机器学习学术速递[6.29]

AI 总结：提出一种新的RANSAC评分函数，通过共轭逆伽马先验边缘化内点尺度，无需用户提供阈值参数，计算复杂度为O(N log N)，在近7万图像对基准测试中优于现有方法。

链接：https://arxiv.org/abs/2606.27385

作者：James Pritts, Felix Seegräber, Kevin Köser

英文摘要：The most widely used RANSAC variants score candidate models by counting inliers or summing per-point scores that saturate beyond a residual threshold. Every such score requires a user-supplied parameter that is a function of the inlier scale, which must itself be estimated from contaminated data. We remove this dependence by reversing the usual order of inference: rather than estimating the scale and then scoring against it, we marginalize the inlier scale analytically in closed form under a conjugate Inverse-Gamma prior for a fixed inlier partition, then optimize over partitions. A single closed-form expression spans the non-informative Jeffreys limit and informative empirical-Bayes priors, so the same score adapts across data-rich and data-scarce regimes without any change to the algorithm. The proposed RANSAC score is the first in which the inlier scale is genuinely absent from the formula. The score admits O(N log N ) computation via sort-and-sweep. On a benchmark of nearly 70 000 image pairs spanning different two-view estimation problems and both engineered and learned feature pipelines, the proposed score exceeds the state of the art (RANSAC, MSAC, GaU, MAGSAC): it stays nearly flat under threshold miscalibration where baselines degrade, reaches near-optimal accuracy from as few as two validation pairs where baselines need ont he order of 100 times more,. and tightens its prior regularization as validation data grows scarce.

AI 总结：提出Prism Transformer，通过逐层增加注意力头数建立局部到全局的表示层次，在参数和计算量不变下提升性能。

链接：https://arxiv.org/abs/2606.27449

作者：Shubham Aggarwal

英文摘要：Multi-head attention conventionally partitions the hidden dimension equally across all heads at every layer, enforcing an identical representational subspace dimension (dh = dmodel/h) throughout the models depth. In this work, we identify this uniform allocation as a fundamental structural bottleneck: due to their restricted dimensional space, early-layer heads are unable to faithfully capture complex, high-dimensional contextual patterns. To resolve this, we introduce the Prism Transformer, a novel architectural paradigm that replaces the static, uniform head configuration with a progressive head schedule. By monotonically increasing the head count across layers, the Prism Transformer naturally establishes a local-to-global representational hierarchy: early layers leverage fewer, exceptionally wide heads to capture complex, local compositional patterns, while deep layers deploy many, narrow heads to decompose these patterns into specialized linguistic features. Crucially, this structural shift is parameter-neutral, compute-neutral, and introduces zero training or inference overhead, preserving identical weight matrices and FLOP budgets as the standard Transformer. Across three model scales (124M, 354M, and 757M), the Prism Transformer consistently outperforms uniform baselines, achieving consistent reductions in validation loss alongside consistent gains on downstream zero-shot benchmarks (including PIQA, HellaSwag, ARC-Easy, and WinoGrande). Our findings demonstrate that non-uniform subspace allocation unlocks latent capacity within the standard Transformer budget, enabling more effective use of model capacity.

AI 总结：针对RLHF中奖励模型忽略评分者个体差异的问题，提出PEBS方法，对每个评分者拟合仿射校准器并应用经验贝叶斯收缩，在PRISM和PluriHarms数据集上分别降低RMSE 8.58%和9.66%。

链接：https://arxiv.org/abs/2606.27578

作者：Arnav Raj

英文摘要：Reward models for Reinforcement Learning from Human Feedback (RLHF) pool preferences across thousands of annotators and fit one global affine calibrator, collapsing raters with systematically different rating-scale offsets and slopes into a single average-rater fit that does not match any individual annotator. PEBS is a per-rater empirical-Bayes shrinkage estimator: it fits per-rater affine calibrators on a held-out slice of each annotator's ratings and applies Morris-James-Stein empirical-Bayes shrinkage toward the population mean, in closed form and without retraining the reward model. On PRISM, PEBS reduces within-user held-out RMSE by 8.58% over the pooled population-slope baseline. The procedure replicates on PluriHarms harm ratings (Qwen-2.5 base, in-family) with a +9.66% RMSE reduction over the same population-slope baseline. PEBS is a closed-form post-hoc estimator for annotator-specific affine calibration in RLHF reward modeling; it leaves the reward base model unchanged and estimates only the rater-level map used at inference time for new ratings.

AI 总结：提出HybridCodec，结合时间压缩离散令牌与降维连续残差，通过混合离散-连续焦点调制编解码器和混合Transformer，在离散域自回归推理并辅以非自回归预测和连续残差上采样，相比纯离散方法显著提升说话人特征保留并减少自回归步数。

链接：https://arxiv.org/abs/2606.27627

机构：Mila, Quebec AI Institute（Mila-魁北克人工智能研究所）； Concordia University（康考迪亚大学）； Sapienza University of Rome（罗马大学）； Inria, Université Grenoble Alpes CNRS, LJK（法国国家信息与自动化研究所，格勒诺布尔阿尔卑斯大学国家科学研究中心，让·库尔曼实验室）

作者：Artem Ploujnikov, Francesco Verdini, Samir Sadok, Mirco Ravanelli

英文摘要：Discrete audio representations have become increasingly popular for building multimodal text-audio systems and integrating audio capabilities into Large Language Models (LLMs). However, numerous studies report performance degradation on various downstream tasks due to information loss during discretization. To address this, we propose a novel approach combining temporally compressed discrete tokens with dimensionality-reduced continuous residuals. Our framework consists of a hybridized discrete-continuous focal modulation codec and a hybrid Transformer. This architecture performs autoregressive inference in the discrete domain, coupled with non-autoregressive prediction and continuous residual upsampling. Experimental results show that our approach significantly improves the retention of speaker characteristics compared to discrete-only methods, while simultaneously reducing the number of required autoregressive steps.

AI 总结：针对Muon优化器在高矩阵参数（如MLP层投影矩阵）中导致行范数不均匀的问题，提出Aurora优化器，通过保持极因子几何的同时强制执行行均匀性，在预训练实验中优于Muon，并达到谱优化器的最优性能。

链接：https://arxiv.org/abs/2606.27715

机构：Tilde Research

作者：Alec Dewulf, Dhruv Pai, Li Yang, Ashley Zhang, Ben Keigwin

英文摘要：We show that for tall matrix parameters, like projection matrices in the MLP layers, the Muon update can have row norms that are arbitrarily non-uniform. This can lead to a self-reinforcing feedback loop whereby neurons receive persistently small updates and eventually do not contribute meaningfully to network outputs. This problem is effectively mitigated by an additional row normalization step, but current methods do this in a way that moves the Muon update geometry away from the polar factor of the momentum matrix, which we find is undesirable. We propose Aurora, an optimizer that enforces row-uniformity of matrix parameter updates while respecting Muon's polar factor geometry. Aurora outperforms Muon in our pre-training experiments and, when combined with existing methods, achieves state-of-the-art performance among spectral optimizers on the optimizer track of the modded-nanoGPT speedrun. Additionally, we find that Aurora's empirical gains over Muon scale with the MLP expansion factor, suggesting that Aurora may allow for effective training of very wide MLP layers.

AI 总结：提出LCA框架，通过可学习信用分配解决结果监督过程奖励模型中的信用分配问题，在多个任务上优于现有方法。

链接：https://arxiv.org/abs/2606.27739

机构：National Engineering Research Center of Software Engineering, Peking University（北京大学软件工程国家工程研究中心）； School of Computer Science, Peking University（北京大学计算机学院）； Key Laboratory of High Confidence Software Technologies, Ministry of Education（教育部高可信软件技术重点实验室）； GRG Banking Equipment Co., Ltd.（广电运通金融电子股份有限公司）； Center on Frontiers of Computing Studies, Peking University（北京大学计算前沿研究中心）； Peking University Information Technology Institute (Tianjin Binhai)（北京大学（天津滨海）信息技术研究院）

作者：Tianyu Jia, Yue Fang, Hongxin Ding, Rihong Qiu, Zhibang Yang, Zhijing Wu, Xu Chu, Junfeng Zhao, Yasha Wang

英文摘要：Process reward models (PRMs) enhance the reasoning capabilities of large language models (LLMs) by providing fine-grained feedback, yet training PRMs typically requires expensive stepwise annotations. Outcome-supervised PRMs offer a scalable alternative by learning from final-answer correctness alone, but this introduces a fundamental *credit assignment* challenge, i.e., attributing outcomes to responsible reasoning steps. Existing approaches rely on either uniform or causal assignment, both of which fail to anchor credit in step correctness and thus hinder process error identification. In this work, we propose Outcome-Supervised Process Reward Modeling via **L**earnable **C**redit **A**ssignment (**LCA**), an outcome-supervised PRM framework that jointly learns credit assignment and reward modeling under the principle of *Weakest Link Assignment: a reasoning chain is as strong as its weakest link*. To address mutual dependence between credit assignment and reward modeling, we formalize outcome-supervised PRM as a Multiple Instance Learning (MIL) problem and introduce Softmax-Weighted-Sum (SWS) pooling, an MIL pooling technique tailored for strong dependence and redundancy among reasoning states. We prove Bayes consistency of our algorithm under mild assumptions. Extensive experiments demonstrate that **LCA** consistently outperforms state-of-the-art outcome-supervised PRMs across multiple tasks and backbones. Code is available at this https URL.

AI 总结：提出Flexformer，通过将随机傅里叶特征中的频谱频率作为可训练参数，实现数据驱动的注意力核学习，在保持线性复杂度的同时提升表达能力，并在语言建模和序列分类任务上优于基线。

链接：https://arxiv.org/abs/2606.27748

机构：Center for Applied Statistics and School of Statistics, Renmin University of China（中国人民大学应用统计科学研究中心与统计学院）

作者：Haoran Zhang, Feng Zhou

英文摘要： Transformer models rely on attention mechanism to capture long-range dependencies but suffer from quadratic complexity, limiting their scalability to long sequences. Kernel-based linear attention reduces this complexity but typically relies on fixed or weakly learnable kernels, restricting expressiveness and performance. In this work, we propose Flexformer, a flexible linear Transformer that learns attention kernels in a fully data-driven manner. Flexformer builds on random Fourier feature-based linear attention and treats spectral frequencies as trainable parameters, enabling the model to learn a broad family of attention kernels. We develop both stationary and nonstationary variants, with the latter offering strictly greater expressiveness. Extensive experiments on language modeling and sequence classification demonstrate that Flexformer consistently outperforms baselines. Moreover, Flexformer can be effectively distilled from pretrained Transformers to recover softmax attention and exhibits strong kernel transferability across domains, achieving both high efficiency and competitive performance on long-sequence tasks.

AI 总结：提出StoMPP方法，通过逐层随机部分掩码渐进二值化，解决STE在深层二值网络中的精度退化问题，在多个数据集上显著提升性能。

链接：https://arxiv.org/abs/2606.27759

机构：Worcester Polytechnic Institute（伍斯特理工学院）

作者：Evan Gibson Smith, Bashima Islam

英文摘要：Training binary neural networks (BNNs) from scratch is dominated by the straight-through estimator (STE), whose forward/backward mismatch produces severe accuracy degradation as networks deepen. We study an orthogonal axis: when and where binarization is enforced during training. We introduce StoMPP (Stochastic Masked Partial Progressive Binarization), which gradually replaces clipped weights and activations with their hard binary counterparts layer by layer from input to output, using stochastic partial masks with soft refresh. StoMPP delivers two complementary benefits. As a standalone training rule, it provides a fully STE-free procedure that improves over vanilla STE with gains that grow with depth (ResNet-50 BNN: +18.0/+13.5/+3.8 on CIFAR-10/100/ImageNet), and the pattern holds across ResNet-18/34/50, MobileNetV2, and BERT fine-tuning. Composed with surrogate gradients by applying STE only to frozen entries, it reaches +27.1/+19.8/+17.7 over vanilla STE on the same setting. Underlying both regimes is a single mechanistic finding: progression order is decisive. Forward layerwise progression prevents depth collapse, reverse progression collapses to near-chance, and binary-weight networks (without binary activations) are insensitive to order. We trace this asymmetry to activation-induced gradient blockades: a committed binary activation severs gradient flow upstream, and ordering controls when these blockades form. To isolate the progression's contribution from any benefit conferred by STE, we conduct all ablations in the STE-free regime; the resulting characterization (schedule, refresh, ordering, dynamics) thus reflects the progression itself rather than its interaction with surrogate gradients.

AI 总结：针对流生成器强化学习后训练中感知质量下降的问题，提出NormGuard方法，通过铰链惩罚约束速度范数，在保持奖励的同时提升图像质量和真实性。

链接：https://arxiv.org/abs/2606.27771

机构：The Hong Kong University of Science and Technology（香港科技大学）； Kuaishou Technology（快手科技）； University of Chinese Academy of Sciences（中国科学院大学）

作者：Tianlin Pan, Lianyu Pang, Cheng Da, Huan Yang, Changqian Yu, Kun Gai, Wenhan Luo

英文摘要：Reinforcement learning (RL) post-training improves the reward alignment of flow-based generators, but often degrades perceptual quality in ways that are not captured by the reward proxy. We identify a simple structural signature of this drift: across three post-training methods (NFT, AWM, DPO), RL fine-tuning inflates the per-step velocity norm $\|v_\theta\|$ by $5\%$ to $15\%$ relative to the reference. A form of norm inflation has been studied in classifier-free guidance (CFG), where rescaling the velocity back to a reference norm at inference time can mitigate the resulting artifacts. However, this inference-time correction does not transfer cleanly to RL: rescaling $v_\theta$ to match $\|v_{\text{ref}}\|$ at inference time neither improves reward nor fixes the quality degradation, because the inflation is co-adapted into the model weights. Furthermore, an adjoint sensitivity analysis shows that velocity magnitude rescaling carries no coherent first-order reward signal at the batch level, indicating that suppressing norm inflation is unlikely to remove a consistently reward-carrying component. Since inference-time renormalization fails while norm suppression carries no reward cost, training-time intervention is the appropriate strategy. Together, these findings motivate \methodname, a hinge penalty that activates only when $\|v_\theta\|$ exceeds $\|v_{\text{ref}}\|$ and composes additively with any velocity-local base loss. Across two base models, three post-training methods, and two reward proxies, \methodname consistently improves MLLM-judged image quality and forensic realism while preserving reward, with gains that amplify under few-step inference and are not explained by early stopping.

