机器学习学术速递[8.25]

AI 总结：该研究探究学习轨迹上费希尔宽度的演化，推导相关分解与界，在特定模型中证明总体梯度流选择的分支满足极限关系，实验验证分支式损失参数化费希尔宽度的合理性。

链接：https://arxiv.org/abs/2608.21561

机构：FPT University（FPT大学）

作者：Vu Khac Ky

英文摘要：Fisher width measures the Gaussian width of a probe set after deformation by the local Fisher geometry. We study its evolution along learning trajectories and ask when training loss can serve as an effective coordinate for this quantity. We first derive an exact trace--shape factorization and a deterministic stability bound for fixed compact probes. In a population Gaussian-teacher logistic model, the teacher-aligned state is extremal on every loss level below $\log 2$: it has minimal parameter norm and maximizes both Fisher trace and Euclidean-ball Fisher width. We then show that population gradient flow asymptotically selects this branch, with explicit rates for the aligned and orthogonal coordinates. This yields, for $d\geq2$, \[ \frac{w_F(B_2^d;\theta(t))} {\sqrt{L(\theta(t))}} \longrightarrow \frac{\sqrt6}{\pi}\mathbb E[\chi_{d-1}]. \] Controlled full-Fisher experiments support the matched-loss branch and the population predictions. In a nonlinear MLP with a diagonal model-Fisher approximation, GD and SGD remain close at matched loss, whereas Adam follows a substantially displaced branch; the fixed probes tested retain highly similar temporal shapes. These results support a branchwise, rather than universal, loss parametrization of Fisher width.

AI 总结：本文针对微控制器上的TinyML系统开展性能与功耗表征，提出估算抽象层成本的模型，其结果可辅助边缘设备的NAS与CNN推理优化。

链接：https://arxiv.org/abs/2608.21646

机构：School of Computing, National University of Singapore（新加坡国立大学计算机学院）

作者：Yujie Zhang, Dhananjaya Wijerathne, Zhaoying Li, Tulika Mitra

英文摘要：TinyML systems are enabling machine learning (ML) inference at the edge. However, there is little quantitative analysis of such systems. This paper presents a systematic performance and power characterization of diverse TinyML applications on microcontrollers (MCUs), spanning neural network models, software libraries, operating systems, and hardware architectures. We focus on the impact of the multiple layers of abstraction that provide higher programmability at the expense of performance and energy efficiency. We propose a model to estimate the costs of different abstraction layers and make recommendations for minimizing those costs. Our findings can help designers with Neural Architecture Search (NAS) and CNN inference optimization on edge devices.

AI 总结：本研究提出MoEXBench基准，系统评估混合专家LLM的可组合压缩技术，发现压缩方法间存在非平凡相互作用，为MoE模型部署提供实用的准确率-内存-延迟比较方案。

链接：https://arxiv.org/abs/2608.21693

机构：University of Virginia（弗吉尼亚大学）； Sony AI（索尼人工智能）

作者：Afsara Benazir, Chen Chen, Rongxiao Qu, Jiabo Huang, Jingtao Li, Lingjuan Lyu

英文摘要：Mixture-of-Experts (MoE) LLMs scale model capacity efficiently through sparse activation, but their large expert parameter footprint, routing imbalance, and long-context KV-cache growth make deployment difficult on commodity hardware. Practical deployment often requires stacking multiple compression techniques: expert pruning removes redundant experts, weight quantization lowers model memory footprint, and KV-cache compression reduces long-context memory pressure. However, these techniques are typically evaluated in isolation, leaving open how they interact when applied together in realistic deployment pipelines. In this work, we present MoEXBench, a systematic benchmark for evaluating composable MoE compression as an end-to-end deployment workflow. MoEXBench studies 10 MoE models ranging from 30B to 235B total parameters across standard-attention, hybrid linear-attention, and sliding window attention architectures. It evaluates 20%-50% expert pruning rates, 1 to 16 bit weight-quantization schemes, and multiple KV-cache precision settings, applied both individually and in combination. MoEXBench introduces an eight-module evaluation suite that jointly measures composable-compression quality, workload and architecture robustness, pruning/quantization/KV cache sensitivity, and deployment efficiency on commodity hardware. Our results reveal non-trivial interactions among compression methods: composable compression cannot be predicted from standalone techniques, compression rate alone does not reliably predict quality loss or runtime gain, expert pruning is the dominant degradation source, and average quality can hide workload and architecture-specific failures. By releasing normalized module scores, compressed artifacts, and reproducible scripts, MoEXBench enables practical accuracy-memory-latency comparison across MoE families and hardware backends.

AI 总结：该研究针对深度多层感知机证明了与原网络宽度无关的可压缩性定理，提出含导数匹配技术与逐层重加权的构造方法，得出压缩宽度量级公式，为神经网络压缩提供理论支撑。

链接：https://arxiv.org/abs/2608.21752

机构：Princeton University（普林斯顿大学）； Peking University（北京大学）； Massachusetts Institute of Technology（麻省理工学院）

作者：Hong-Yi Wang, Mingze Wang, Liu Ziyin

英文摘要：It has long been known that well-trained neural networks can be compressed very strongly without affecting their performance, an important phenomenon that remains poorly understood. We prove a uniform compressibility theorem for deep multilayer perceptrons with analytic activations. For a deep, wide fixed teacher network, there exists a narrow (same depth) network that approximately represents the same function as the original. The reachable compressed width is strikingly independent of the original width, but is $O((\log(1/\varepsilon))^{d_{in}})$, where $\varepsilon$ is the error budget and $d_{in}$ is the effective input dimension. Our construction involves a novel derivative-matching technique which is aware of the low-dimensional input, and a layer-wise reweighting that preserves the input-output mapping.

AI 总结：针对现有医疗AI智能体无法内化临床经验的问题，提出MSM-Mem框架，整合多类型临床经验并逐步更新，经MoE-LLaVA评估可提升性能，助力医疗AI智能体随实践进化推理能力。

链接：https://arxiv.org/abs/2608.21810

机构：University of Georgia（佐治亚大学）； New Jersey Institute of Technology（新泽西理工学院）

作者：Md Asaduzzaman Jabin, Khoa Le, Lin Zhao, Tianming Liu

英文摘要：Clinical decision-making is inherently experience-driven: physicians progressively refine their reasoning by synthesizing patient history, multimodal observations, and prior diagnostic experiences across interactions. In contrast, current multimodal large language model (MLLM)-based medical AI agents largely operate as stateless inference systems, generating decisions independently for each interaction without retaining or internalizing experiential knowledge. This discrepancy limits their ability to progressively improve reasoning reliability through usage and adapt to longitudinal patient contexts in real-world clinical workflows. In this study, we propose Medical Structured Multimodal Memory (MSM-Mem), an agentic memory framework that enables medical AI agents to evolve through accumulated clinical experiences. MSM-Mem organizes heterogeneous clinical experiences into semantic, episodic, and visual memory and incrementally updates them during inference, allowing the agent to retrieve prior experiences to inform current reasoning and progressively refine decision-making over time. Evaluations on MoE-LLaVA backbones demonstrate consistent performance improve- ments with further gains observed through continued usage. In general, MSM-Mem offers a viable pathway toward medical AI agents capable of evolving their reasoning competence in a manner analogous to the way clinicians learn from practice over time.

