In the mushroom body, the real Kenyon-cell output wiring beats a cross-subtype identity shuffle by 0.09 under P1 and P2 but loses to it by 0.09 under P3 (95% CIs exclude zero; held-out runs confirm). The verdict is set by the share of alpha'beta' cells among active Kenyon cells, which P1 selects low; at the share implied by imaging (about 44%) the real-wiring model fails to learn, exposing a model inadequacy rather than a deficit of the fly's wiring. In flyvis, P1 penalises a mirrored connectome, which has a constructed solution within 0.07 (median) of the real error, more than a spatially shuffled one in 50 of 50 members; the ranking survives ten-fold longer retraining but depends on the checkpoint at fifty-fold. A credible structural claim needs matched optimisation, data-constrained physiology and a model that still works there.

AI 总结：TorchCurves是一个PyTorch库，将参数曲线作为可学习基元，支持输入与系数求导，提供多种模块与接口，经实验验证其在部分场景计算实用，可用于多种机器学习相关应用。

链接：https://arxiv.org/abs/2611.14516

作者：Alex Shtoff

英文摘要：We present TorchCurves, a PyTorch library that treats vector-valued parametric curves as learnable primitives. Its central operation maps batches of scalar inputs to points on curves represented in B-spline or Legendre polynomial bases, with differentiation through both inputs and coefficients. Curve coefficients can be optimized jointly with the surrounding model or produced by another network. The library provides trainable modules, reusable basis modules, and a functional interface, supporting applications ranging from continuous feature embeddings to Kolmogorov-Arnold networks and functional regression. Its implementation combines classical evaluation algorithms with vectorized PyTorch operations, emphasizing readability and maintainability without custom hardware kernels. Timing experiments compare curve evaluation against NumPy and SciPy and KAN training against EfficientKAN across a range of problem sizes. The results illustrate where this generality remains computationally practical and where overhead is substantial. Library is available on GitHub at this https URL.