AI 总结：提出TA-SparseMG轻量级模型，结合趋势感知归一化、尺度自适应门控去噪和多尺度门控注意力MLP模块，解决长期时间序列预测中的非平稳性、高频干扰和跨周期依赖问题，在多个基准上取得稳定优异性能。

链接：https://arxiv.org/abs/2606.27908

机构：School of Mathematical Sciences, Guizhou Normal University（贵州师范大学数学科学学院）； School of Big Data and Computer Science, Guizhou Normal University（贵州师范大学大数据与计算机科学学院）

作者：Wenchao Liu, Hongbing Wang, Youji Zhu, Xiaodong Liu, Xiangguang Xiong

英文摘要： Long-term time series forecasting finds extensive applications in domains such as power demand, traffic flow, meteorological observation, and renewable energy dispatch. Forecasting dynamically varying long-term time series poses inherent challenges, including statistical nonstationarity, local high-frequency disturbances, and coupled cross-period dependencies, which make it difficult for lightweight models to balance parameter efficiency and forecasting performance. To address this issue, this study presents TA-SparseMG, a lightweight cross-period forecasting model built on SparseTSF's sparse cross-period modeling framework. It incorporates three key modules: a trend-aware reversible instance normalization module, a scale-adaptive gated denoising module, and a multiscale gated-attention MLP forecasting module. The trend-aware normalization module captures input-window statistics and calibrates forecast-window distributions, effectively mitigating distribution shift. The scale-adaptive gated denoising module performs feature smoothing and residual suppression before period rearrangement, thereby reducing interference from high-frequency perturbations. The multiscale gated attention prediction module strengthens the prediction head's adaptive representational capacity via conditional gating and feature modulation. Extensive experiments across multiple LTSF benchmarks demonstrate that the proposed TA-SparseMG consistently achieves superior, stable performance. Ablation studies confirm that each module independently improves distribution adaptation, input robustness, and cross-period feature mapping capability.

AI 总结：提出MixTTA模块，通过低秩跨通道变换增强归一化层，解决分布偏移下通道间结构变化问题，提升测试时自适应方法的鲁棒性。

链接：https://arxiv.org/abs/2606.28142

机构：Seoul National University（首尔大学）； Kookmin University（国民大学）； HodooAI Labs（HodooAI实验室）

作者：Mansoo Jung, Youngwook Kim, Jungwoo Lee

英文摘要：Test-Time Adaptation (TTA) methods commonly update the affine parameters of normalization layers to adapt deployed models under distribution shifts. However, per-channel affine parameters perform axis-aligned scaling and shifting, making them geometrically incapable of correcting cross-channel structural changes induced by distribution shift. To address this limitation, we propose MixTTA, a lightweight plug-in module that equips normalization layers with a low-rank cross-channel transformation, enabling inter-channel mixing at each layer. To ensure that the low-rank branch captures only cross-channel interactions, we also propose Decoupling Projection that enforces strict separation from the diagonal affine path, along with Spectral Projection that prevents rank-1 collapse under non-stationary test streams. MixTTA can be seamlessly integrated into any existing normalization-based TTA method. Experiments in both standard and wild TTA settings show consistent improvements over strong baselines while mitigating adaptation failure under challenging conditions. The source code is publicly available at this https URL.

AI 总结：针对蛋白质共折叠模型中的成对表示，提出PairSAE方法，通过N-mode SVD总结成对张量为token级交互角色，再用稀疏自编码器学习共享特征，实现可解释性并避免传统SAE的二次爆炸问题。

链接：https://arxiv.org/abs/2606.27440

机构：University of California, Irvine（加州大学尔湾分校）； Flagship Pioneering

作者：Giosue Migliorini, Aristofanis Rontogiannis, Grigori Guitchounts, Nicholas Franklin, Axel Elaldi, Olivia Viessmann

英文摘要：Foundation models for structural biology have achieved remarkable performance in predicting biomolecular structure and show promise for the design of proteins and small molecules. Yet understanding which internal features drive their outputs remains challenging. Standard sparse autoencoders (SAEs), effective on transformer-style sequence embeddings, do not transfer cleanly to pairformer-like architectures: naively operating on pairwise representations yields a quadratic blow-up of features and obscures concepts distributed jointly across sequence and pair representations. We introduce PairSAE, which summarizes pairwise tensors via an N-mode SVD into token-wise interaction roles, then uses a sparse autoencoder to learn a shared set of token-level features that decode into both sequence and pair representations. Evaluated on Boltz-2 activations for PLINDER protein-ligand complexes, PairSAE yields interpretable features that align with UniProt annotations and predict Boltz-2 affinity values. These results indicate that PairSAE links the latent space of foundation models for structural biology to interpretable structural concepts, clarifying what the model "knows" while avoiding pairformer-induced pitfalls that limit conventional SAEs.

AI 总结：提出严格中介原则解决LLM中历史旁路导致的潜在状态不可识别问题，引入离散文本潜在状态和因子化GRPO方法，在TextWorld和ScienceWorld上实现表示质量和滚动性能显著提升。

链接：https://arxiv.org/abs/2606.27681

机构：Intuit AI Research（Intuit AI研究）

作者：Xiang Gao, Kaiwen Dong, Yuguang Yao, Padmaja Jonnalagedda, Kamalika Das

英文摘要：World models in partially observed environments rely on latent representations that summarize interaction history, but in many modern LLM-based architectures predictive performance fails to reflect representation quality due to history bypass, rendering the latent state unidentifiable. Strict latent state mediation, requiring predictions to depend only on the latent state and action, is a classical principle that resolves this, but enforcing it in text-based settings is an open challenge: textual latent states are discrete and non-differentiable, precluding variational training, and expressive LLM decoders readily ignore the bottleneck. We show how to make strict mediation work in the text domain. We formalize why it is necessary, showing that strict mediation makes representation quality empirically testable while history-leaky architectures break this connection. We then introduce textual latent states, which are discrete, interpretable, and variable-length, and factorized GRPO (fGRPO), a tree-structured reinforcement learning method that enforces strict mediation during training. Experiments on TextWorld and ScienceWorld show preserved one-step prediction accuracy alongside up to 57\% gains in representation quality and 98\% improvements in rollout performance, increasing with task complexity and horizon.

AI 总结：提出OperatorSHAP，一种网格无关的归因方法，用于高效估计神经算子的Shapley值，支持不规则网格和跨分辨率迁移。

链接：https://arxiv.org/abs/2606.28065

机构：LMU Munich（慕尼黑大学）； MCML（慕尼黑机器学习中心）； DFKI（德国人工智能研究中心）

作者：Joshua Stiller, Santo M. A. R. Thies, Felix Czaja, Eyke Hüllermeier

英文摘要： Understanding model predictions is essential for physical applications, where outputs often inform safety-critical decisions, such as structural load assessment, weather warnings, and clinical diagnosis. Shapley values satisfy many desirable properties as an attribution method, but their computational cost during inference hinders their practical use. Current amortized explainers, such as FastSHAP, are limited to homogeneous inputs, which is problematic for physical applications where data often comes from irregular grids and geometries. We introduce OperatorSHAP, a grid-agnostic attribution method and training procedure that allows us to train FastSHAP-like explainers for neural operators. We establish a theoretical framework for attributions in function space, connecting to Aumann-Shapley values. We further show that OperatorSHAP's explanations are consistent with state-of-the-art discrete Shapley values across resolutions and transfer across grid sizes without retraining.

AI 总结：针对连续时间潜在随机微分方程模型的可辨识性问题，利用环境引起的扩散协方差偏移，证明在漂移无稀疏性假设下，对角扩散机制可识别潜在坐标至置换和缩放，并估计潜在解缠与因果图。

链接：https://arxiv.org/abs/2606.28228

机构：Mohamed bin Zayed University of Artificial Intelligence（穆罕默德·本·扎耶德人工智能大学）； The University of Melbourne（墨尔本大学）； Peking University（北京大学）； Carnegie Mellon University（卡内基梅隆大学）

作者：Yuanyuan Wang, Wenjie Wang, Haoxuan Li, Mingming Gong, Kun Zhang

英文摘要：Causal representation learning for time series has developed strong identifiability results in discrete-time latent causal models, but identifiability in continuous-time latent stochastic differential equation (SDE) models remains largely open. We address this gap using environment-induced shifts in diffusion covariance. We study additive-noise latent SDEs observed through an unknown nonlinear diffeomorphism, with shared drift but environment-specific diffusion covariance. We show that two diagonal diffusion regimes with pairwise distinct coordinate-wise variance ratios identify the latent coordinates up to permutation and scaling, without any sparsity assumption on the drift. We first prove this result for linear Ornstein--Uhlenbeck systems and then extend it to general additive-noise latent SDEs. Under mild smoothness, the instantaneous drift-Jacobian causal graph is identifiable up to the same permutation. We propose a two-stage estimator for latent disentanglement and optional graph recovery; experiments on synthetic systems confirm the predicted identifiability boundary, and an application to Hardanger Bridge monitoring data illustrates the approach on real sensor trajectories.

AI 总结：研究状态不可观测且决策时段可能受限的马尔可夫强盗问题，提出自退化马尔可夫强盗模型，设计UCB-NOM算法实现近对数遗憾，并证明无先验知识时遗憾超对数下界。

链接：https://arxiv.org/abs/2606.27448

机构：IRIT, Université de Toulouse, CNRS, Toulouse INP（图卢兹大学IRIT实验室，法国国家科学研究中心，图卢兹国立理工学院）； Ikerbasque-UPV/EHU, University of the Basque Country（巴斯克科学与技术联盟-巴斯克大学）

作者：Thomas Hira, Victor Boone, Urtzi Ayesta, Ina Maria Verloop

英文摘要：This paper studies the problem of regret minimization in Markovian bandits with \emph{non-observable states} and possibly \emph{constrained} decision epochs. The focus is restricted to a ``pure'' regret benchmark, that compares the performance of the learning algorithm to the best \emph{pure policy} which -- akin to optimal policies of stochastic bandits -- picks the optimal arm from start to finish without ever switching. We introduce a generalization of rested Markovian bandits, \emph{self-degrading Markovian bandits}, for which pure policies are always asymptotically this http URL show that without prior knowledge on the underlying bandit, the regret of algorithms that switch arms rarely necessarily scales super-logarithmically for every bandit, i.e., as $\omega(\log(T))$, where $T$ is the learning horizon. Despite the unreachability of the logarithmic regime, we design UCB-NOM, an optimistic algorithm inspired by UCB, of which the regret is nearly logarithmic. Lastly, we show that given prior knowledge on the Markovian bandit in the form of a bound on the bias functions of its arm, a proper instantiation of UCB-NOM achieves $O(\log(T))$ regret. We further show that this prior knowledge allows for a $O(\sqrt{T \log(T)})$ worst-case regret bound for UCB-NOM. Notably, our regret bounds do not depend on the number of states of the underlying Markov chains. Our findings suggest that the non-observability of states is a mild inconvenience in self-degrading Markovian bandits.

AI 总结：针对RLHF中奖励信号延迟的问题，提出回溯优势修正（RAC），通过非负核函数队列化延迟反馈并注入优势估计，理论证明其无偏性，在表格MDP实验中偏差降低达47.9倍。

链接：https://arxiv.org/abs/2606.27580

作者：Arnav Raj

英文摘要：Reinforcement learning from human feedback (RLHF) in production does not always have a synchronous reward signal. Code-execution verifiers, slow judge ensembles, and queued human review can return several gradient steps after the rollout that produced them, breaking the synchronous-reward assumption underlying standard PPO. We address this gap with Retroactive Advantage Correction (RAC): each pending slow completion is queued, aged through a non-negative kernel, and reinjected as a clipped residual into the next optimiser step's advantage. We prove that under an unbiased clipped importance ratio, the cumulative RAC correction is exactly unbiased when the effective delay kernel reinjects all of its mass, and carries a bias linear in the unreinjected fraction otherwise; at the no-delay identity kernel it reduces to V-trace. On a tabular Markov decision process (MDP) proof-of-concept, RAC reduces the closed-form policy bias by up to 47.9x at the two-slow-channel configuration, beating wait-for-slow at lower wall-clock cost. RAC integrates with PPO and GRPO through a two-line reward-manager patch.

AI 总结：本文研究在通信受限下，通过强化学习训练智能体策略，使其动作更易用于状态估计，并在飞机跟踪问题中验证了可观测性提升对任务性能影响极小。

链接：https://arxiv.org/abs/2606.27609

机构：Department of Aerospace and Mechanical Engineering（航空航天与机械工程系）

作者：Andres Enriquez Fernandez, John J. Bird

英文摘要：Physical or operational constraints often impose communications limitations on autonomous agents. Such limitations complicate monitoring or multiagent coordination. Even when strong communications are absent, some information may still be available. The remainder of the relevant agent state may be reconstructed via estimation. The actions taken by an agent are a potential source of information -- as the agent interacts with the environment, these actions may be observed even in the absence of explicit communication. We investigate using actions to estimate the state of an agent, using reinforcement learning to develop policies which make the estimation problem more tractable. Policy observability is encouraged through the training reward and is analyzed using simulation of the trained agent. In an aircraft tracking problem a policy with enhanced observability is found that has minimal impact on nominal task performance.

AI 总结：提出RS-Diffuser，一种结合扩散轨迹生成与分布价值评论家的风险敏感离线规划框架，通过尾部感知目标引导去噪过程，实现灵活的风险偏好行为，在风险敏感D4RL和机器人导航基准上达到最优性能。

链接：https://arxiv.org/abs/2606.27766

机构：Northwestern Polytechnical University（西北工业大学）

作者：Shiqiang Gong

英文摘要：Offline reinforcement learning enables policy learning from fixed datasets without additional environment interaction, making it appealing for safety-critical applications where online exploration is costly or unsafe. Diffusion-based decision-making methods have recently achieved strong performance in offline RL by modeling rich, multimodal trajectory distributions. However, existing diffusion planners are typically risk-neutral and therefore may overlook rare but catastrophic outcomes that are crucial in real-world deployment. In this work, we propose RS-Diffuser, a risk-sensitive offline diffusion planning framework that combines diffusion-based trajectory generation with distributional value critics. RS-Diffuser learns a diffusion planner over future state trajectories, a separate inverse dynamics model for action decoding, and a Monte Carlo distributional critic that estimates the full return distribution of candidate plans through quantile regression. At sampling time, we incorporate a risk-sensitive guidance signal into the denoising process, using gradients computed from tail-aware objectives such as Conditional Value at Risk to steer generation toward desired risk profiles. As a result, a single trained model can flexibly produce risk-averse, risk-neutral, or risk-seeking behaviors by changing only the inference-time risk parameter. Extensive experiments on risk-sensitive D4RL and risky robot navigation benchmarks demonstrate that RS-Diffuser achieves state-of-the-art performance, improving both overall return and worst-case robustness while reducing safety violations.