AI 总结：该研究在合成环境中评估混合专家状态空间模型，发现增加专家数量会导致逆缩放等问题，算子级混合模型未优于单专家基线，需几何感知评估模式切换动力学系统。

链接：https://arxiv.org/abs/2608.21840

机构：Delhi Technological University（德里理工大学）

作者：Chandresh Pandey

英文摘要：Mixture-of-Experts (MoE) architectures are commonly motivated as a way to increase expressivity by decomposing complex systems into simpler local dynamics. This intuition has recently been extended to spectral state-space models, where mixing stable operators is assumed to enable adaptation to heterogeneous or regime-switching time series. We critically evaluate this assumption in a controlled synthetic setting designed to isolate dynamical rather than representational challenges. We study a next-step prediction task on sequences composed of three regimes: chaotic dynamics generated by the Mackey-Glass system, a stable oscillatory regime, and a noise-dominated autoregressive regime. Across extensive ablations including capacity scaling, oracle routing, frozen-expert variants, and comparisons to output-level MoE baselines, operator-level mixture models consistently fail to outperform a single-expert baseline. Increasing the number of experts leads to inverse scaling, routing collapses or fails to induce meaningful specialization, and even perfect regime supervision does not prevent degradation in global performance. Furthermore, we show that apparent improvements in mean squared error on chaotic trajectories can be misleading. Phase-space analysis reveals that lower error often arises from temporal smoothing that destroys the geometry of the underlying attractor rather than from faithful modeling of the dynamics. These results identify a likely limitation of operator interpolation under the studied parameterization and training protocol, and underscore the need for geometry-aware evaluation when assessing regime-switching dynamical systems.

AI 总结：该研究提出Bi-EZP双层框架，通过大语言模型与CMA-ES解耦聚合结构发现和参数校准，在NATS-Bench等数据集上验证了其构建集成零成本代理的有效性

链接：https://arxiv.org/abs/2608.21927

机构：School of Mathematics and Statistics, Guangdong University of Technology（广东工业大学数学与统计学院）

作者：Yutao Lai, Kezhao Lai, Hai-Lin Liu

英文摘要： Zero-cost proxies enable neural architecture search (NAS) to rank candidate networks from statistics computed at initialization, avoiding repeated training. However, different proxies capture different properties and often produce inconsistent rankings across search spaces. Ensemble proxies can combine complementary signals, but automated discovery must optimize both discrete aggregation structures and their continuous coefficients, making structural quality difficult to separate from parameter calibration. We propose Bi-EZP, a bilevel framework that decouples these decisions. At the upper level, a large language model generates executable aggregation programs over four complementary base proxies with program-specific parameter bounds. At the lower level, covariance matrix adaptation evolution strategy (CMA-ES) optimizes the continuous parameters of each fixed program on an inner training split. The calibrated programs are then evaluated using Kendall's rank correlation on a disjoint validation split, enabling evolutionary selection to favor structures that generalize beyond their calibration data. Experiments on NATS-Bench and Network Design Spaces evaluate ranking performance across heterogeneous search spaces, and DARTS experiments assess downstream architecture search. Results show that separating program discovery from numerical calibration provides an effective approach to automated ensemble zero-cost proxy construction. The source code is available at: this https URL

AI 总结：该研究提出符号神经常微分方程框架，通过多步训练结合稀疏正则化，从时间序列数据中学习稳定、可泛化的稀疏可解释动力系统模型，可准确恢复多类系统的动力学特性。

链接：https://arxiv.org/abs/2608.22112

机构：University of California, Santa Barbara（加利福尼亚大学圣巴巴拉分校）

作者：Nibodh Boddupalli, Jeff Moehlis

英文摘要：We present a machine learning framework for identifying sparse, interpretable models of dynamical systems directly from time-series data. Our approach parameterizes the underlying vector field using a neural architecture and trains it by minimizing a multi-step prediction loss over a finite horizon. To ensure numerical tractability, we optimize a mean absolute error objective averaged across prediction steps, and progressively increase the horizon during training. A key feature of this formulation is that it enforces consistency under repeated composition of the learned dynamics. As a result, the identified models exhibit significantly improved stability compared with approaches based on one-step regression of the vector field. When combined with sparsity-promoting regularization, this leads to parsimonious models that generalize beyond the training data. We demonstrate accurate recovery of systems exhibiting a wide range of behaviors, including stable and unstable fixed points, periodic orbits, and chaotic attractors. For chaotic systems, while long-term trajectory prediction is inherently limited by sensitivity to initial conditions, we show that multi-step training yields models with accurate short-term dynamics and strong agreement in long-time statistical properties, including mean, variance, and Lyapunov exponents. Moreover, we establish theoretical bounds linking trajectory error to statistical accuracy, providing a step toward a principled explanation for this behavior.

AI 总结：本文提出DSSM-CRF架构，通过双尺度状态空间模型与说话人感知动态CRF分离对话情感识别的跨说话人影响与说话人内部演化，在IEMOCAP和MELD数据集上取得了优于对照的识别性能。

链接：https://arxiv.org/abs/2608.22399

机构：National Taiwan University of Science and Technology（台湾科技大学）

作者：Guan-Hua Wen, Kuan-Yu Chen, Hou-Chiang Tseng

英文摘要：Conversational speech emotion recognition must reconcile acoustic evidence across temporal scales with two interaction processes: cross-speaker contextual influence and within-speaker emotion evolution. We propose DSSM-CRF, an audio-only architecture that explicitly separates these processes. Bidirectional state-space models encode fused self-supervised speech representations at frame and dialogue scales, so each utterance representation captures local prosody and context from all speakers. The decoder then orders each speaker's utterances into an independent dynamic conditional random field chain. Consecutive utterances in a speaker's chain form a transition pair whose score combines a corpus-level transition matrix with a residual predicted from the two contextualized utterances. An auxiliary objective supervises whether each pair changes emotion but does not participate in Viterbi inference. Thus, interlocutor turns affect contextual emotion scores without being treated as transitions in another speaker's emotion trajectory. DSSM-CRF achieves 75.81% UA and 74.90% WA on IEMOCAP, and 54.72% WA and 49.31% WF1 on MELD. Matched controls demonstrate complementary gains from speaker-wise factorization and CRF modeling.