AI 总结：提出GraphDR-LinUCB算法，将臂特征投影到图的低频谱子空间，实现从d到k的维度缩减，并证明首个~O(k√T)遗憾界，通过扰动分析扩展到噪声图，在六个真实数据集上减少15倍累积遗憾。

链接：https://arxiv.org/abs/2606.27917

作者：Joyanta Jyoti Mondal, Ibne Farabi Shihab, Anuj Sharma

英文摘要：Contextual bandits with graph-structured arms arise in recommendation, citation retrieval, and social advertising, where arms connected on a graph tend to share reward signal. Standard dimensionality reduction ignores this structure, inflating exploration cost by a factor of $d/k$. We propose GraphDR-LinUCB, which projects arm features onto the graph's low-frequency spectral subspace and runs linear UCB in the resulting $k$-dimensional space. We prove the first $\wtO(k\sqrt{T})$ regret bound for spectral-projection-based contextual bandits, reducing dimension dependence from $d$ to $k$; a perturbation argument extends this to noisy graphs, with an explicit penalty for reward-smoothness mismatch and graph-estimation error. Our central theoretical finding is that the high-frequency reward component need not incur a worst-case linear-in-$T$ penalty: its actual cost depends on its realized impact along the played path, not on its total energy. A simple spectral comparison between subspaces ($\Gamma_k$) predicts which reducer wins on a given dataset, correctly calling five of six real-dataset outcomes without any fitted threshold. Across a synthetic benchmark and six real datasets (MovieLens, Amazon, LastFM, ogbn-arxiv, MIND), GraphDR-LinUCB reduces cumulative regret by $15\times$ over full-dimensional LinUCB and outperforms competing graph-aware methods on five of six; the single failure is precisely where the graph's spectral subspace is misaligned with the reward.

AI 总结：提出KL耦合策略正则化(KCPR)框架，通过策略间动态先验交互实现奖惩强化学习中的策略协调，并衍生出KL耦合软最优性(KCSO)及深度实现klDMP，在网格世界和Gazebo机器人导航任务中提升了安全性和学习稳定性。

链接：https://arxiv.org/abs/2606.28152

机构：Dept. of Brain Robot Interface, ATR Computational Neuroscience Laboratories（ATR计算神经科学实验室脑机接口系）

作者：Jiexin Wang, Eiji Uchibe

英文摘要：We propose KL-Coupled Policy Regularization (KCPR), a policy coordination framework for Reward-Punishment Reinforcement Learning (RPRL). Based on KCPR, we derive KL-Coupled Soft Optimality (KCSO) and develop its deep realization, klDMP. Unlike existing RPRL approaches that optimize reward-seeking and punishment-related policies largely independently, KCPR enables direct interactions between companion policies by treating each as a dynamically learned prior for the other. KCSO yields coupled soft-optimal policies and KL-regularized Bellman operators, allowing reward and punishment information to jointly influence value propagation. To improve learning stability, we introduce a companion-prior softening mechanism and evaluate separate replay-buffer designs for balancing reward- and punishment-related experience. Experiments in grid-world and Gazebo robotic navigation tasks demonstrate that klDMP improves safety and learning stability while maintaining competitive task performance compared with DQN, SQL and softDMP. These results suggest that policy-level coordination provides an effective mechanism for integrating multiple behavioral objectives and may serve as a useful design principle for reinforcement learning systems with interacting motivational processes.

AI 总结：提出LLawCo框架，通过反思失败提取行为模式并推导高层合作法则，结合监督微调融入推理，提升具身多智能体在分散部分可观测环境中的合作效率与任务成功率。

链接：https://arxiv.org/abs/2606.28182

作者：Qinhong Zhou, Chuang Gan, Anoop Cherian

英文摘要：Embodied agents operating in decentralized and partially observable environments have attracted growing attention in recent years. However, existing large language model (LLM)-based agents often exhibit behaviors that are misaligned with their partners or inconsistent with the environment state, leading to inefficient cooperation and poor task success. To address this challenge, we propose a novel framework, Learning Laws of Cooperation (LLawCo), that enables embodied agents to autonomously align with both their partners and task objectives. Our framework allows agents to reflect on past failures to extract misaligned behavioral patterns, which are used to derive high-level behavioral laws, such as "Talk when necessary" and "Wait for partner." These laws are explicitly incorporated into the agents' chains of thought via supervised fine-tuning, aligning their reasoning with task requirements and the behavior of other agents. To evaluate our approach, we introduce PARTNR-Dialog, a large-scale multi-agent communicative and cooperative planning benchmark built on the PARTNR environment. Experiments on existing tasks and our new benchmark demonstrate significant improvements in cooperative efficiency and task success rates. Across four backbone LLMs, our method achieves average success rate improvements of 4.5% on the PARTNR-Dialog benchmark and 6.8% on the TDW-MAT benchmark over state-of-the-art open-source communicative agent frameworks. See the LLawCo project page for details: this https URL

AI 总结：提出首个量子生成扩散模型QDiffusion-TS，通过量子神经网络替换去噪Transformer中的前馈组件，在IQM量子处理器上验证，在金融时间序列合成中Wasserstein距离降低44%，下游预测RMSE提升71%。

链接：https://arxiv.org/abs/2606.27561

机构：Kingston University London（伦敦金斯顿大学）； University of Florence（佛罗伦萨大学）； Brunel University of London（伦敦布鲁内尔大学）

作者：Jack Waller, Filippo Caruso, Dimitrios Makris, Rajagopal Nilavalan, Xing Liang

英文摘要：Generative models have achieved remarkable success in data synthesis, though recent advances driven by increasing model scale have introduced challenges in computational cost and efficiency. Quantum machine learning offers a promising alternative, representing complex data distributions using compact, highly expressive models. Here, we propose QDiffusion-TS, the first quantum generative diffusion model for time series synthesis, and validate it on the IQM quantum processor. The framework extends a classical diffusion architecture by replacing feed-forward components within the denoising transformer with quantum neural networks, yielding a hybrid quantum transformer that reduces the number of trainable parameters in each replaced component by nearly three orders of magnitude. Evaluated on financial time series from Apple and Amazon, the model generates synthetic data that more accurately reproduces the real distributions, reducing Wasserstein distance by approximately 44% relative to its classical counterpart across both datasets. In a downstream forecasting task, augmentation with the generated data improves predictive performance by up to 71% in RMSE over a baseline trained solely on real data. These results show that quantum enhanced architectures can consistently match and frequently surpass classical performance with substantially fewer parameters, establishing a practical framework towards more efficient and scalable data-driven generative modelling.

AI 总结：针对低频类别在扩散模型中因低密度区域导致评分估计不准和生成质量差的问题，提出类别频率引导的噪声调度（CFRG），为低频类别分配更大尺度噪声，在CIFAR-100-LT和ImageNet-LT上显著提升生成质量。

链接：https://arxiv.org/abs/2606.27696

机构：Hefei University of Technology（合肥工业大学）； University of Science and Technology of China（中国科学技术大学）； The University of Hong Kong（香港大学）； The Chinese University of Hong Kong（香港中文大学）； Nanyang Technological University（南洋理工大学）

作者：Jiequan Cui, Beier Zhu, Qingshan Xu, Xiaojuan Qi, Bei Yu, Hanwang Zhang

英文摘要：In this paper, we are the first to examine the correlations between class frequency and the multi-scale noise schedule within diffusion models. For score-based generative models, low-density regions often lead to inaccurately estimated scores, thereby compromising the generation quality. Although the multi-scale noise schedule can alleviate this issue during the diffusion process, low-frequency classes still face the challenge of large low-density regions, resulting in more inaccurate estimated scores than high-frequency classes. Furthermore, high-frequency classes tend to dominate the score space, causing a convergence of most data points towards generating samples from these classes. Consequently, samples generated within low-frequency classes exhibit suboptimal quality and limited diversity. To address this challenge, we propose the \textit{Class-frequency Guided (CFRG)} noise schedule, leveraging the insight that low-frequency classes should be endowed with larger-scale noises. To illustrate the effectiveness of our method, we conduct experiments on various tasks, including image generation, image classification, and text-to-image generation, using imbalanced datasets, \textit{i.e.}, CIFAR-100-LT, and ImageNet-LT. By employing the CFRG noise schedule, we achieve substantial improvements over baselines, manifesting the crucial role of frequency statistics in noise schedule design.

AI 总结：提出Pepti-drift框架，通过单一抗原条件漂移步骤生成肽候选，在嵌入空间中吸引抗原匹配结合肽并排斥毒性相关区域，采用预热策略稳定竞争目标。

链接：https://arxiv.org/abs/2606.27824

机构：SB Intuitions

作者：Takashi Fujiwara, Hikaru Shindo, Kaushalya Madhawa, Jun Jin Choong, Keisuke Ozawa

英文摘要： Peptides are a promising therapeutic modality that combine the chemical tunability of small molecules with the target specificity of macromolecular therapeutics. However, designing antigen-specific binding peptides while avoiding toxicity remains a major challenge for therapeutic peptide discovery. Here, we present Pepti-drift, a toxicity-aware latent refinement framework that generates peptide candidates through a single antigen-conditioned drift step. In a peptide embedding space, Pepti-drift learns to attract generated peptide latents toward antigen-matched binding peptides while repelling them from toxicity-associated regions. This is challenging because binding-promoting physicochemical features often overlap with toxicity-associated features in peptide representation space. To address this, we introduce a warm-up strategy to stabilize this competing objective by first learning binding-oriented attraction and then increasing toxicity repulsion.

AI 总结：提出MDM-VGB，一种结合奖励引导重掩码的离散扩散采样器，通过扩展回溯随机游走到掩码状态图，实现高效高奖励生成和低奖励样本修复，理论上具有二次复杂度且对过程验证器噪声鲁棒。

链接：https://arxiv.org/abs/2606.28301

作者：Kijung Jeon, Thuy-Duong Vuong, Molei Tao

英文摘要：Inference-time scaling is a promising paradigm to improve generative models, especially when outputs must satisfy structural constraints or optimize downstream rewards. We consider Masked Diffusion Model (MDM) and introduce MDM-VGB, a discrete diffusion sampler that augments unmasking generation with theoretically principled reward-guided remasking. Inspired by the recent success of the classical Jerrum-Sinclair backtracking Markov chain in reward-tilted generation, MDM-VGB extends the backtracking random walk from a fixed prefix tree to a masked-state graph, allowing tokens to be unmasked and remasked at arbitrary positions. The resulting sampler favors unmasking and remasking moves that lead to higher-value partial configurations, enabling both effective high-reward generation and efficient repair of low-reward samples. We prove that MDM-VGB is robust to process-verifier noise and achieves quadratic complexity, while popular test-time heuristics such as best-of-$N$ can incur exponential complexity due to error accumulation. Our theoretical findings are corroborated by strong empirical performance, particularly on popular constraint-satisfaction and scientific benchmarks such as Sudoku and QM9.

AI 总结：通过递归分解序列为子问题并组合解决方案，课程学习在模拟半自动机任务中实现了比亚直接方法更优的统计复杂度，显著降低了监督需求和参考模型覆盖条件。

链接：https://arxiv.org/abs/2606.27721

作者：Nived Rajaraman, Audrey Huang, Miroslav Dudik, Robert Schapire, Dylan Foster, Akshay Krishnamurthy

英文摘要：Compositional generalization, the ability to solve complex problems by combining solutions to simpler sub-problems, is a fundamental capability of both natural and artificial intelligence, and a key mechanism underlying chain-of-thought reasoning. However, the theoretical underpinnings of compositional generalization remain poorly understood: when and why does decomposing a problem into parts yield more efficient learning than solving it directly? We study this question through the canonical problem of learning to simulate semiautomata (predicting the outcome of $T$ steps of sequential computation), a model that captures state tracking, regular language recognition, and modular arithmetic. We show that an autocurriculum-based approach building on Part I of this series, recursively decomposing longer sequences into shorter sub-problems, learning to solve them, and composing the solutions, achieves dramatically better statistical complexity than direct methods. (i) For a setting inspired by supervised fine-tuning (SFT) where the learner receives interactive feedback on intermediate states of the computation, curriculum facilitates learning from only $2^{\mathcal{O}(\sqrt{\log T})}$ tokens of supervision; i.e., subpolynomial in the sequence length $T$, overcoming the $\Omega(T)$ token barrier required by direct simulation. (ii) For a setting inspired by reinforcement learning with verifiable rewards (RLVR), where the learner improves a pre-trained reference model using an outcome verifier, we show that curriculum reduces the requirement on the reference model from coverage at the full sequence length $T$ to coverage at a shorter block length $B \ll T$, an exponentially weaker condition.

AI 总结：研究Wasserstein空间中具有凸差分解的目标函数，提出Wasserstein CCCP算法，证明其收敛性，并应用于MMD和能量距离优化，实验表明比Wasserstein梯度下降更快更稳定。

链接：https://arxiv.org/abs/2606.27767

机构：CMAP, CNRS, Ecole Polytechnique, IP Paris（CMAP，CNRS，巴黎综合理工学院，巴黎-萨克雷大学）； CERMICS, CNRS, ENPC, IP Paris（CERMICS，CNRS，国立桥路学校，巴黎-萨克雷大学）； IBM Research（IBM研究院）

作者：Clément Bonet, Pierre-Cyril Aubin-Frankowski, Youssef Mroueh

英文摘要：Optimizing functionals over the space of probability measures is now ubiquitous in machine learning. A widely used approach is to perform the optimization directly over the Wasserstein space, but many objective functionals of practical interest are non-convex along Wasserstein geodesics, making the analysis of standard first-order methods challenging. In this work, we study a class of objectives over the Wasserstein space that admit a difference-of-convex (DC) decomposition and we lift the classical convex-concave procedure (CCCP) to this setting. Under smoothness and strong convexity assumptions on the convex components of the decomposition, we prove almost stationarity along the iterates of the resulting algorithm. Our main focus is on the Maximum Mean Discrepancy (MMD) and the Energy Distance (ED) functionals, for which we develop explicit Wasserstein DC decompositions, and establish local convergence of the scheme under mild assumptions. Empirically, we show that well-chosen DC decompositions yield faster and more stable convergence than Wasserstein gradient descent on these MMD objectives.