AI 总结：该研究提出功能兼容性是持续神经学习的因果决定因素，通过受控实验验证其可推广至多类模型与任务，为解决神经网络稳定性-可塑性问题提供了新方向。

链接：https://arxiv.org/abs/2608.22462

机构：Dublin City University（都柏林城市大学）

作者：Hossein Javidnia

英文摘要：Artificial neural networks can acquire new capabilities but often damage existing ones when they continue to learn. This stability-plasticity problem has motivated replay, regularization and constrained-update methods, yet it remains unclear whether a property of incoming learning itself determines what can be retained without disrupting protected behaviour. Here we show that functional compatibility, the extent to which new learning can coexist with behaviour that must be preserved, is a causal determinant of persistent learning. To our knowledge, this is the first controlled causal demonstration in which compatibility is deliberately changed from matched neural states and persistent learning is measured under a common retention requirement. The effect generalizes across independent learning directions, convolutional and transformer architectures, vision and text, and additional seeds. Learning rules differ in how efficiently they exploit available compatibility, while retention constraints limit how much can be stored. At larger finite updates, nonlinear geometry changes the available learning opportunity and ultimately prevents the matched compatibility continuum from being realized. These results establish functional compatibility as an experimentally controllable principle of persistent neural learning, shifting the problem from preventing forgetting towards identifying which components of new learning can safely become permanent.

AI 总结：本研究针对终端智能体泛化问题，提出智能体组合泛化假设，开发训练方案River，其用不足30%的TMax环境使多规模模型在两个基准上RL增益平均提升超100%，且性能优于开源8B模型、可跨多维度泛化。

链接：https://arxiv.org/abs/2608.22631

机构：Salesforce AI Research（Salesforce人工智能研究院）； Carnegie Mellon University（卡内基梅隆大学）

作者：Yihang Yao, Bo Pang, Xuan Phi Nguyen, Ding Zhao, Shafiq Joty, Semih Yavuz

英文摘要：Terminal agents are a compelling application of large language models (LLMs), with the potential to integrate deeply into users' daily workflows. Reinforcement learning (RL) is a key technique for improving their capabilities, making scalable training environments a central challenge. Since public real-user interaction data are scarce, synthetic environments provide a practical alternative, but often suffer from domain gaps and limited fidelity, leading to poor generalization. Existing work mainly scales the quantity and diversity of synthetic environments, while reward-signal quality and the mechanisms governing generalization remain under-explored. We study how RL improves terminal agents and propose the Agentic Compositional Generalization hypothesis: rather than teaching new domain-specific skills from scratch, RL primarily shapes high-level decision-making behaviors that compose and route low-level skills acquired during pre-training and supervised fine-tuning (SFT). This account is consistent with our empirical results and suggests that verifier quality, which determines which behaviors are reinforced, is more important than simply increasing environment quantity or diversity. Motivated by this insight, we propose River, a simple training recipe that improves reward quality by filtering low-quality environments and augmenting outcome rewards with process-level behavior regularization. Using this recipe, our RL-trained agent achieves the best performance among evaluated open-source RL-trained 8B models across four terminal-agent benchmarks. River also generalizes across model families, scales, agent harnesses, and RL objectives. Using fewer than 30% of the TMax training environments, River improves RL gains by 106% and 30% on average for models ranging from 2B to 27B on Terminal-Bench-Lite and Terminal-Bench-v2.1, respectively.

AI 总结：本文针对混合专家模型的路由诱导偏差问题，提出子组重加权与门控熵正则化结合的端到端框架，在提升公平性的同时维持了预测性能。

链接：https://arxiv.org/abs/2608.22820

机构：Dongyang Mirae University（东阳未来大学）

作者：Sunhee Hwang

英文摘要：Deep learning models often produce performance disparities across demographic groups, due to the training data imbalance with respect to sensitive attributes such as gender or age. To address this problem, existing work has explored fair representation learning, data re-sampling, and adversarial training, which can be broadly categorized into two main approaches. Single-stage methods typically learn a shared representation for fairness, but often struggle to handle heterogeneous subgroup distributions. Two-stage methods learn representations separately from the final prediction task, which can lead to misalignment between fairness objectives and downstream predictions. We identify routing-induced bias, a failure mode in which subgroup imbalance drives the gating network to route subgroups onto a few experts, and propose an end-to-end Mixture-of-Experts (MoE) framework that corrects it. Specifically, we apply subgroup reweighting to correct data imbalance, and introduce gate entropy regularization to prevent routing from collapsing onto subgroup attributes, keeping expert utilization both balanced and interpretable. Beyond improving fairness, the routing distribution offers an interpretable view of how subgroups are allocated across experts. Experimental results demonstrate that the proposed approach improves fairness while maintaining competitive predictive performance.

AI 总结：该研究开发了PolyChirp方法，结合多技术实现低功耗传感器的多物种鸟鸣分类，性能优于现有技术，可同时分类10种鸟类，满足野外长期运行的资源要求。

链接：https://arxiv.org/abs/2608.23101

机构：École Polytechnique（巴黎综合理工学院）； Freie Universität Berlin（柏林自由大学）； Berlin University of Applied Sciences (BHT)（柏林应用科技大学）； HES-SO Valais-Wallis（瓦莱州瓦利斯高等专业学院）； Inria（法国国家信息与自动化研究所）

作者：Nathan Duboisset, Zhaolan Huang, Felix Bießmann, Roudy Dagher, Antoine Lavandier, Emmanuel Baccelli

英文摘要：Recent progress in the field of TinyML has demonstrated that low-power hardware based on microcontrollers can achieve bird species monitoring in real time based on acoustic sensor data for an entire breeding period on a single battery charge. However, the state of the art on low-power microcontrollers was so far limited to binary classification of a single species. In contrast, real fauna monitoring deployments often target multiple species simultaneously. To address this challenge we develop PolyChirp, an approach combining biological domain expertise, automated dataset curation, neural architecture optimization and novel hardware to achieve multiclass bird species detection in the wild. PolyChirp is based on newly designed tiny multiclass models that leverage recent microcontrollers and hardware acceleration with a neural processing unit (NPU). We evaluate the predictive performance of these models, and we measure their computational performance -- memory footprint, latency, energy consumption -- on common microcontroller hardware. Our results demonstrate that PolyChirp not only outperforms state-of-the-art on single species binary classification, but also achieves robust classification of up to 10 species simultaneously, while still fitting with the resource envelope of a sensor that must remain operational in the field for a full season on a single battery charge.