AI 总结：本文证明非仿射聚合规则破坏梯度单调性，导致收敛不稳定和泛化退化，并给出恢复单调性的充分条件。

链接： https://arxiv.org/abs/2606.28123

机构：PreMeDICaL team, Inria, Idesp, Inserm, Université de Montpellier, Montpellier, France（PreMeDICaL团队，法国国家信息与自动化研究所，Idesp，法国国家健康与医学研究院，蒙彼利埃大学，法国蒙彼利埃）； Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189, CRIStAL, F-59000 Lille, France（里尔大学，法国国家信息与自动化研究所，法国国家科学研究中心，中央里尔高等电力工程学院，UMR 9189，CRIStAL，法国里尔）； Department of Computer Science, University of Copenhagen, Copenhagen, Danemark（哥本哈根大学计算机科学系，丹麦哥本哈根）

作者：Thomas Boudou, Batiste Le Bars, Nirupam Gupta, Aurélien Bellet

英文摘要：Last-iterate convergence and generalization guarantees in first-order convex learning hinge on the monotonicity of the update operator. While linear averaging preserves the monotonicity of gradient updates, this property is often violated when gradients are aggregated non-affinely, as in modern pipelines enforcing constraints like adaptivity, privacy, robustness or fairness. Whether it is possible to design non-affine aggregation rules that maintain monotonicity has remained an open question. We answer this question negatively: we prove that the monotonicity of aggregated gradients is preserved if and only if the aggregation rule is positively affine. Consequently, non-affine aggregation prevents steady convergence and substantially degrade algorithmic stability. We quantify these drawbacks and propose a path forward by identifying sufficient conditions under which monotonicity can be restored. Our results provide a unified theoretical framework explaining the disparate failure modes observed in modern learning systems.

AI 总结：研究二次两层网络中泛化误差随样本数、宽度和正则化的缩放规律，揭示由目标谱结构控制的数据相关幂律和相变。

链接：https://arxiv.org/abs/2606.28242

机构：Statistical Physics of Computation Laboratory, École Polytechnique Fédérale de Lausanne (EPFL)（统计物理与计算实验室，洛桑联邦理工学院）； Information, Learning and Physics Laboratory, École Polytechnique Fédérale de Lausanne (EPFL)（信息、学习与物理实验室，洛桑联邦理工学院）

作者：Julius Girardin, Emanuele Troiani, Yizhou Xu, Vittorio Erba, Florent Krzakala, Lenka Zdeborová

英文摘要：Understanding how performance scales jointly with model size and data is a central problem in modern machine learning. Existing theoretical works on scaling laws typically describe generalization as a function of data or compute, often in fixed-feature or infinite-width regimes and for online SGD. Here, we instead study how generalization scales with the number of trainable parameters and the number of samples in a feature-learning model. We analyze $\ell_2$-regularized empirical test error minimization in a quadratic two-layer network in a finite-sample setting with structured data. This setting allows for an explicit characterization of the generalization error as a function of the number of samples, model width, and regularization. Our results reveal a phase diagram with distinct scaling regimes as the number of parameters varies. In particular, the generalization error follows data-dependent power laws controlled by the spectral structure of the target. We further characterize the transitions between regimes, including the onset of interpolation, and their impact on generalization.

AI 总结：针对层次稀疏预测编码中迭代推理成本高的问题，提出混合摊销推理方法，结合LISTA式编码器快速初始化与少量能量校正步骤，在保持稀疏性和重建质量的同时显著降低延迟。

链接：https://arxiv.org/abs/2606.27802

机构：Department of Clinical Engineering, Komatsu University（小松大学临床工程系）

作者：Kazuhisa Fujita

英文摘要：Hierarchical predictive coding provides an interpretable framework for perception as error-driven inference in multi-layer generative models, while sparse coding imposes parsimonious latent representations through explicit sparsity constraints. Their combination yields hierarchical sparse predictive coding models with appealing computational and neuroscientific properties, but practical use is often limited by the cost of iterative latent inference. In such models, each input may require many recurrent refinement steps before a useful sparse representation is obtained, and this burden becomes more severe as the hierarchy deepens. We study this bottleneck by holding the hierarchical sparse energy fixed and varying the inference procedure. The comparison includes four schemes: classical iterative inference based on ISTA, an accelerated MFISTA reference, structurally informed amortized inference using a LISTA-style bottom-up encoder adapted to the hierarchical model, and a hybrid method in which this fast amortized initialization is followed by a small number of corrective energy-based refinement steps. Under this shared objective, we measure reconstruction quality, sparsity, latency, and stability on static image benchmarks. The results show that a shallow LISTA-style initializer plus short corrective recurrence improves over pure amortization while remaining much faster than long iterative inference.

AI 总结：提出一种任务无关的逐层能耗估计模型，在超过10万层、295种架构的数据集上中位误差19.6%，优于现有方法，并展示逐层分解可通过共享层泛化到新任务。

链接：https://arxiv.org/abs/2606.27841

机构：Bell Labs, Nokia Networks France（诺基亚网络法国贝尔实验室）； Inria, Université Côte d’Azur, France（法国国家信息与自动化研究所，蔚蓝海岸大学）； Université Côte d’Azur, CNRS, Inria, I3S, France（蔚蓝海岸大学，法国国家科学研究中心，法国国家信息与自动化研究所，信息与系统科学实验室）

作者：Adrien Sardi, Marie-Line Alberi Morel, Sara Alouf, Frédéric Giroire, Joanna Moulierac

英文摘要： The widespread adoption of Artificial Intelligence (AI) has led to increasing concerns about energy consumption, yet there is a lack of standardized methodologies to accurately estimate AI inference energy consumption, particularly across various tasks and architectures. In this study, we propose a task independent, layer-wise energy estimation model for AI architectures. Our model is evaluated on a large dataset of more than 100,000 layers for 295 neural network architectures across 3 widely-used tasks and 3 distinct hardware platforms. Our approach achieves a median error of 19.6%, outperforming state-of-the-art methods. We further show that layer-wise decomposition generalize to new tasks without complete retraining, by leveraging shared layers across architectures. It offer tools, insights and a precise methodology to empower stakeholders in designing energy-efficient AI systems.

AI 总结：提出FlexMoE方法，通过专家通道重要性排序和离散动作学习，将预训练MoE大语言模型转换为嵌套子网络族，支持跨预算部署，在Qwen2-57B-A14B上剪枝50%路由专家参数仍保留约99.8%性能。

链接：https://arxiv.org/abs/2606.27866

机构：National University of Singapore（新加坡国立大学）

作者：Fan Mo, Yuxuan Han, Geng Zhang, Wangbo Zhao, Yang You

英文摘要：Mixture-of-Experts (MoE) language models scale model ability with sparsely activated experts, making this architecture a standard recipe for modern large models. However, sparse activation does not remove the deployment burden of storing and serving all experts, and the available deployment budget can vary substantially across devices, users, and workloads. Existing MoE compression methods are still largely fixed-budget, typically optimizing one compressed endpoint at each chosen target budget. We study a different setting: converting a large pretrained MoE LLM into a nested family of deployable subnetworks across budgets. Our method first ranks expert FFN channels by their importance, then lets each expert learn a discrete action to prune its channels. By gradually increasing cost pressure, a single action-training run exports a series of action masks from high to low budgets, each of which identifies a reliable smaller subnetwork nested in the ranked base model. Moreover, we use a single recovery fine-tune at a mid pruning budget (40%) to recover degraded model quality and transfer the recovered model to other unseen budgets. Overall, our framework surpasses recent MoE compression baselines. Specifically, on Qwen2-57B-A14B, our method retains ~99.8% of base performance while pruning 50% of routed expert parameters even without fine-tuning. For deployment, our pruned subnetworks deliver real memory reduction and throughput gains, and further support realtime online budget switching with kernel-level co-design.

AI 总结：提出参数高效混合Transformer（PEHT），通过低秩适应和多模态融合整合城市移动性与拥堵信息，在减少参数的同时提升网络流量预测精度。

链接：https://arxiv.org/abs/2606.28274

机构：National Science Foundation（国家科学基金会）； U.S. Department of Transportation（美国交通部）； University Transportation Center (UTC)（大学交通中心）； Transportation Cybersecurity Center for Advanced Research and Education (CYBER-CARE)（交通网络安全高级研究与教育中心）； NVIDIA Academic Grant Program（NVIDIA学术资助计划）

作者：Abdolazim Rezaei, Mehdi Sookhak, Mahboobeh Haghparast

英文摘要：Accurate network traffic prediction is a critical element for efficient resource allocation in dynamic urban cellular networks. However, prediction remains challenging because network demand is influenced by complex mobility patterns, congestion dynamics, and heterogeneous user behavior. This paper introduces the Parameter-Efficient Hybrid Transformer (PEHT), a network traffic prediction framework that integrates urban mobility and congestion information into a Transformer-based architecture. PEHT separates primary network communication features from secondary urban mobility features and incorporates Low-Rank Adaptation (LoRA) into the Transformer encoder to reduce the number of trainable parameters while maintaining high predictive accuracy. A multimodal fusion strategy then injects external mobility and congestion features into the decoder to improve traffic forecasting. Experiments on the Telecom Italia Milan dataset and multiple synthetic congestion scenarios show that PEHT outperforms state-of-the-art baselines in terms of RMSE, MAE, and $R^2$. The implementation is available in the GitHub repository.

AI 总结：提出隐私保护的概率种族/民族估计方法，结合隐私技术实现美国LinkedIn成员种族/民族公平性度量，并展示其在候选人和观众侧公平性测量中的应用。

链接：https://arxiv.org/abs/2606.27558

机构：LinkedIn（领英）

作者：Osonde A. Osoba, Yuzi He, Saikrishna Badrinarayanan, Varun Mithal, Sakshi Jain, Natesh S. Pillai

英文摘要：Fairness measurements in the form of disaggregated evaluations often rely on demographic signals that are legally constrained or culturally sensitive. Race and ethnicity signals are among the more difficult signals to curate and use for this task. This paper presents Privacy-Preserving Probabilistic Race/Ethnicity Estimation (PPRE) as a method for enabling fairness measurements with respect to race/ethnicity for U.S.\ LinkedIn members in a privacy-preserving manner. PPRE applies privacy technologies (specifically: secure two-party computation, differential privacy, and additive homomorphic encryption) on top of two race/ethnicity demographic signal sources (the Bayesian Improved Surname Geocoding estimator and a sparse golden survey set of self-reported demographics) to power a fairness measurement solution with respect to US-based race/ethnicity demographics. We detail its privacy guarantees and demonstrate its application on candidate- and viewer-side fairness measurements. We close with a transferable framework for institutions seeking to implement similar privacy-preserving measurement infrastructure.

AI 总结：提出FoggyTrust，通过分层信任网络将信任计算本地化到雾节点，结合本地信任聚合与异构感知全局优化器，在异构数据下提升鲁棒性，在CIFAR-10上比FLTrust提升超50%。

链接：https://arxiv.org/abs/2606.27622

作者：Emmanuel Rassou, Tomas Gonzalez

英文摘要： Byzantine-robust federated learning seeks to protect distributed model training from malicious or corrupted clients without requiring access to their private data. FLTrust addresses this challenge by introducing a trusted server-side root dataset that assigns trust scores to client updates for more robust aggregation. In this work, we propose FOGGYTRUST, a hierarchical extension of FLTrust that localizes trust computation to fog nodes, allowing the framework to better handle globally heterogeneous data while preserving robustness within locally homogeneous client groups. We further show that this two-level architecture can simultaneously address distribution mismatch in trust estimation and client drift across groups by combining local trust-based aggregation with heterogeneity-aware global optimizers such as FedAdam and SCAFFOLD. Across benchmark datasets, FOGGYTRUST achieves its strongest gains on more challenging heterogeneous settings, particularly on CIFAR-10 under Krum and Trim attacks, where it achieves an over 50% improvement over FLTrust. We also test FOGGYTRUST in a real-world safari dataset to show the promise of hierarchical trust networks for robust federated learning in socially impactful, safety-critical settings such as distributed wildlife monitoring.

AI 总结：针对仅API黑盒访问下大模型遗忘难题，提出受控行为差异（CBD）框架，利用辅助模型构建保留与遗忘输入的行为差异并路由遗忘提示，通过基于梯度统计的判别基提升相似数据区分度，实现更优的遗忘-效用权衡。

链接：https://arxiv.org/abs/2606.27683

机构：State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications（北京邮电大学网络与交换技术国家重点实验室）； Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing（北京未来区块链与隐私计算高精尖创新中心）； Department of Electronic and Electrical Engineering, Sungkyunkwan University（成均馆大学电子与电气工程系）

作者：Zhiqiang Xie, Yijing Lin, Zhipeng Gao, Dong In Kim

英文摘要：Edge devices increasingly invoke large language models (LLMs) through API services for context aware edge intelligence, while edge generated data may be collected to improve LLMs and may introduce sensitive, copyrighted, harmful, or outdated information into model behavior. Machine unlearning offers a practical way to remove the influence of undesired data without retraining LLMs. However, existing methods still face two gaps. The first is API only black box access, where target model parameters and internal logits are unavailable. The second is how to preserve retained utility when unlearning target data and retained data share highly similar prompt structures or semantic patterns. To address these challenges, we propose Controlled Behavioral Divergence (CBD), an API only black box unlearning framework. CBD uses two auxiliary models to create controlled behavioral divergence between retained inputs and unlearning target inputs, converts this divergence into an unlearning relevance score, and routes unlearning related prompts away from the target LLM. To improve discrimination accuracy under high similarity between target and retained data, CBD constructs a gradient statistics based discriminative basis by estimating empirical Fisher matrices and solving a regularized generalized eigenvalue problem, guiding the unlearning signal toward target specific information rather than shared prompt structures. Compared with eleven white box and gray box unlearning baselines, CBD achieves a better unlearning utility trade off and its performance varies little across settings. On ToFU forget10, CBD approaches the retrained reference on the forget set while raising model utility to 74.90, about 15% above the second best baseline. On WMDP, it lowers hazardous knowledge accuracy to 25.68, near random guessing, while preserving MMLU accuracy of 52.67. Code is at this https URL.

AI 总结：提出在完全委托的AI合作社中，基于异质价值约束筛选更新，利用遍历学习实现细粒度信用分配，以解决数据估值和联邦学习中的贡献归属问题。

链接：https://arxiv.org/abs/2606.28217

机构：Hongik University（弘益大学）

作者：Young Yoon, Jimin Kim, Soyeon Park

英文摘要：We propose a framework for reward allocation in fully delegated AI cooperatives where humans are represented by agents that contribute data and participate in model updates under heterogeneous value constraints. The key idea is to credit only those updates that remain admissible after screening them against each principal's value profile. We formulate value-conditioned gradient filtering, online marginal contribution signals, and cumulative revenue settlement within a traversal learning (TL) substrate. TL is especially attractive here because it performs decentralized backpropagation without the quality loss associated with aggregation-centric distributed learning and, we argue, offers a finer attribution substrate than FedAvg-style federated learning by preserving explicit traversal and gradient paths. The framework is positioned against data valuation, federated contribution estimation, personalized federated learning, and pluralistic alignment.

AI 总结：提出一种元学习框架实现随时有效的认证鲁棒性，通过轻量级元学习器预测图像特定先验，将样本复杂度降低20倍，同时保持严格统计保证。

链接：https://arxiv.org/abs/2606.27694

机构：University of Melbourne（墨尔本大学）； DST Group, Adelaide（国防科技集团阿德莱德分部）

作者：Andrew C. Cullen, Paul Montague, Benjamin I.P. Rubinstein

英文摘要：Randomized Smoothing (RS) provides rigorous robustness guarantees for neural networks without architectural constraints, yet its adoption is limited by extreme computational costs. Standard RS requires tens of thousands of model evaluations per input and forces practitioners to commit to fixed sample sizes a priori. In this work, we present a novel meta-learning framework for anytime-valid certified robustness that adaptively deploys computational resources. By using a lightweight meta-learner to predict image-specific priors for a sequential E-process, we achieve a 20-fold reduction in sample complexity compared to traditional methods while maintaining rigorous statistical guarantees. Beyond raw efficiency, we demonstrate how anytime-validity enables adaptively allocating compute based upon application-specific risk thresholds, a form of resource triage impossible under classic certification frameworks. That this is achievable while also providing similar certification performance demonstrates that our approach provides a pathway for real-time, safety-critical certification deployments.

AI 总结：针对自动语音识别系统对扰动敏感且缺乏真实转录的问题，提出双门诊断管道（双侧原子审计和基于排名的锦标赛），实现词错误率降低55%并提供词级和句级认证。

链接：https://arxiv.org/abs/2606.27698

机构：University of Melbourne（墨尔本大学）； DST Group（国防科技集团）

作者：Andrew C. Cullen, Neil Marchant, Jiani Xie, Paul Montague, Benjamin I.P. Rubinstein

英文摘要：Automatic Speech Recognition systems are notoriously both sensitive to adversarial and benign perturbations. While this has been repeatedly demonstrated using reference datasets, detecting such behaviors in deployed systems is incredibly challenging, due to the absence of oracle knowledge of the true transcription. We demonstrate that employing a certification-inspired mechanism can significantly decrease WER, increase recall, and decrease the Spearman correlation between confidence and WER. We achieve this through a dual-gate diagnostic pipeline: a Two-Sided Atomic Audit that accumulates statistical wealth to certify both token existence and adversarial exclusion, and a Rank-Based Tournament that selects the winning sequence. Our evaluations across four diverse architectures demonstrate up to a 55% relative reduction in Word Error Rate, while also providing granular word- and sentence-level certifications to enhance acoustic security.

AI 总结：提出USAD方法，通过方差差异和扰动协方差差异两个新统计量捕捉对抗样本的全局与局部不确定性特征，在多种攻击下优于现有基线。

链接：https://arxiv.org/abs/2606.27832

机构：The University of Melbourne（墨尔本大学）

作者：Zhijian Zhou, Xunye Tian, Jiacheng Zhang, Zesheng Ye, Yiyi Guo, Donghao Zhang, Liuhua Peng, Feng Liu

英文摘要：Statistical adversarial detection (SAD) treats detection as a two-sample test. Given a reference set of clean examples (CEs) and a batch of queries, potentially containing an unknown mixture of CEs and adversarial examples (AEs), SAD decides whether the query distribution drifts away from the CE distribution while controlling the false-alarm rate. Existing SAD-based methods mainly use maximum mean discrepancy (MMD) to measure the distributional discrepancy. However, MMD's distributional properties limit its ability to capture characteristic uncertainty patterns of AEs that are crucial for detection: AEs typically exhibit abnormal feature spread (i.e., global uncertainty) and instability under perturbations (i.e., local uncertainty). To close the gap, we propose Uncertainty-aware Statistical Adversarial Detection (USAD), which explicitly captures these uncertainty patterns with two new statistics: (1) Variance Discrepancy (VD), which measures the difference in feature spread between AEs and CEs to capture global uncertainty differences. (2) Perturbation-based Covariance Discrepancy (PCD), which compares feature covariance under Gaussian perturbations to capture local uncertainty differences. By aggregating VD and PCD, USAD achieves superior detection performances over baseline methods against various adversarial attacks, highlighting the importance of considering characteristic behaviors of AEs for effective SAD. Our code is available at: this https URL.

AI 总结：提出RECAST方法，利用Wasserstein重心原型和反事实解释，在有限数据和受限访问下重建黑箱模型，解决决策边界偏移和过拟合，实现高保真度和公平性审计。

链接：https://arxiv.org/abs/2606.27948

作者：Xuan Zhao, Lena Krieger, Zhuo Cao, Arya Bangun, Hanno Scharr, Ira Assent

英文摘要：Counterfactual explanations (CFs) help understand machine learning models by identifying minimal input changes that would lead to alternative model outcomes. Recent work demonstrates their utility for reconstructing black-box models, enabling third-party auditing of opaque decision systems for fairness and accountability. Still, CF-based reconstruction may suffer from decision boundary shifts, overfitting, and restrictive assumptions requiring online query access to target platforms. We propose REconstruction via Counterfactual-Aware waSserstein opTimization (RECAST) under limited data and restricted access, a behavioral surrogate model based on Wasserstein barycentric prototypes. Our approach addresses decision boundary shifts by incorporating CFs as informative, though less representative, samples for both classes, maintaining high surrogate fidelity in low-sample regimes without requiring online access during reconstruction. To enhance fairness auditing, our method enables systematic group fairness diagnostics. Experiments on real-world datasets and various setups show that RECAST effectively achieves high fidelity and query efficiency, as well as stable results even when the access is limited and noisy.

AI 总结：提出一种基于梯度优化的公平分类算法，通过学习有效特征表示，在无需重新训练的情况下实现与处理中方法相当甚至更优的公平性-准确性权衡效率。

链接：https://arxiv.org/abs/2606.28097

作者：Maaya Sakata, Kazuto Fukuchi

英文摘要：Post-hoc controllability of fair machine learning models, the ability to control the trade-off between fairness and accuracy after training, is valuable for practical deployment. Existing post-processing methods provide such post-hoc controllability but often suffer from significant accuracy degradation, whereas in-processing methods achieve efficient trade-offs but require computationally expensive retraining for each change in trade-off ratio. To achieve both post-hoc controllability and efficient trade-offs, we propose a novel fair classification algorithm that learns effective feature representations to improve the trade-off efficiency of post-processing fair classifiers, by a gradient-based optimization approach. Experimental results on real-world datasets demonstrate that our method achieves trade-off efficiency comparable to, or even surpassing, in-processing methods, without requiring any retraining.

AI 总结：针对TeRo模型在关系映射特性和时间信息表达上的不足，提出TeRoR方法，通过解耦实体嵌入的时间演化、独立旋转头尾实体增强时间建模，并利用半径约束头实体在尾实体圆形区域内以捕捉关系映射特性，在四个数据集上取得竞争性能。

链接：https://arxiv.org/abs/2606.27651

机构：South China Normal University（华南师范大学）

作者：Peijia Xie, Yike Liu, Chao He, Huiling Zhu

英文摘要：In recent years, with the emergence of Temporal Knowledge Graphs (TKGs), research on learning entity and relation representations in TKGs has attracted increasing attention, giving rise to a large number of TKG embedding methods. TeRo is a simple and efficient temporal knowledge graph embedding approach. However, TeRo does not do well in modeling the mapping properties of various relations, such as one-to-many, many-to-one, and many-to-many. Meanwhile, it also has limitations in the expression of temporal information. To address these issues, we propose a novel TKG embedding method named TeRoR. This method divides the temporal evolution of entity embeddings, and conducts independent rotation transformations on head and tail entities in the complex vector space to strengthen temporal information modeling capacity. In terms of relational characteristics, we train a radius to constrain the rotated and translated head entities within a circular region centered on the tail entity, which effectively captures the diverse mapping properties of relations. Experimental results demonstrate that TeRoR achieves competitive performance against state-of-the-art models on four distinct TKG datasets.

AI 总结：提出GNBAN架构，通过异构图表示学习和可解释的基分解头，实现大规模零售需求预测，在M5和Favorita数据集上WRMSSE降低4-5%。

链接：https://arxiv.org/abs/2606.27863

机构：Phi Labs, Quantiphi

作者：Janak M. Patel, Anirudh Deodhar, Dagnachew Birru

英文摘要：Demand forecasting at the bottom of a retail hierarchy requires predicting tens of thousands of correlated long-horizon series across products, stores, and regions. Modern systems must scale across massive catalogs, capture shared demand dynamics, and remain interpretable enough to be trusted. Classical statistical methods need a separate model per series and are hard to manage at scale; deep autoregressive models struggle as the joint state grows to tens of thousands of dimensions; and recent graph-based forecasters, while capturing cross-entity dependencies, often produce opaque long-horizon forecasts. We propose GNBAN (Graph Neural Basis Attention Network), an end-to-end architecture combining heterogeneous graph representation learning with an interpretable basis-decomposition head. Retail data are represented directly as a heterogeneous graph derived from the relational schema, so a single model serves the entire catalog. Rather than predicting the horizon directly, GNBAN decomposes each forecast into trend, seasonal, and generic components. Its key innovation is a per-basis attention mechanism: each basis function keeps its own learnable query and retrieves information independently from the entity's historical neighborhood, letting different bases specialize to distinct temporal patterns while preserving interpretability. On two large-scale benchmarks, M5 Walmart and Favorita Grocery Sales, evaluated under matched protocols, GNBAN improves volume-weighted WRMSSE by roughly 4-5% over a matched graph baseline. Qualitative analysis shows the learned decomposition exposes trend, seasonal, and residual demand drivers without post-hoc explanation methods. These results demonstrate that scalable relational forecasting and interpretable forecast decomposition can be achieved together in a unified graph-based framework.

AI 总结：针对多模态数据的不完整和时间异步问题，提出基于对偶学习的惩罚多对齐聚类模型（DLPMAC），通过对偶学习保持语义和结构一致性，并利用惩罚机制实现多对多数据对齐，提升对齐精度并避免数据聚合。

链接：https://arxiv.org/abs/2606.27984

机构：Dalian University of Technology（大连理工大学）； Hainan University（海南大学）

作者：Liang Zhao, Shubin Ma, Bo Xu, Qingchen Zhang

英文摘要：Multimodal feature fusion can effectively capture complex patterns in real-world data by integrating complementary information from different modalities. However, in many applications, such as boiler combustion monitoring, equipment failure, inconsistent sensor sampling frequencies, and network delays often cause missing modalities and temporal asynchrony. These issues lead to incomplete and disorderly multimodal data. To address them, previous studies have proposed several data fusion methods that align cluster centers before fusion. However, these methods have two key limitations. First, they cannot guarantee accurate sample-level alignment of data pairs. Second, they do not address significant discrepancies in data sizes across different classes, which may affect subsequent fusion performance. To address these problems, we propose a dual-learning based penalized multi-align clustering model, named DLPMAC. The dual-learning mechanism enables the model to learn prior knowledge from each modality, including semantic and structural information. This helps preserve semantic consistency and structural similarity across modalities at both local and global levels. In addition, the penalized multi-align module performs multi-to-multi data alignment through a penalty mechanism. It allows one sample to form data pairs with different samples from other modalities, thereby improving data-pair alignment accuracy. The penalty mechanism also prevents data aggregation, avoiding the case where excessive samples are linked to a single sample. Experimental results demonstrate the effectiveness of DLPMAC in addressing data alignment and fusion challenges from both sampling and clustering perspectives.

AI 总结：针对图欺诈检测中标签稀疏不平衡和表示稀释问题，提出ADC-GNN框架，结合扩散引导特征增强、对比学习和多跳谱注意力，在1%训练数据下显著提升检测性能。

链接：https://arxiv.org/abs/2606.28134

机构：Central South University（中南大学）； University of Technology Sydney, Australian Artificial Intelligence Institute（悉尼科技大学，澳大利亚人工智能研究所）

作者： Liming Liu, Chao Hu, Mingfei Lu, Yiwei Ge, Xingle Li, Heyuan Shi

英文摘要：Graph-based fraud detection is essential for safeguarding large-scale transaction systems, where undetected anomalies may lead to substantial financial losses and security risks. Real-world fraud graphs pose two coupled challenges: sparse and imbalanced supervision, where verified fraudulent labels are scarce and heavily skewed toward benign accounts, and representation dilution, where spatial message passing may oversmooth camouflaged anomalies while spectral filters may suppress fraud-relevant mid- and high-frequency irregularities. To address these challenges, we propose ADC-GNN, short for Attention-guided Diffusion-Contrastive Graph Neural Network, a unified framework that combines diffusion-guided feature augmentation, contrastive representation learning, and multi-hop spectral attention for few-shot graph fraud detection. The diffusion component is formulated as a feature-space denoising augmentation mechanism rather than a full topology-generative graph diffusion model: it constructs noise-perturbed node-feature views under a cosine schedule and uses contrastive learning to stabilize node representations across perturbations. The spectral attention module further adaptively emphasizes fraud-relevant hop-level and relation-level cues. We evaluate ADC-GNN primarily on three public benchmarks and additionally report a proprietary real-world telecom transaction dataset with approximately 60,000 records as a private case study. Under the 1% training setting, ADC-GNN achieves consistent improvements over original graph fraud baselines and four protocol-consistent recent graph anomaly/fraud baselines on the public benchmarks. Additional analyses on split stability, training ratios, oversampling alternatives, module-level ablations, diffusion schedules, and runtime and memory-consumption comparisons further characterize the effective operating regime of ADC-GNN.