AI 总结：该研究提出BPCO流程稳定高效训练评价模型，在数学推理任务上，其优于基线，且采样1个响应时可匹配或超过基于组的GRPO基线。

链接：https://arxiv.org/abs/2608.23566

机构：National University of Singapore（新加坡国立大学）； Tencent Hunyuan（腾讯混元）

作者：Penghui Qi, Xiangxin Zhou, Wee Sun Lee

英文摘要： Group-based reinforcement learning methods such as GRPO for large language models avoid training a critic by sampling multiple responses for each prompt. A reliable critic could instead estimate token-level advantages from one response, but standard critic-based training recipes are often unstable. We study this instability and develop \textbf{Best-Practice Critic Optimization (BPCO)}, a recipe that combines DPPO, value predictions bounded to the reward range, Monte Carlo value targets, unnormalized policy advantages, and length-adaptive generalized advantage estimation. Because the critic is used only during training, BPCO can also condition it on reward-defining information, such as a reference answer or grading rubric, that is hidden from the policy. Controlled experiments isolate the effect of each design choice. Across mathematical reasoning tasks with models ranging from 1.5B parameters to 30B-A3B mixtures of experts, BPCO improves a strong critic-based baseline consistently, and matches or exceeds a group-based baseline while sampling one response per prompt. The same recipe also improves learning with rubric-based rewards. These results show that a carefully designed critic provides a reliable alternative to group-relative advantage estimation. Code is available at this https URL

AI 总结：本研究以TPC数据为测试平台，通过冻结编码器的探针方法，发现架构是TPC嵌入中任务相关结构的主要来源，其诱导的表示可跨实验和探测器系统复用。

链接：https://arxiv.org/abs/2608.21756

机构：Michigan State University（密歇根州立大学）； Grand Valley State University（大峡谷州立大学）； Davidson College（戴维森学院）； University of Kentucky（肯塔基大学）

作者：Tyler Wheeler, Michelle P. Kuchera, Raghuram Ramanujan, William Sieland, Ryan Krupp, Daniel Bazin, Connor L. Cross, Hoi Yan Ian Heung, Andrew J. Jones, Ruchi Mahajan, Saiprasad Ravishankar, Pranjal Singh, Benjamin Votaw, Chris Wrede

英文摘要：Deep-learning efforts have increasingly shifted toward foundation model approaches. In experimental physics, this allows models and learned representations to be reused beyond the experiments in which they were developed. This work evaluates the reusability of representations across experiments and detector systems using probes on frozen encoders. These probes reveal task-relevant structure before downstream adaptation, complementing fine-tuning. Together with random-weight controls, they distinguish contributions from architecture and encoder training that downstream performance alone cannot resolve. Time projection chamber (TPC) data provide a useful testbed because events from TPC systems can be represented as variable-length sparse tensors, while detector geometries, event topologies, and scientific tasks can differ substantially. We investigate whether fixed-dimensional TPC event representations can be reused across classification tasks, experiments, and detector systems. Sparse ResNet and PointNet-style encoders produce 512-dimensional embeddings for four datasets from the GADGET II TPC and AT-TPC. Randomly initialized encoders isolate the contribution from architecture before supervised training. We then train each encoder on a classification task, freeze its parameters, and train a linear or nonlinear probe for each downstream task. We find that this architecture-induced structure remains useful across experiments and detector systems. The randomly initialized PointNet-style representation is highly informative on several tasks. The two architectures organize their embedding spaces differently, but neither exhibits a large, systematic loss of utility cross-detector. These results show that architecture is a major source of task-relevant structure in TPC embeddings and should be treated explicitly when assessing representation learning and developing reusable detector models.

AI 总结：该研究提出仅从权重生成语言模型所有潜在通信通道的通信图，通过实验验证其可恢复诱导电路并消除模型的诱导能力，发布了相关工具套件。

链接：https://arxiv.org/abs/2608.22007

机构：St. John Fisher University（圣约翰费舍尔大学）

作者：Richard Zhe Wang

英文摘要：The components of a transformer communicate by writing to and reading from a shared residual stream, and mechanistic interpretability has mapped these connections by hand, one circuit at a time. We present the communication map, which charts every potential communication channel in a language model from weights alone, generalizing the composition score of Elhage et al. (2021) into a single coupling coefficient covering all 18 connection classes, from entire attention head circuits to single neurons. The census of all candidate channels, from $6.3\times10^{8}$ in GPT-2 to $1.3\times10^{11}$ in Pythia-6.9B, finds that 70-89% of head pairs are oriented far from chance, some coupled strongly and others actively avoiding each other. The full map costs 15 seconds for GPT-2 and 11 minutes for Pythia-6.9B on one consumer GPU. Two applications demonstrate the utility of the map. In Application 1, the strongest head-to-head couplings recover the known induction circuits blind and group them into communities, and ablating one such community destroys the model's in-context copying. In Application 2, pooling every head's coupling coefficients identifies a distinct two-dimensional stream subspace, whose deletion abolishes the induction capability in six models up to Pythia-6.9B. This subspace is different from those identified by either activation PCA or outlier dimensions. We release the map, the statistical machinery, and the intervention suite.

AI 总结：本研究提出ReMAP自监督学习方法，基于EEG揭示大脑表征与脆弱性，可准确预测麻醉深度，其紧凑模型性能可媲美大规模EEG基础模型，还能关联临床结局与神经生理学特征。

链接：https://arxiv.org/abs/2608.22042

机构：AP-HP（巴黎公立医院集团）； Inserm（法国国家健康与医学研究院）； Université Paris Cité（巴黎城市大学）； Université Paris-Saclay（巴黎-萨克雷大学）； Inria（法国国家信息与自动化研究所）； CEA（法国原子能和替代能源委员会）； Sorbonne Université（索邦大学）

作者：Jade Perdereau, Virginie Loison, Kanssa El Ayeb, Louis Gervais, Melvin Berto Strouc, Fabrice Vallée, Thomas Moreau, Jérôme Cartailler

英文摘要： General anesthesia offers a rare opportunity to observe the human brain under a standardized, controlled perturbation. Yet intraoperative electroencephalography (EEG) is almost always reduced to a single proprietary depth index, collapsing a rich trajectory into one number and discarding how a brain moves between states. Here we ask whether the geometry of that trajectory, not merely the depth it reaches, carries clinically meaningful information. Using similarity-based self-supervised learning on raw, two-electrode frontal EEG, with no labels, we place each recording within a low-dimensional space in which anesthetic depth becomes one readable axis while the shape of a patient's path encodes additional structure. We validate the representation across two cohorts and two acquisition systems totaling more than 1,000 patients. Depth of anesthesia is predicted accurately (BIS mean absolute error = 3.2, R2 = 0.82), and in the sparse-montage setting our compact ( 68k parameter) model remains competitive with EEG foundation models orders of magnitude larger (4M-157M parameters), indicating that matching the representation to the recording dominates raw scale. The learned space organizes age along its own gradient, independent from depth, without supervision. The same space also aligns with interpretable anesthetic signatures like frontal alpha, slow-delta, and burst suppression, linking this data-driven representation to established neurophysiology. On an independent cohort with longitudinal follow-up, the geometry of the early trajectory separates 30- month cognitive and mortality outcomes complementary to age (AUROC 0.86). These results suggest that the path a brain traces through anesthesia is a label-efficient correlate of latent vulnerability, motivating prospective validation.