AI 总结：研究小语言模型在顺序LoRA个性化中的灾难性遗忘问题，提出基于检查点的监测协议，利用轻量级参考集分布诊断揭示模型特定不稳定性模式。

链接：https://arxiv.org/abs/2606.27634

机构：MALTA, Machine Learning Theory and Applications Lab, PUCRS（MALTA，机器学习理论与应用实验室，PUCRS）； Kunumi Institute（Kunumi研究所）

作者：Thomas S. Paula, Lucas S. Kupssinskü, Rodrigo C. Barros

英文摘要：Small Language Models (SLMs) are increasingly being considered for deployment on edge devices such as laptops, enabling private, low-latency, and locally personalized applications. However, personalization requires models to adapt over time to evolving user- or task-specific data, placing them in a continual learning setting. This creates the risk of catastrophic forgetting, where learning new information degrades performance on previously learned tasks or broader model capabilities. Recent benchmarks such as TRACE have shown that continual fine-tuning can significantly degrade the general abilities of aligned large language models. In this work, we present a study for sequential LoRA personalization of SLMs. We save model checkpoints after each adaptation stage and evaluate them on current tasks, previously seen tasks, and a fixed reference set. This checkpoint-level protocol enables us to monitor task performance, forgetting, and reference set drift over time. We show that lightweight reference set distributional diagnostics can reveal model-specific instability patterns during sequential LoRA personalization of SLMs, including cases where task-level metrics alone hide harmful adaptation. We hope this can highlight new research avenues for monitoring stability of SLMs in a continual learning setting.

AI 总结：本文发现持续学习中的LoRA适配器存在显著低秩冗余，提出LiteLoRA门控机制，通过复用现有适配器减少20-70%活跃适配器，同时保持或超越现有性能。

链接：https://arxiv.org/abs/2606.28117

作者：Tanguy Dieudonné, Giulia Lanzillotta, Enis Simsar, Louis Barinka, Thomas Hofmann

英文摘要：Low-Rank Adaptation (LoRA) has become the standard tool for parameter-efficient fine-tuning of large pretrained models. When applied sequentially across tasks in Continual Learning (CL), the standard assumption is that each new task requires a dedicated low-rank adapter. In this work, we challenge this assumption empirically and structurally. We show that task-specific LoRA adapters in CL exhibit significant low-rank redundancy: the subspaces spanned by adapters trained on different tasks substantially overlap, and in many cases earlier adapters can faithfully represent later tasks. Building on this observation, we propose LiteLoRA, a plug-and-play gating mechanism that learns at train time whether to recruit a new adapter or reuse existing low-rank representations. Our method reduces the number of active adapters by 20-70% while matching or exceeding state-of-the-art performance on standard CL benchmarks, revealing that structural redundancy is pervasive and that selective learning is sufficient to achieve stability without sacrificing plasticity.

AI 总结：针对基础模型接口碎片化问题，Darts 开发了统一 FoundationModel 类集合，提供标准化全周期预测接口，实现零样本或微调预测、不确定性估计和回测。

链接：https://arxiv.org/abs/2606.27438

作者：Zhihao Dai, Dennis Bader, Alain Gysi

英文摘要： Since its initial release in 2020, Darts has become a widely used open-source Python library for time series analysis. A series of foundation models have recently claimed accuracy improvements in zero-shot forecasting, promising a paradigm shift from training custom models to harnessing pre-trained general-purpose forecasters. Foundation models, however, are often released as isolated packages with fragmented interfaces and limited interoperability with common tooling, making joint evaluation and integration within complete pipelines difficult. In Darts, we developed a unified $\texttt{FoundationModel}$ class collection (Chronos-2, TimesFM 2.5, TiRex, PatchTST-FM) that provides standardized, full-cycle forecasting interfaces with minimal external dependencies for integrating foundation models into the ecosystem. Existing Darts pipelines can now use foundation models with only a name change; new pipelines can use them for zero-shot or fine-tuned forecasting, uncertainty estimation, and backtesting, combined with data processing and evaluation tooling, all within a unified framework.

AI 总结：提出KARMA方法，通过构建马尔可夫替代模型捕捉时间依赖，从三个核心方面（最小历史长度、最优转移核、五级全局解释层次）解释时间序列预测器，在真实和合成数据上验证其恢复因果结构和识别时间依赖的能力。

链接：https://arxiv.org/abs/2606.27599

机构：XITASO GmbH

作者：Amadeo Tunyi

英文摘要：While many explainable AI (XAI) methods have been proposed, most are not designed for time-series forecasting models and often rely on the implicit assumption that timestamp features are independent. This assumption ignores the fundamental property of temporal dependence and can lead to explanations that violate the sequential and causal structure of the data. We introduce \textsc{KARMA}, a method for explaining time-series predictors by constructing a Markov surrogate model that captures the temporal dependencies learned by the predictor. Our approach revolves around three main aspects: identifying the minimal history length $K$ that is predictively sufficient for the model, estimating the best-fitting $K$-order Markov transition kernel from the discretized history space, and a five-level global explanation hierarchy that can be derived from the Markov transition kernel, which we illustrate using real-world weather data (Beijing PM 2.5). We also certify using complex synthetic data with known true causal edges that KARMA (i) recovers the data causal structure as learned by the model via a controlled experiment and (ii) identifies temporal dependencies better than established attribution methods such as TimeSHAP.

AI 总结：本文系统评估了Chronos-2和MOMENT等时间序列基础模型在电子鼻数据上的嵌入表示，发现微调是必要步骤，且融合专用模型表示可提升气体识别和浓度预测性能。

链接：https://arxiv.org/abs/2606.27672

作者：Taeyeong Choi, Mohammed Kamruzzaman

英文摘要：Inspired by advances in natural language processing and computer vision, "time-series foundation models" (TSFMs) have recently been introduced with the promise of strong generalization across diverse time-series tasks, including forecasting, classification, and anomaly detection, as well as across domains such as healthcare, climate science, and manufacturing. However, their utility for gas-sensing data remains largely unexplored. To address this gap, this paper systematically evaluates recent TSFMs on electronic nose (E-Nose) data. In particular, we investigate whether embeddings produced by representative TSFMs, including Chronos-2 and MOMENT, provide effective representations for gas identification and concentration prediction. Specifically, we show that fine-tuning is necessary to achieve satisfactory performance on E-Nose data, and fusing TSFM embeddings with representations learned by specialized predictive models can further improve the performance, suggesting both the potential and limitations of current TSFMs for gas-sensing applications.

AI 总结：提出一种通过决策理论预训练学习时间序列估计器的神经网络框架，通过指定生成世界和决策目标，训练网络逼近最优决策规则，在零样本推断中实现近乎最优风险、偏差控制和均匀校准。

链接：https://arxiv.org/abs/2606.27711

机构：Google（谷歌）

作者：Pablo Montero-Manso, Marcel Scharth

英文摘要：We introduce a neural network-based framework for learning time series estimators through a process we term decision-theoretic pretraining. Analysts specify a generative world, a distribution over data-generating processes, and a target decision objective. A neural network trained on stratified simulations from this world approximates the corresponding optimal decision rule, yielding a neural estimator that provides forecasts, parameter estimates, predictive intervals, or model-selection for zero-shot inference on previously unseen time series. The joint specification of the generative world and objective enables the estimators to directly approximate process-level, finite-sample properties: near-optimal risk, bias control, minimax performance, and uniform calibration. Our experiments demonstrate that these neural estimators can outperform traditional baselines such as maximum likelihood estimation and model selection via AICc, for the same model structural model classes. Furthermore, even when trained purely on simulations of structural models, they achieve competitive or state-of-the-art forecasting accuracy on major real-world benchmarks, compared with statistical, neural or large pre-trained models. We illustrate the framework by addressing two longstanding challenges: finite-sample bias and miscalibration in AR(p) models, and the forecast combination puzzle. These applications highlight the approach's main advantage: its ability to approximate solutions to analytically intractable or computationally prohibitive time series problems, including complex structural equations or optimality criteria. Ultimately, by enabling explicit control over decision-theoretic trade-offs, the framework equips analysts with highly efficient estimation tools tailored to their specific analytical needs.

AI 总结：系统比较七种传感器融合方法在HARMES多模态数据集上的性能，发现门控多模态融合在留一参与者评估中宏F1分数达0.82，优于拼接式后期融合基线（0.76）。

链接：https://arxiv.org/abs/2606.27886

机构：University of Siegen（锡根大学）

作者：Ahmed Mohamady, Robin Burchard, Kristof Van Laerhoven

英文摘要： Recent advances in Human Activity Recognition (HAR) from wearable sensors have shown that multi-modal deep learning models consistently outperform their uni-modal counterparts. Modalities can include IMUs, RGB cameras, audio signals, and others. One important aspect of multi-modal deep learning is the sensor fusion approach we apply. Over recent years, multiple fusion paradigms have been proposed for multi-modal HAR. However, to the best of our knowledge, no head-to-head comparison of these paradigms exists on a common multi-modal HAR benchmark dataset. To address this research gap, we systematically compare seven state-of-the-art sensor fusion methods on the recently released HARMES dataset, which comprises 61 hours of fully labeled IMU, audio, and ambient humidity data. The chosen dataset focuses on 15 household and personal hygiene activities of daily living (ADLs). By applying the seven different fusion techniques to a state-of-the-art multi-modal model architecture, we show that Gated Multi-modal Fusion achieves the highest macro F1-score (0.82), surpassing the concatenation-based late fusion HARMES paper baseline of 0.76 by +6pp under leave-one-participant-out evaluation. All code used in our experiments is made publicly available on GitHub.

AI 总结：提出一个框架，通过评估不同选择策略如何保持全局模型排名，从大量数据集中选择代表性子集，并利用自举聚合提供置信区间。在时间序列分类和NLP基准上，简单策略如FAFI优于随机选择；在推荐系统中改进不显著。

链接：https://arxiv.org/abs/2606.27997

机构：Applied AI Institute Moscow（莫斯科应用人工智能研究所）

作者：Rostislav Gusev, Alexey Zaytsev

英文摘要：Benchmarks of machine learning models often include many datasets, making evaluation expensive. For efficiency, it is preferable to perform evaluations on small, representative datasets instead. The selection of such subsets typically relies on heuristics and is rarely analyzed for the robustness of the resulting model rankings. We introduce a framework to perform the task of selecting datasets subsets with an evaluation of how different selection strategies preserve the global model rankings. Our framework includes bootstrap aggregation, which provides valid confidence intervals, allowing a principled comparison of selection strategies. We consider clustering, design criteria (A/D-optimality), random baselines, and greedy farthest-first (FAFI). For the latter, we derive upper bounds on selection quality in terms of ranking errors as a function of the number of selected datasets. Empirically, in time series classification (TSC, 112 datasets) and in a supplementary natural language processing benchmark derived from MTEB (57 tasks), several selection strategies improve rank preservation compared with random subsets, including simple FAFI. In contrast, in recommender systems (30 datasets), the improvement of strategies over random selection is small and typically statistically insignificant. For TSC, our best-performing strategy achieves a Spearman correlation of 0.95 with the full benchmark model rankings using only five selected datasets. Additional experiments indicate that the effectiveness of selection approaches depends on both the quality of dataset representations and the scale of the benchmarking regime.

AI 总结：提出COCOLogic-V2数据集，通过正样本、近边界和远边界负样本分类，评估模型在真实图像上的视觉归纳推理能力，发现模型在近边界样本上表现差，视觉归纳推理仍是开放挑战。

链接：https://arxiv.org/abs/2606.28194

机构：Computer Science Department, TU Darmstadt（达姆施塔特工业大学计算机科学系）； Hessian Center for AI (hessian.AI)（黑森人工智能中心）； German Research Center for AI (DFKI)（德国人工智能研究中心）； Max Planck Institute for Informatics, SIC（马克斯·普朗克信息学研究所SIC）； RTG Neuroexplicit Models（RTG神经显式模型）

作者：David Steinmann, Antonia Wüst, Kristian Kersting, Wolfgang Stammer

英文摘要：While interpretable models such as concept bottleneck models (CBMs) and program synthesis methods enable verification of model decisions, their evaluation is typically limited to simple tasks, leaving complex reasoning on real-world images largely unexplored. We introduce COCOLogic-V2, an object-centric dataset for visual inductive reasoning on real-world images covering a broad subset of first-order logic. By categorizing samples into positive variants, near-boundary (NB), and far-from-boundary (FB) negatives, COCOLogic-V2 enables fine-grained diagnosis of model accountability. Our evaluations show that models tend to separate positive and FB samples well but fail on NB samples, while perceptual noise and large rule-induced search spaces pose additional challenges in few-shot settings. Together, these results highlight that visual inductive reasoning remains an open challenge and COCOLogic-V2 provides a concrete foundation for advancing methods in this direction.

AI 总结：针对概率时序图中预测误差与模型不确定性混淆的问题，提出因果框架生成具有瞬态边和真实因果结构的时序图，推导Cramér-Rao界验证参数估计误差与不可约预测损失之间的权衡。

链接：https://arxiv.org/abs/2606.28225

作者：Aniq Ur Rahman

英文摘要：Temporal link prediction is usually evaluated by predictive performance on unseen edges, but in probabilistic temporal graphs this criterion can conflate model error with irreducible uncertainty. We study this issue by characterising an inherent estimation--prediction tradeoff in binary logistic models where regimes that maximise Fisher information and improve parameter recoverability are also those with the highest entropy, making individual predictions intrinsically harder even under perfect parameter recovery. We propose a probabilistic causal framework for generating temporal graphs with transient edges and known ground-truth causal structure, allowing temporal link prediction to be evaluated jointly with causal parameter recovery. For the proposed binary logistic parametrisation, we derive the Cramér--Rao bound and validate the tradeoff between parameter estimation error and irreducible predictive loss. Our results show that predictive accuracy alone may not reflect whether a model has learned the underlying causal mechanism, motivating benchmarks that distinguish reducible model error from intrinsic process uncertainty.

AI 总结：提出四层AI-人类协作分类法，并介绍论文助手工具（PAT），通过推理缩放技术实现深度科学评审，在SPOT基准上数学错误召回率提升34%，并在STOC和ICML会议中成功识别关键错误。

链接：https://arxiv.org/abs/2606.28277

机构：Google Research USA（谷歌美国研究）

作者：Rajesh Jayaram, Drew Tyler, David Woodruff, Corinna Cortes, Yossi Matias, Vahab Mirrokni, Vincent Cohen-Addad

英文摘要： Artificial intelligence is driving a revolution in scientific discovery, accelerating everything from hypothesis generation to mathematical theorem proving. However, this rapid acceleration is creating a systemic challenge: traditional human peer review cannot scale to match the influx of AI-assisted science. Ultimately, to resolve this tension, we must also deploy AI to accelerate the verification and review process itself. To frame the discussion around this transition, we propose a taxonomy consisting of four progressive levels of AI-human collaboration in scientific evaluation, and discuss various trade-offs involved with each. As a step toward this future, we introduce the Paper Assistant Tool (PAT), an agentic AI framework built for deep scientific review and verification. PAT ingests full scientific manuscripts and produces a comprehensive evaluation, checking theoretical results, validating experiments, suggesting improvements, and identifying potential flaws. By utilizing inference scaling techniques, PAT is able to identify deeper issues than a single model call alone, achieving a 34% improvement over zero-shot recall on mathematical errors in the SPOT benchmark. Pilot deployments of PAT as a pre-submission tool for authors at two major Computer Science conferences -- STOC and ICML -- demonstrate its ability to identify critical errors and suggest substantive improvements to research papers. By catching errors early, PAT eases the cognitive burden placed on referees, while preserving their control over the outcomes of the review process.