AI 总结：针对员工职业抱负对个人与组织成长的重要性，提出结合职业路径相似度（CPS）与抱负相似度（AS）的数据驱动方法推荐个性化职业路径，定义了领域知识驱动（DKD）和无监督表示学习与对齐（URLA）两种CPS计算方法，URLA方法无需领域知识且包含时间维度，在DCG指标上优于DKD。

链接：https://arxiv.org/abs/2608.22056

作者：Kuleshwar Sahu, Girish Keshav Palshikar, Rajiv Srivastava

英文摘要：Fulfilling career aspirations is important for growth of employee and organization. We propose a data driven methodology to recommend personalized career path for a given aspirant's career path and aspirations. The pro-posed method uses the career path similarity (CPS) between aspirant's career and candidate career path, and 'aspirational similarity' (AS) between aspiration and candidate career paths to find suitable career path. CPS ensures personalized recommendation while AS ensures aspiration fulfillment. We defined two methods to compute the CPS between career paths which are (a) domain knowledge driven (DKD) and, (b) unsupervised representation learning and alignment (URLA) based, along with different AS measures. The DKD based similarity is defined in the terms of features extracted and summarized over career paths. In the URLA, we use the sequence of event names present in the career paths of the employees to learn the embedding for each event name. In URLA we use learned embedding vector of the career path event names and as-sociated event attributes (skill cluster and domain) to find the best alignment between two career paths. We hypothesized that relative position of event names in the sequence represents semantics of event name and that can be learned. We use LSTM neural network to learn the embedding vector of each career event name. We also define the matching method to compute the AS be-tween aspiration and career path in both proposed methods. We combine CPS and AS to rank available 'candidate career paths' of employees to find the suitable one. We get better DCG value in URLA as compare to DKD. We also showed that ranking are coherent using both the methods. URLA method is better since it does not require domain knowledge to model the similarity and includes temporal aspect by optimal Levenshtein alignment using weighted cosine distance.

AI 总结：该研究对比了直接行为克隆策略等三类策略，明确世界-动作模型学习与直接行为克隆的能力差异，指出观测演示无法识别动作效果，干预可实现零遗憾值。

链接：https://arxiv.org/abs/2608.22197

机构：Nanjing University（南京大学）

作者：Yang Yu

英文摘要：World-action models predict a future outcome and then infer an associated action. Although this factorization can improve representation learning and data efficiency, it is unclear whether it provides stronger control capability than direct behavior cloning when both are trained from the same observational demonstrations. We compare a direct behavior-cloning policy, an imitation-trained world-action policy, and a policy optimized with an action-conditioned world model. At the controller-class level, every world-action policy can be flattened into a direct stochastic policy with the same closed-loop trajectory distribution. At the population level, under realizability, exact optimization, common deployment information, and distribution-preserving deployment, direct behavior cloning and world-action imitation both recover the observational behavior policy. Thus, future prediction changes the learning factorization but not the unrestricted external policy class or ideal imitation target. Action-conditioned world-model learning differs by predicting outcomes under specified actions and comparing them through a control objective. We characterize the irreducible action-specific prediction error of future models that do not condition on the candidate action, identify conditions under which a world-action joint can recover an interventional forward model, and show that observational demonstrations do not identify action effects in general. Finally, we construct an environment family in which every observational learner has positive worst-case regret, whereas one informative intervention permits zero regret. The key distinction is therefore between predicting futures associated with observed behavior and predicting consequences of specified actions for policy optimization.

AI 总结：本文针对高度混合粒度分布数据的解混问题，提出了 AnalySize 的推广方法——最大距离非负矩阵分解，通过最大化端元间距离结合分层交替最小二乘算法，实现了对高度混合数据的有效分解。

链接：https://arxiv.org/abs/2608.22681

机构：Hangzhou Dianzi University（杭州电子科技大学）； Utrecht University（乌得勒支大学）； University of Southampton（南安普顿大学）

作者：Qianqian Qi, Zhongming Chen, Peter G. M. van der Heijden

英文摘要：Nonnegative matrix factorization (NMF) decomposes a nonnegative matrix into the product of two nonnegative matrices. This property makes NMF well suited for unmixing grain-size distribution data, which are inherently nonnegative and have row sums equal to one. Previous studies have shown that AnalySize, an NMF-based method, performs well on poorly mixed grain-size distribution data but struggles when the data is highly mixed, where no observed samples are close to the true end members. To overcome this limitation, we introduce a maximum-distance NMF that encourages the estimated end members to be as distinct as possible and develop a hierarchical alternating least squares algorithm for optimization. The proposed formulation can be regarded as a generalization of AnalySize, where AnalySize minimizes the distance among end members while the proposed method maximizes it. Experimental results demonstrate that the method effectively decomposes highly mixed grain-size distribution data.

AI 总结：针对不平衡时间序列分类中现有过采样方法泛化与多样性不足的问题，提出FreMGP方法，结合对比学习的频域判别表示引导多树遗传规划，实验证明其性能优于现有方法并提升各类分类器表现。

链接： https://arxiv.org/abs/2608.22804

机构：Dalian University of Technology（大连理工大学）； Nanyang Technological University（南洋理工大学）

作者：Wenbin Pei, Yunrong Hao, Zhen Liu, Guan Wang, Bing Xue, Yiu-Ming Cheung, Qiang Zhang

英文摘要：Real-world time-series classification tasks often exhibit class imbalance, which can be extremely severe in some applications. To avoid training biased classifiers on imbalanced data, sampling is one of the most popular data pre-processing techniques because of its classifier-agnostic nature. However, due to the complex temporal dependencies in original time-series data and the scarcity of minority-class samples, existing sampling methods, including interpolation-based oversampling methods and deep learning-based generative models, usually suffer from limited generalization and poor diversity when generating new time-series samples. This paper proposes a Frequency-domain representation-guided Multi-tree Genetic Programming-based oversampling approach (FreMGP) to imbalanced time-series classification, where each individual represents a set of synthetic samples for the minority class. A frequency-domain class-discriminative representation module based on contrastive learning is also developed, guiding the evolutionary search toward high-quality synthetic time-series samples. Experiments on imbalanced time-series datasets demonstrate that FreMGP outperforms existing oversampling methods and consistently improves the performance of different classifiers, including both general machine learning and deep learning models.

AI 总结：本研究提出支持10240nt上下文的长RNA基础模型RIBOSPAN，结合多种技术实现高分辨率长RNA建模，在多任务评估中达最优性能，并开发基于其主干的mRNA生成与优化框架。

链接：https://arxiv.org/abs/2608.22849

机构：Shanghai Innovation Institute（上海创新研究院）； Center for Excellence in Molecular Cell Science, CAS（中国科学院分子细胞科学卓越创新中心）； University of Chinese Academy of Sciences（中国科学院大学）； Shanghai Jiao Tong University（上海交通大学）

作者：Ziyuan Wang, Bohao Tang, Fei Zhang, Shuo Han, Pengfei Liu

英文摘要：Full-length RNAs, particularly messenger RNAs, often exceed the context lengths used to pretrain existing RNA foundation models, limiting complete-transcript modeling at single-nucleotide resolution. We present RIBOSPAN, a 1.61-billion-parameter bidirectional RNA foundation model natively pretrained with context lengths up to 10,240 nt. RIBOSPAN combines dense bidirectional self-attention, single-nucleotide tokenization, and attention-isolated sequence packing to enable high-resolution modeling of complete long RNAs. We evaluate the model through nucleotide reconstruction, a controlled long-context representation benchmark, and frozen RNA-type representation analysis. Native 10K pretraining preserves strong reconstruction at 10,240 tokens, while continued pretraining with 40% masking improves recovery under heavy corruption while preserving representation quality. The long-context benchmark further shows that native 10K models maintain strong contextual responsiveness and context-specific representation separation while keeping perturbation-induced representation changes highly localized. Inference-time YaRN scaling recovers much of the contextual organization lost by direct extrapolation of short-context models, but induces substantially greater distal representation diffusion. Frozen-representation evaluations further demonstrate state-of-the-art RNA representation quality, with RIBOSPAN achieving the strongest overall performance across diverse RNA types and retaining a clear advantage on long RNAs. Building on the same backbone, we develop a multidimensionally conditioned discrete-diffusion framework for full-length mRNA generation and redesign, including synonymous-codon diffusion for protein-preserving CDS optimization. Together, RIBOSPAN establishes a powerful long-context foundation for transferable RNA representation learning and full-transcript mRNA design.

AI 总结：本文针对两类对象嵌入流形，提出并证明了随机可分性定理，通过投影测度集中分析技术，揭示了对象嵌入流形的几何与统计特性，为深度网络表示学习提供新机制。

链接：https://arxiv.org/abs/2608.22874

机构：School of Computer Science（计算机学院）； Shanghai Jiao Tong University（上海交通大学）

作者：Liqing Zhang

英文摘要：Neurobiological studies and representation learning have observed that representations of objects belonging to the same category in high-dimensional neural spaces exhibit low-dimensional object manifold characteristics, and different object manifolds are linearly separable in these neural spaces. However, these experimentally observed phenomena lack rigorous theoretical validation to date. This paper proposes a new stochastic separability theorem for embedding manifolds of two different object categories. First, we establish a projection measure concentration theorem for embedding manifolds under general conditions. We develop a new two-layer measure concentration analysis technique, which unifies two estimation bounds via the law of total expectation to derive measure concentration inequalities. Based on the measure concentration theorem, we further prove a stochastic separability theorem for embedding manifolds of two different object categories. If two datasets have distinct means and bounded total variances, their samples become linearly separable with high probability, provided that the projection direction satisfies a non-singularity condition. The main contributions of this paper are twofold: 1. We prove the projection concentration properties of embedding manifolds in high-dimensional spaces by using two-lawyer tail-bound inequalities. 2. We identify a non-singularity condition for the stochastic separability between embedding manifolds, and rigorously prove the stochastic projection separability theorem. The theorem not only uncovers geometric and statistical properties of the object embedding manifolds, but also provides a novel mechanism for representation learning in deep networks.

AI 总结：该研究对比零样本、上下文注意力等四种模型专业化机制，发现生成专家模型在匹配质量下成本更低，但在高维序列建模中不及上下文注意力，还提出可证伪论点界定各机制适用场景。

链接：https://arxiv.org/abs/2608.21386

机构：Nineonefour

作者：John C. Howell

英文摘要：Given a task described by a few examples, how should a model be specialized to it? Four mechanisms are available -- zero-shot, in-context attention, test-time gradient adaptation, and emitting specialist weights from a hypernetwork -- yet the operating regime of the last is rarely mapped. We run the identical four-way comparison across six tasks spanning regression, generation, language modeling, reinforcement learning, and clinical and genomic classification, holding the specialist, the context, and (where we can) the training budget fixed. The clearest wins for emission are about cost at matched quality: it ties the state-of-the-art amortized tabular model (TabPFN) on clinical few-shot classification while emitting a reusable specialist instead of re-attending the support set per query, and reaches noise-floor shape generation with a $132$-float per-instance program. On few-shot sinusoid regression it is $2$--$3$ orders of magnitude below MAML at zero test-time gradient steps -- a margin that narrows to $\sim$$30\times$ but persists once training budgets are equalized. Emission cannot match in-context attention on high-dimensional sequence modeling: under matched-budget pre-training a one-pass adapter recovers only a minority of the in-context gain ($14.0\pm0.9\%$ at $5$M, $11.2\pm0.5\%$ at $15$M), and a LoRA-rank sweep shows this shortfall is a partial capacity limit -- capture climbs from $5\%$ to $21\%$ as rank grows but plateaus far below full recovery. Mechanism ablations confirm the emitted specialist is genuinely task-conditioned, not a memorized prior; and, more speculatively, emitted specialists compose in weight space -- interpolating two of them tracks the corresponding blend of their functions. We close with a falsifiable thesis, operationalized through a per-task resolution measure, bounding when each conditioning mechanism should be preferred.

AI 总结：该研究提出运行时动作干预（RAI）机制，将其应用于AlphaStar复现版并开展《星际争霸II》人类实验，发现向用户披露AI对手能力会显著影响人类对其公平性、毒性的感知，需将执行栈控制与能力披露分开评估。

链接：https://arxiv.org/abs/2608.21398

机构：University of New South Wales（新南威尔士大学）； CSIRO（澳大利亚联邦科学与工业研究组织）

作者：Jaymari Chua, Chen Wang, Liming Zhu, Lina Yao

英文摘要：A trained reinforcement learning policy does not determine the complete behavior that users encounter: deployment code still schedules, admits, suppresses, or replaces its proposed actions. We contribute \emph{runtime action interference} (RAI), an AI control mechanism that preserves policy parameters while regulating action pacing and filtering configured action patterns after inference. RAI releases a proposed action only when its cooldown condition is satisfied and its content detector does not flag the action; otherwise, it dispatches a no-op. The detector covers specified toxic behaviors, including worker-unit harassment, while the cooldown controls action rate. We implement RAI in a replication of AlphaStar this http URL and make the implementation and reproducibility materials available through an open source code repository. We deployed RAI in a \textit{StarCraft~II} human participant study that compared two presentations of the same opponent with high capability and rate limited actions; we withheld its capability claim in one presentation and disclosed it in the other. On response scales from 1 to 5, we observed pooled fairness, trust, and toxicity means of 3.90, 3.50, and 2.00 under claim withholding, compared with 2.62, 4.31, and 2.85 under disclosure. Disclosure corresponded with lower perceived fairness and higher perceived toxicity across every expertise group, whereas trust increased among novices and experts but decreased among intermediate participants. Our human evaluation therefore shows that perceptions of an opponent controlled through RAI can vary substantially with the capability information presented to users, even when the configured control remains constant. We conclude that human-computer evaluations must separate control within the execution stack from capability disclosure and assess fairness, trust, and toxicity as distinct dimensions of human experience.