AI 总结：针对城市交通中队列溢出导致级联锁死的问题，提出OverFlowLight框架，利用多模态传感实时检测溢出并动态插入专用相位，结合规则与强化学习混合控制，实际部署43个路口减少60.4%溢出事件并提升18.2%网络吞吐量。

链接：https://arxiv.org/abs/2606.27381

机构：Beijing University of Posts and Telecommunications（北京邮电大学）； University of Electronic Science and Technology of China（电子科技大学）； South China University of Technology（华南理工大学）； Beijing Xiaocheng Intelligent Computing（北京小城智能计算）； Tsinghua University（清华大学）； Lanzhou University（兰州大学）

作者：Mingyuan Li, Boyang Huang, Tianqi Jiang, Chenpu Li, Chunyu Liu, Yang Li, Ruimin Li, Qiang Wu

英文摘要：Queue overflow, a severe consequence of urban traffic congestion, occurs when vehicle queues exceed intersection capacity, obstructing upstream traffic and triggering cascading gridlocks. Prevailing traffic signal control (TSC) algorithms, primarily optimized for throughput, often fail to address overflow during peak hours, exacerbating congestion and creating safety hazards. We propose OverFlowLight, a real-time framework designed to preemptively resolve overflow and enhance overall TSC performance. It first introduces a mechanism to accurately detect overflow in real-time by leveraging multi-modal sensing from cameras and radars. Upon detection, it dynamically generates and inserts dedicated overflow phases into the signal cycle to clear the blocking queues. This is orchestrated by a hybrid control design that combines rapid rule-based overflow intervention with controller back ends such as reinforcement learning (RL) for longer-horizon efficiency. We conducted extensive real-world deployments of OverFlowLight across 43 intersections in three major cities. The framework demonstrates seamless integration with existing RL-based TSC agents, highlighting its modularity and practical applicability. Empirical results show that OverFlowLight reduces overflow incidents by 60.4% and increases network throughput by 18.2% compared to deployed baselines. Furthermore, it substantially diminishes the need for manual intervention common with expert-tuned signal plans. This work presents the first practical, scalable, and data-driven framework for actively preventing traffic gridlock, offering a crucial component for building resilient and efficient urban transportation systems. Our demonstration videos, codes and datasets are available at the anonymous URL, this https URL.

AI 总结：针对不同长宽比的二维平环面上的三次非线性薛定谔方程，提出几何条件傅里叶神经算子，学习一步解算子，并捕获有理与无理环面上不同的Sobolev范数行为。

链接：https://arxiv.org/abs/2606.27459

机构：Oregon State University（俄勒冈州立大学）

作者：Emmanuel E. Oguadimma, Victory C. Obieke, Xueying Yu

英文摘要：We consider the cubic nonlinear Schrödinger (NLS) equation on two-dimensional flat tori with varying aspect ratios. In this formulation, the choice of aspect ratio governs the Fourier resonance structure, so rational and irrational geometries can exhibit different high-frequency cascade behaviors. We present a geometry-conditioned Fourier neural operator (FNO) for the cubic defocusing NLS equation, where the input consists of the real and imaginary parts of the solution together with the aspect-ratio parameter \(\omega^2\). The model is trained to approximate the one-step solution operator and is evaluated on unseen trajectories generated from random-phase initial data using Fourier pseudospectral method. Our numerical experiments show that the learned operator captures the main solution dynamics on both tori and reproduces the distinct Sobolev norm behavior of the two geometries, with stronger \(H^2\)-growth on the rational torus and more constrained behavior on the irrational torus, consistent with the findings of \cite{hrabski2021energy}. We perform ablation studies to examine the roles of retained Fourier modes, activation functions, Fourier-layer depth, and explicit geometry conditioning. The results indicate that including $\omega^2$ improves long-time predictive accuracy, especially for the rational geometry, and supports the use of geometry-aware neural operators for learning spectral-transfer phenomena in nonlinear dispersive partial differential equations.

AI 总结：研究通过对比十种截断域边界处理方法与全域参考模拟，评估边界条件保真度对井底压力和CO2羽流预测的影响，发现保留角点孔隙体积是关键，而透射率修正并非普遍有益，渐进修正结合透射率修正效果最佳。

链接：https://arxiv.org/abs/2606.27515

机构：Department of Earth and Planetary Sciences, Jackson School of Geosciences, The University of Texas at Austin（德克萨斯大学奥斯汀分校杰克逊地球科学学院地球与行星科学系）； Bureau of Economic Geology, The University of Texas at Austin（德克萨斯大学奥斯汀分校经济地质局）； Hildebrand Department of Petroleum and Geosystems Engineering, The University of Texas at Austin（德克萨斯大学奥斯汀分校希尔德布兰德石油与地质系统工程系）

作者：Romal Ramadhan, Seyyed A. Hosseini, Larry W. Lake

英文摘要： Accurate prediction of bottom-hole pressure (BHP) and CO2 plume migration is essential for safe geological carbon storage, yet practical simulations often rely on truncated domains where artificial boundaries distort pressure diffusion and CO2 saturation footprints. In this study, we evaluate how boundary-condition fidelity affects BHP and CO2 plume prediction by comparing ten reduced-domain boundary treatments against full-domain reference simulations in homogeneous and heterogeneous reservoirs. We test uniform pore-volume multipliers, transmissibility modifiers, corner-adjusted pore-volume corrections, layered corrections, and gradual modifiers using BHP RMSE, NRMSE, peak pressure deviation, and plume Intersection over Union (IoU) as performance metrics. Our results show that conserving corner pore volume is the most important requirement for truncated-domain modeling. We find that uniform treatments which neglect corner storage generate large pressure errors, with BHP RMSE of 362 to 382 psi in the homogeneous model and 250 to 304 psi in the heterogeneous model, and yield plume IoU values near 0.80 to 0.84, indicating roughly 16 to 20% of the combined plume area is misrepresented. Corner-adjusted scenarios substantially reduce pressure errors and raise plume IoU above 0.94, but we observe that transmissibility correction is not universally beneficial. In homogeneous reservoirs, uniform transmissibility adjustment improves pressure fidelity; in heterogeneous reservoirs, it can over-restrict flow across variable-permeability boundary faces, increasing BHP error and contracting the predicted plume. We find the gradual modifier with transmissibility correction provides the most consistent performance, achieving BHP NRMSE below 3.7% and plume IoU above 0.97 in both reservoir types.

AI 总结：提出HIA-GAT，一种双流异构图注意力网络，通过专用注意力路径处理纵向和横向交互，并利用冲突类型感知门控机制融合，在NGSIM数据集上实现最佳风险排名性能（AUC最高0.867），并提供可解释的车辆级冲突归因。

链接：https://arxiv.org/abs/2606.27577

机构：DENSO International America, Inc.（电装国际美国公司）

作者：Mahshid Malazizi, Seyedmehdi Khaleghian, Mina Sartipi, Toru Hirano, Yunfei Xu, Hoang H. Nguyen

英文摘要：This paper formulates frame-level freeway risk assessment as a multi-agent scene graph-level binary classification problem, where each video or trajectory frame is labeled risky if any TTC- or PET-based conflict violates a specified severity threshold. We construct a relation-aware graph per frame with vehicles as nodes and two interaction types as edges: same-lane (longitudinal) and adjacent-lane (lateral), augmented with physics-informed edge features aligned to rear-end and lane-change conflict mechanisms. Building on a structured benchmarking suite of non-graph models and graph baselines, we propose HIA-GAT, a dual-stream heterogeneous graph attention network that processes longitudinal and lateral interactions through dedicated attention pathways and fuses them via a conflict-type-aware gating mechanism with event-level gate supervision derived from SSM conflict attribution. Experiments on the NGSIM I-80 and US-101 freeway datasets across nine TTC and PET threshold configurations show that HIA-GAT achieves the best average risk-ranking performance (AUC 0.835 on I-80 and 0.867 on US-101), with the largest gains on PET-only (lane-change) settings where relational structure is essential. Beyond accuracy, the learned gate provides interpretable per-vehicle attribution of dominant conflict type, supporting actionable, real-time freeway safety monitoring. We show that graph structure is critical for modeling lateral conflict risk, while longitudinal risk can often be captured by non-relational aggregation.

AI 总结：提出COOPA模块化LLM智能体架构，通过迭代置信度建模、元素级溯源与多求解器路由，提升运筹问题建模准确性与可解释性，在多个基准上优于基线方法。

链接：https://arxiv.org/abs/2606.27611

机构：Tsinghua University（清华大学）； Duke University（杜克大学）； Yale University（耶鲁大学）

作者：Chuanhao Li, Xiaoan Xu, Dirk Bergemann, Ethan X. Fang, Yehua Wei, Zhuoran Yang

英文摘要：Operations Research (OR) provides a rigorous framework for high-stakes decision-making, but effective OR modeling requires substantial domain knowledge, mathematical abstraction, and solver expertise. Recent LLM-based systems automate parts of this pipeline, yet remain limited by low accuracy on complex problems, opaque outputs, and narrow solver support. We propose COOPA (COoperative OPerations Agent), a modular LLM-agent architecture for interpretable and scalable OR decision support. It combines three components: iterative confidence-based modeling, which generates multiple candidate formulations, self-evaluates them across modeling dimensions, and selects one using a max-min confidence criterion; element-level provenance and confidence explanations, which link variables, parameters, constraints, and objectives to quoted source text and provide an audit trail for human verification; and multi-solver routing to specialized optimizer agents for different OR problem classes. Across three OR benchmarks, eight LLM backbones, and four baselines under identical conditions, COOPA achieves the best macro-average accuracy on six of eight backbones and improves over the strongest baseline by up to 6.7 percentage points. A within-system ablation isolates the contribution of iterative confidence-based modeling, while additional analyses and case studies illustrate the value of source traceability and multi-solver dispatch.

AI 总结：研究多期波动率预测中，通过改变推理时的展开规则来提升预测性能，发现验证集选择的部署策略能有效改善MIMO模型性能，且策略选择对评估指标敏感。

链接：https://arxiv.org/abs/2606.27688

机构：The University of Bristol（布里斯托大学）

作者：Riku Green, Zahraa S. Abdallah, Telmo M Silva Filho

英文摘要： In financial forecasting, predictive performance depends not only on which model is trained, but also on how the trained model is deployed. We study this issue in multi-horizon volatility forecasting. Our starting point is that a trained multi-output (MIMO) forecaster does not define a single deployable predictor: by changing the inference-time rollout rule, the same trained model induces a family of forecasts with different accuracy and cost profiles. Across 20 stock-volatility series, three forecast horizons, and architectures ranging from linear models to PatchTST, we find that non-default rollout rules often improve over standard MIMO deployment. However, the best fixed rule varies substantially across architectures and horizons, making any single static replacement unreliable. We therefore evaluate validation-based deployment policies over the induced rule family. Under the primary MSE objective, validation-selected singletons provide a low-cost improvement over default MIMO, while small rule subsets recover much of the benefit of larger ensembles at substantially lower inference cost. We also find that policy rankings are metric-sensitive: MSE-selected policies do not transfer uniformly to QLIKE, a finance-standard volatility loss. These results show that inference-time deployment is a meaningful source of adaptiveness in financial forecasting, and that trained volatility forecasters should be evaluated not only by their architecture, but also by their deployment policy.

AI 总结：提出PerturbCellRL框架，通过强化学习后训练预训练的单细胞转录组生成器，利用细胞级验证器奖励确保生物一致性，在多个基准上提升预测性能。

链接：https://arxiv.org/abs/2606.27752

机构：Stanford University（斯坦福大学）； Peking University（北京大学）

作者：Dongxia Wu, Mingyu Li, Yuhui Zhang, Anurendra Kumar, Emma Lundberg, Serena Yeung-Levy, Emily B. Fox

英文摘要：Single-cell perturbation models can reduce costly wet-lab screening by predicting how cells respond transcriptionally to interventions. While recent generative models improve population-level prediction, individual generated cells are not explicitly checked for biological consistency. We introduce PerturbCellRL, a reinforcement learning (RL) framework that post-trains a pretrained single-cell transcriptomic generator using a suite of cell-level verifiers as rewards. These verifiers define four rewards: Pearson top-k similarity, RMSE top-k proximity, DE Spearman, and Pathway activity. The Pathway activity verifier rewards cells whose pathway responses match known perturbation biology. We evaluate PerturbCellRL on multiple genetic and chemical perturbation benchmarks. Across these benchmarks, PerturbCellRL improves over the pretrained flow-matching generator on reward-aligned evaluation metrics and a held-out evaluation metric. Moreover, PerturbCellRL remains competitive with state-of-the-art methods on population-level metrics. Together, these results frame trustworthy single-cell prediction as verifier-guided generative alignment, moving beyond matching expression distributions toward predictions whose single-cell perturbation effects are explicitly checked for biological consistency.

AI 总结：研究利用ReLU激活统计的记忆化指标，在低样本sEMG校准中无需验证集即可早期检测过拟合，实验表明测试精度下降伴随激活率特征变化。

链接：https://arxiv.org/abs/2606.27855

机构：German Aerospace Center (DLR), Institute for AI Safety and Security（德国航空航天中心（DLR），人工智能安全与安全研究所）； Faculty of Electrical Engineering and Information Technology, Ruhr University Bochum（波鸿鲁尔大学电气工程与信息技术学院）； Institute for Neural Computation, Ruhr University Bochum（波鸿鲁尔大学神经计算研究所）； Institute of Computer Science, Ruhr West University of Applied Science（鲁尔西部应用科学大学计算机科学研究所）

作者：Stephan J. Lehmler, Tobias Glasmachers, Ioannis Iossifidis

英文摘要：Deep learning models for surface electromyography (sEMG) can benefit substantially from subject-specific (re-)calibration, since no sufficiently large and diverse datasets are available to train fully generic decoders. However, for user acceptance, the number of repetitions that can realistically be collected during calibration is severely limited, which increases the risk of overfitting and, in extreme cases, can even degrade performance compared to the uncalibrated model. Classical overfitting indicators such as validation performance and regularization with early stopping are difficult to apply in this low-sample regime, as they require additional held-out data that is rarely available in practical calibration scenarios. In this work, we investigate a recently proposed class of memorization indicators based solely on the activation statistics of rectified linear units (ReLU) in deep neural networks, which can be computed directly from training data without any extra validation set. We conduct a transferlearning experiment on a benchmark sEMG dataset, where a convolutional neural network is first pre-trained on multiple subjects and subsequently fine-tuned on individual users using only a small number of repetitions. During calibration, we monitor both decoding performance and the activation behaviour of the last hidden layer. Our results provide first evidence that decreases in test accuracy during fine-tuning are ac companied by characteristic changes in activation rates, indicating that activation-based memorization indicators are a promising tool for early spotting of unsuccessful learning in low-sample sEMG calibration settings.