AI 总结：该研究发现，基于良性事实的强化学习可放大指令模型对已记忆个人身份信息的泄露，且推理能力与拒绝率得以保留，为攻击者提供了无需接触私人数据即可提取的途径。

链接：https://arxiv.org/abs/2608.21727

机构：Carnegie Mellon University（卡内基梅隆大学）

作者：Renfei Zhang, Niloofar Mireshghallah

英文摘要：Reinforcement learning with verifiable rewards (RLVR) is deployed to make models better at reasoning tasks, but its side effect on what models will divulge is under studied. Here we show that RLVR on facts increases extraction of personally identifiable information (PII) the instruct model had already memorized. We first confirm that instruct models have already memorized PII but leave them latent, rarely surfacing one when asked. We then apply RL on benign factual data that contains no PII of any kind, and re-probe: a targeted probe over name->email pairs, and an untargeted free-recall prompt that simply asks the model to list the addresses it knows. PII extraction rises sharply under both: on DeepSeek-V3.1, verbatim recall@k increases from 0.155 to 0.370, a 2.4x gain. The effect scales with model size: across three models spanning 8B to 671B parameters, absolute leakage is largest in the biggest model. Meanwhile model's reasoning abilities and refusal rates are retained, indicating that RL selectively changes which memorized information is accessible rather than broadly altering the model. In summary, memorized private data can be made markedly more extractable by training that never touches it. This gives an adversary a route to memorized data that requires no privacy-relevant training signal and no access to the data itself -- only the ability to fine-tune on something innocuous.

AI 总结：该研究以Qwen3-0.6B为对象，通过实验验证了推理与记忆可通过单一方向分离，但经GRPO后该方向几何结构显著重组，不过任务组的AUROC仍保持1.00，挑战了固定方向稳定性的假设。

链接：https://arxiv.org/abs/2608.21919

作者：Shaheen Nabi

英文摘要： Recent work has proposed that reasoning and memorization in language models can be characterized by a single representation direction, including methods that keep this direction fixed during reinforcement learning. We test two assumptions behind this view. First, are reasoning-oriented and factual-recall task groups approximately single-direction separable? Second, does the resulting geometry remain stable after GRPO? Using Qwen3-0.6B and a controlled 400-example dataset, we find that a one-dimensional projection can match a full 1024-dimensional linear probe with AUROC = 1.00 on the studied task groups. However, after GRPO, the corresponding direction is substantially reorganized: mean-direction cosine averages 0.453, probe-direction cosine 0.445, while direct representation drift reaches 0.511 at the final layer. Probe AUROC nevertheless remains 1.00. The evidence therefore supports single-direction decodability for the studied task groups but challenges fixed-direction stability: the information persists while its geometric realization changes.

AI 总结：该研究提出门控解耦组合多臂老虎机（GDCB），证明其可统一多个工业系统，核心定理可将非平稳问题转化为近似平稳问题，还推广了相关结果，配套论文已验证短期租赁动态定价实例。

链接：https://arxiv.org/abs/2608.21993

作者：Oleg Miroshnichenko

英文摘要：We introduce Gated Decoupled Compositional Bandits (GDCB), a family of contextual bandit algorithms with three structural innovations that jointly fall outside the taxonomy of LinUCB, LinTS, HierTS, factored bandits, neural contextual bandits, and RLHF. In a GDCB system: (i) the action delivered to the environment is the composition of a nominal arm, drawn by a discrete or hierarchical bandit, with a context-dependent scaler; (ii) the scaler parameter is learned in a separate supervised loop, not jointly with arm selection; and (iii) every action passes through a pre-execution gate that may modify or veto the composed action before it reaches the environment. We formalise this class of algorithms, prove four structural theorems characterising its statistical behaviour, and show that six industrially significant systems -- short-term rental dynamic pricing, clinical drug dosing, credit origination, grid demand response, content moderation, and LLM tool-use agents -- are all instances of GDCB, differing only in the composition operator, scaler family, and gate. The central result is the Decoupling Variance Reduction theorem: a well-calibrated scaler removes context-induced variance from the arm-to-reward mapping, turning a non-stationary bandit problem into an approximately stationary one. The Gate-Induced Equivalence theorem shows that under a stationary gate, historical data collected under any prior policy is a valid warm-up initialiser without importance-sampling correction, generalising the companion P-HITL result ( arXiv:2606.02595 ) from human approval to arbitrary gates. In regulated, high-stakes domains, constraints usually treated as deployment frictions -- approval gates, compliance rules, safety shields -- are the mechanism that makes fast deployment possible, not an obstacle to it. The companion paper validates instance 1 (STR dynamic pricing) on real production data.

AI 总结：该研究提出 VarDE 方法，将其应用于 BAI、MCTS、BPI 问题，在方差衰减和简单遗憾上有理论保证，在高度随机环境中性能优于现有方法。

链接：https://arxiv.org/abs/2608.21995

机构：Hanoi University of Science and Technology（河内科技大学）； Quantum AI & Cyber Security Institute, FPT Corporation（FPT公司量子人工智能与网络安全研究院）

作者：Khang Luong, Nam Nguyen, Hoang Ta, Hung The Tran, Tuan Dam

英文摘要：We propose Variance Driven Exploration (VarDE), a principled approach for pure exploration in highly stochastic environments, where the exploration process is dominated by stochastic variance. VarDE is built on a fundamental principle: sampling effort should be allocated to minimize the uncertainty of the final decision. We formalize the uncertainty of the final decision through a smooth decision function and derive allocation rules that explicitly capture how stochastic noise in individual components affects the reliability of the final output. We apply this methodology to three core problems of pure exploration -- Best Arm Identification (BAI), Monte Carlo Tree Search (MCTS), and Best-Policy Identification (BPI) -- with theoretical guarantees on variance decay and simple regret. Empirically, we demonstrate consistent and significant improvements of VarDE over existing methods, with especially strong gains in highly stochastic environments.

AI 总结：本研究针对传感器融合中周期性结构污染导致的卡尔曼滤波器协方差校准问题，提出FFT-GDCB四角色预滤波方案，在六个领域验证其性能且计算开销极低。

链接：https://arxiv.org/abs/2608.22023

作者：Oleg Miroshnichenko

英文摘要：Context-adaptive Kalman filters calibrate their noise covariance matrices Q and R from innovation residuals via online regression. When the underlying sensor or signal carries periodic structure -- mechanical LiDAR rotation harmonics, engine vibration, ground multipath, weekly and annual demand cycles, dosing-interval rhythms, weekly media-buying cadence -- the regression input is contaminated and the fitted covariance models structural modes rather than genuine state uncertainty. We introduce a four-role FFT pre-filter that solves this problem at $O(N\log N)$ cost and serves three additional roles "for free": (i) it whitens coloured noise before the Kalman update, restoring the optimality assumption; (ii) it cleans innovations before covariance regression, preventing periodic contamination of $\hat{R}$ and $\hat{Q}$; (iii) it generates spectral context features that enrich the downstream bandit's regime-selection state; (iv) it deseasonalises the input feature vector before any supervised regression that produces a sensitivity coefficient (beta, dose offset, bid modifier). We position the algorithm inside the Gated Decoupled Compositional Bandits (GDCB) family, where it acts as a preprocessing layer for the supervised scaler. The single $O(N\log N)$ FFT call thereby serves four downstream consumers, fits in <0.1% of the sensor-fusion or pricing-pipeline compute budget, and is a drop-in addition with no changes to the Kalman filter, bandit, or runtime composition operator. We summarise empirical validation across six independent domains (rocket descent, autonomous-vehicle tracking, short-term rental pricing, clinical drug dosing, airline fare distribution, and ad-operations bid calibration), all returning a PROVES verdict under a pre-registered evaluation protocol.