AI 总结：提出两阶段微调方法，先领域自适应微调，再强化学习迭代微调，以生成符合目标氨基酸组成且保持序列质量的蛋白质序列。

链接：https://arxiv.org/abs/2606.27939

作者：Violeta Basten-Romero, Rubén Muñoz-Tafalla, Anna María Díaz-Rovira, Bertran Miquel-Oliver, Isaac Filella-Merce, Víctor Guallar

英文摘要：Protein language models are standard priors for biological sequence generation, but steering them toward explicit distributional design targets remains largely unexplored. We study a constrained protein generation problem in which sequences must match a desired amino-acid (AA) composition profile while preserving plausible sequence statistics and diversity. The motivating application is synthetic feed protein design, where the AA composition of dietary proteins directly determines their nutritional value. We propose a two-stage pipeline in which domain-adaptive fine-tuning (FT) on an in-domain protein dataset is followed by iterative reward-weighted FT via reinforcement learning (RL) anchored against the FT model as a frozen reference. We evaluate the pipeline on two AA compositions and find that FT brings the average composition close to the target, while the subsequent RL enforces specific sequence constraints that FT alone cannot satisfy. We additionally evaluate the design choices of the proposed composition reward term against two baselines and an ablated variant, isolate the contribution of each training stage, and verify that AA composition alignment is achieved without degrading sequence quality.

AI 总结：本文评估五种降维模型（含自编码器变体）将九维传感器数据压缩为单维潜在分数，通过重构误差与可解释性复合指标选择最优模型，发现深层自编码器表现最佳，跑步配速、有氧解耦和平均心率为关键驱动因素。

链接：https://arxiv.org/abs/2606.28145

作者：Mateusz Kubita, Jan Zubalewicz, Krzysztof Siwek

英文摘要：Wearable devices produce large, high dimensional training logs for everyday runners, and interpretation rather than data collection is now the limiting step. This paper evaluates five dimensionality reduction models, three autoencoder variants, PCA, and a Variational Autoencoder, on their ability to compress nine sensor runner profiles into a single scalar performance indicator, the latent score. Because the setting is fully unsupervised, model quality is assessed along two complementary axes: reconstruction error (Mean Squared Error) and latent score interpretability, measured via Spearman and Kendall rank correlations, Mutual Information, and Permutation Importance. These are combined into a composite selection criterion that prevents selecting models on reconstruction accuracy alone. Feature rankings from the four metrics are aggregated via a modified Borda count, and their stability is confirmed by bootstrap validation. A two feature linear baseline is included to anchor the comparison. Deep autoencoder achieved the lowest reconstruction error and the highest composite score. Once the PCA hidden layers were widened, the deeper variants became closely competitive with Deep AE on the composite criterion, indicating that the limiting factor was hidden layer capacity rather than the one dimensional bottleneck. Running pace, aerobic decoupling, and average heart rate emerged as the dominant latent score drivers across all models and resampling runs, consistent with established physiology.

AI 总结：提出基于多尺度边界激励和傅里叶特征编码的物理信息神经网络框架，从有限边界数据重建电导率，相对误差约3%-12%，并分析坐标表示和激励设计的影响。

链接：https://arxiv.org/abs/2606.28158

机构：Institut de Ciències del Cosmos (ICCUB), Universitat de Barcelona（巴塞罗那大学宇宙科学研究所）； Departament de Física Quàntica i Astrofísica, Universitat de Barcelona（巴塞罗那大学量子物理与天体物理学系）； ICREA； School of Mathematics, University of Minnesota（明尼苏达大学数学学院）

作者：Ali AlHadi Kalout, Pablo Tejerina-Pérez, Konstantin Karchev, Pedro Tarancón-Álvarez, Leonid Sarieddine, Raul Jimenez, Max Engelstein, Guy David

英文摘要：Physics-informed neural networks (PINNs) have recently emerged as a promising framework for addressing the Calderón inverse problem from limited boundary data. In this work, we revisit neural Calderón inversion by introducing multiscale boundary excitations based on randomized wavelet functions and investigating the role of Fourier-feature encoding (FFE) for representing sharp conductivity variations. We propose a physics-informed reconstruction framework that represents the unknown conductivity and the associated family of electric potentials with separate neural networks conditioned on the applied boundary excitations. The governing elliptic PDE is enforced through physics-informed residuals, while finite Dirichlet-to-Neumann (DtN) data are incorporated through boundary losses. Using synthetic data from a finite-difference forward solver, we evaluate the method on conductivity fields with inclusions, sharp interfaces, smooth profiles, and heterogeneous media. Results show that the framework recovers dominant conductivity structures from finite boundary measurements with relative errors between $3\%-12\%$ approximately. We show that FFE improves the reconstruction of localized sharp features, particularly for inclusions and interfaces, but are not universally optimal, with raw-coordinate networks performing competitively for smoother fields. These results highlight coordinate representations and boundary excitation design as key factors in neural Calderón inversion.

AI 总结：提出CPAgents框架，通过多智能体协作自动构建可解释的复合表型（如多项式、比值和交互形式），在群体规模心脏影像队列中显著提升疾病判别性能。

链接：https://arxiv.org/abs/2606.28179

机构：Department of Mechanical Engineering, University College London（伦敦大学学院机械工程系）； CSIG Group, Tencent（腾讯CSIG组）； Department of Computing, Imperial College London（帝国理工学院计算系）； Department of Brain Sciences, Imperial College London（帝国理工学院脑科学系）； Data Science Institute, Imperial College London（帝国理工学院数据科学研究所）； FAU Erlangen–Nürnberg（埃尔朗根-纽伦堡大学）； UK Dementia Research Institute, Imperial College London（英国痴呆症研究所帝国理工学院）； Rosalind Franklin Institute（罗莎琳德·富兰克林研究所）

作者：Zuoou Li, Wenlong Zhao, Kelly Yu, Weitong Zhang, Paul M. Matthews, Wenjia Bai, Bernhard Kainz, Mengyun Qiao

英文摘要：Identifying robust associations between cardiac imaging phenotypes and clinical diseases is fundamental to population-scale cardiovascular research and reliable risk stratification. However, current phenome-wide association studies rely on pre-defined, single-variable phenotypes or expert-crafted features, which limits their ability to capture clinically meaningful non-linear effects and cross-phenotype interactions. To address this, we propose CPAgents, an iterative phenotype-Composition framework for cardiovascular Phenome-wide association study (PheWAS) that automatically constructs and validates interpretable composite phenotypes (e.g., polynomial, ratio, and interaction forms) from base imaging features. Specifically, our system coordinates three agents: (i) an Analyst that identifies statistical pathologies and nominates candidate transformations; (ii) a Proposer that generates constrained, medically and statistically motivated expressions under numerical safety rules; and (iii) a Verifier that evaluates candidates using multi-stage criteria and produces transparent evidence trails for accepted phenotypes. Evaluated on a population-scale cardiac imaging cohort, the discovered composite phenotypes markedly improve disease discrimination: across 72 classifier-disease-metric combinations, our variants achieve the top rank in 56 cases versus 18 for baselines, with gains observed across all nine clinical disease categories. Our framework yields compact, clinically interpretable phenotype formulas with transparent evidence trails, enabling scalable discovery of stronger phenotype-disease associations beyond expert-driven feature selection.

AI 总结：基于32年数据，通过时间序列建模和机器学习分析斯里兰卡汇款，发现汇率和油价是主要驱动因素，Ridge回归比SARIMA精度提升73.8%，预测2026年汇款90.01亿美元。

链接：https://arxiv.org/abs/2606.28190

作者：Dhinanjaya Fernando, Dinura Ginige, Kalana Lakshan, Chanupa Gurusinghe, Lasana Pahanga, Subavarshana Arumugam, Sandeepa Weerasekara, Sandareka Wickramanayake, Nisansa de Silva

英文摘要： This study analyzes Sri Lankan migration and remittances over 32 years (1994-2025). Using a 384-month harmonized dataset, we apply exploratory data analysis, stationarity corrected time-series modeling (ADF, Johansen, VAR/VECM), and supervised learning. Results reveal remittance inflows are primarily driven by external macroeconomic variables, specifically exchange rate dynamics and global oil prices, rather than domestic indicators. Impulse response analysis confirms the asymmetric impact of currency depreciation and oil price shocks. Predictively, multivariate machine learning models outperform traditional univariate approaches; Ridge Regression achieves a 73.8% accuracy improvement over SARIMA (Annualized RMSE: USD 494.8 Mn). The optimized framework projects 2026 remittances at USD 9,001 million under stable conditions. These findings highlight the structural dependence of remittances on global economies, emphasizing the need for robust exchange rate policies, skilled migration, and formal financial channels to enhance long-term economic resilience.

AI 总结：提出一种结合物理信息神经网络与迁移学习的框架，用于锂离子电池单粒子模型（含电解质）的状态估计，通过预训练和微调实现快速收敛与准确电压预测。

链接：https://arxiv.org/abs/2606.28220

机构：Texas Tech University（德克萨斯理工大学）

作者：Gift Modekwe, Qiugang Lu

英文摘要：Physics-informed neural networks (PINNs) have emerged as a powerful tool for solving nonlinear partial differential equations (PDEs), including battery electrochemical models. They typically en-force conservation laws within the loss function to ensure physically consistent solutions. Tradi-tional numerical methods such as finite difference, finite volume, and finite element techniques, re-ly on discretization and can be computationally expensive for nonlinear systems. To address this challenge, PINNs offer improved scalability, particularly for reduced-order models like the single particle model with electrolyte (SPMe). The SPMe describes lithium-ion battery dynamics through coupled diffusion, transport, reaction kinetics, and voltage equations. Despite these advantages, training SPMe-based PINNs from scratch for different battery chemistries or operating conditions is demanding and often leads to slow convergence. To overcome this limitation, this work introduces a transfer learning framework for SPMe-PINNs. The model is first pretrained to learn general elec-trochemical dynamics and then adapted to a target battery by transferring weights, freezing se-lected layers, and fine tuning the remaining parameters, including estimating key electrochemical variables. Validation using PyBaMM demonstrates accurate voltage prediction, indicating that the proposed approach preserves electrochemical consistency while reducing training time and ena-bling efficient generalization across batteries.

AI 总结：本文从因果中介分析重新推导激活补丁估计量，发现自然间接效应包含交互效应，导致组件因果重要性评估偏差，并提出交互效应作为可解释性诊断工具。

链接：https://arxiv.org/abs/2606.27510

机构：University of Massachusetts Amherst（马萨诸塞大学阿默斯特分校）； Adobe Research（Adobe研究院）； Boston University（波士顿大学）

作者：Sankaran Vaidyanathan, David Arbour, Aaron Mueller, Scott Niekum, David Jensen

英文摘要：Activation patching is the primary tool in mechanistic interpretability. It attributes causal responsibility for a model behavior to each of its individual components by estimating its natural indirect effect (NIE). Re-deriving the activation patching estimand from causal mediation analysis, we find that the NIE does not solely capture the causal effect through the specific component. It also contains interaction effects (INT) that measure how much the component's causal effect itself depends on the state of other components in the model. A natural response may be to try to eliminate INT by adjusting the estimator or unit of analysis, but each of these potential remedies has predictable failure modes. We demonstrate these failure modes in the GPT-2 IOI circuit; components whose causal importance is conditional on the state of other components are either invisible or artificially inflated, and INT variance explains the previously documented instability of faithfulness scores. We prove that INT scales with the distance between clean and patched component activations, is negligible when the model is locally affine, and decomposes combinatorially into pairwise and higher-order group interactions. Despite its inevitability, INT is not a nuisance to be eliminated, but rather a diagnostic for interpretability studies. Its individual and group-level magnitude and sign signal when causal conclusions are prompt-dependent, and when greedy NIE-based component ranking will miss mechanisms only discoverable through combinatorial search.

AI 总结：针对概率化学反应用网络规模过大的问题，通过恢复因子图结构并迁移约简结果，在保持信念传播不动点的前提下显著缩小网络。

链接：https://arxiv.org/abs/2606.27737

作者：Mauricio Montes, Gregoire Sergeant-Perthuis

英文摘要：Programming adaptive behaviors at the cellular level is a long-standing goal that raises the question of how probabilistic computation can be implemented in biochemical systems. Chemical reaction networks (CRNs) provide such a substrate and have been shown to realize probabilistic models, including hidden Markov models and factor graphs, with dynamics reproducing Bayesian inference and belief propagation. However, encoding these algorithms typically requires prohibitively large reaction networks, and classical CRN reduction techniques do not directly apply. By recovering the factor graph structure encoded in Napp--Adams-compiled CRNs, we transport recent factor-graph reduction results to their chemical implementations, obtaining significantly smaller CRNs while preserving the belief-propagation fixed points on surviving variables.

AI 总结：提出Democratic ICAI方法，通过结构化角色辩论收集多元理由，从偏好中提取更全面的指导原则，提升偏好预测准确性。

链接：https://arxiv.org/abs/2606.28294

机构：TCS Research（TCS研究）

作者：Kevin Kingslin, Anish Natekar, Ashutosh Ranjan, Vivek Srivastava, Savita Bhat, Shirish Karande

英文摘要：Preference-based alignment often struggles to capture the reasoning that underlies human judgments. Many evaluations rely on multiple interacting criteria, yet pairwise labels reveal only the final choice rather than the considerations that shape preferences. Inverse Constitutional AI (ICAI) improves interpretability in decision making by summarizing preferences into natural-language principles, but its single-pass explanations miss much of the nuance involved in complex decisions. We introduce Democratic ICAI, a novel approach that gathers multiple competing rationales through structured persona debate, offering a broader and more expressive account of the factors influencing each comparison. From these richer signals, we derive clearer and more comprehensive steering principles and use them to guide decision modeling through both LLM-based and decision-tree judges. Experiments on creative preference benchmarks, MuCE-Pref and LiTBench, across multiple creative task categories show that Democratic ICAI yields a more faithful preference structure. It improves average preference prediction across tasks relative to deliberative prompting and principle-based baselines, while producing constitutions that LLM annotators prefer.