AI 总结：该研究提出反事实商模型，通过移除动作间共享的未来分量，直接从同步反事实回滚学习动作依赖效应，在物理环境实验中验证其可抑制无关变异、提升动作排序效果。

链接：https://arxiv.org/abs/2608.22092

机构：School of Computer Science and Engineering, Beihang University（北京航空航天大学计算机科学与工程学院）

作者：Junlin Chen, Ruijie Wang, Jianxin Li

英文摘要：Reinforcement-learning models commonly predict complete future states, observations, or feature occupancies, even though action selection depends only on differences between the consequences of candidate actions. As a result, these models may devote substantial statistical and representational capacity to high-dimensional phenomena that evolve independently of the agent's current choice. We introduce the Counterfactual Quotient Model, which treats action-conditioned futures as equivalent when they differ only by a component shared across actions. Its canonical centered representation removes this common component while preserving every pairwise action comparison expressible by the modeled reward family. The implemented model learns these action-dependent effects directly from synchronized counterfactual rollouts, so shared stochastic dynamics cancel before function approximation rather than after complete futures have been predicted. We establish the decision sufficiency, identifiability, common-mode invariance, approximation behavior, and regret properties of the resulting representation. Controlled experiments in physics-based environments provide initial evidence for these properties: direct effect learning suppresses action-independent variation, supports previously unseen reward queries, and improves action ranking relative to models trained to predict absolute futures.

AI 总结：该研究针对多智能体多臂老虎机的Top-K联合臂识别问题，设计无通信的UCB-Intervals算法，量化分析了去中心化移除通信的统计代价，证明完全非对称下样本复杂度有4倍固定惩罚。

链接：https://arxiv.org/abs/2608.22120

机构：University of California, Los Angeles（加利福尼亚大学洛杉矶分校）

作者：Larissa Xu, Jasmine Nguyen, William Chang

英文摘要：Cooperative teams often need to agree on the best few options rather than simply accumulate reward, and they must do so while each member sees only a fragment of the team's collective experience. We study this as top-$K$ joint-arm identification in multi-agent multi-armed bandits: at every round $M$ agents simultaneously choose individual actions that compose a joint arm, and the team must ultimately return the $K$ joint arms of highest mean reward. The difficulty is that an agent may not observe the actions of others, their rewards, or either. We treat three observability regimes---(A) shared rewards with hidden actions, (B) observed actions with private rewards, and (C) full asymmetry---and design communication-free elimination algorithms (UCB-Intervals) that reconstruct implicit coordination from whatever signal each regime leaves intact: a shared arm ordering in (A), observable deviations in (B), and enlarged confidence radii under (C). We give matching analyses in both the fixed-budget and fixed-confidence objectives, then fold all three regimes into a single meta-guarantee indexed by a multiplicity $c$ and a consensus factor $\rho$. Our central result is quantitative rather than merely algorithmic: change-of-measure lower bounds show that shared-reward identification is optimal up to one universal logarithmic factor, and that the entire statistical price of removing communication is a multiplicative $\rho^2$ in sample complexity---a fixed $4\times$ penalty under full asymmetry. The resulting stopping time scales as $O\!\left(\sum_{\mathbf{a}} \frac{\log(A^M/\delta)}{\Delta_{\mathbf{a}}^2}\right)$ and the fixed-budget error as $\exp(-\Theta(T/H_1))$, with the dependence on the joint-action count $A^M$ shown to be unavoidable.

AI 总结：该研究提出解耦物理建模与执行的统一框架，通过两阶段后训练策略提升小型大语言模型物理推理性能，在多个基准上实现约3%的平均性能提升。

链接：https://arxiv.org/abs/2608.22126

机构：University of Pennsylvania（宾夕法尼亚大学）； William & Mary（威廉玛丽学院）； University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）； Amazon（亚马逊公司）； Stevens Institute of Technology（史蒂文斯理工学院）； Northwestern University（西北大学）

作者：Ye Zhang, Xuehang Guo, Rui Pan, Pengfei Yu, Denghui Zhang, Manling Li, Qingyun Wang

英文摘要：Physics reasoning requires constructing a consistent model of the underlying physical system rather than relying solely on symbolic or formula-based manipulation. Although large language models have shown strong ability in solving math and coding problems, they still struggle with physics problems, as these problems entangle the physical modeling process with mathematical calculations. Humans approach physics by first building a representation of the system before performing calculations. Inspired by this, we introduce a unified framework that distills intermediate representations that explicitly encode the physical modeling process and adopt a two-stage post-training strategy, where supervised fine-tuning establishes structured modeling, and reinforcement learning with rubric-based feedback improves the quality of the modeling process. Experiments on multiple multimodal physics benchmarks show that our approach leads to consistent improvements in reasoning performance across different models and datasets. On PhysReason, PhyX and SeePhys benchmarks, physical modeling output performs GRPO by an average ~3%, showing that explicit physical modeling is an efficient strategy of improving physics reasoning for small LLMs.

AI 总结：本文针对强化学习分位数目标的不稳定性问题，提出UCB-BQRL算法，引入缓冲分位数准则与EVI-BQ过程，建立了其regret界与信息论下界，并证明相关评估问题为PP难。

链接：https://arxiv.org/abs/2608.22227

机构：Virginia Tech（弗吉尼亚理工大学）

作者：Mohammad Alipour-Vaezi, Huaiyang Zhong, Sajad Khodadadian

英文摘要： Reinforcement Learning (RL) has achieved tremendous success in recent years. However, the classical foundations of RL do not account for the risk sensitivity of the objective function, which is critical in various fields, including healthcare, finance, etc. A popular approach to incorporate risk sensitivity is to optimize a specific quantile of the cumulative reward distribution. However, exact quantile objectives are non-smooth and can change abruptly under small perturbations of the return distribution, making them difficult to optimize reliably when the transition model must be learned from data. Motivated by this instability, we develop UCB-BQRL, a model-based optimistic learning algorithm that maintains confidence sets for the transition kernel and plans using a lower-buffered quantile criterion. The buffered criterion smooths the exact quantile objective by averaging nearby lower quantiles, thereby improving stability under transition-estimation error. To compute the buffered-quantile policy at each episode, we introduce EVI-BQ, an exact dynamic-programming procedure. We establish a high-probability regret bound for UCB-BQRL, which up to logarithmic factors scales as $\mathcal{O}(\mathrm{e}^{\tau/\rho_\tau}+H^2\sqrt{SAT})$, where $\rho_\tau$ is denoted as the root-level left-plateau threshold, which is a problem-dependent constant. Further, we establish an information-theoretic lower bound of $\Omega(H/\rho_\tau\sqrt{AT})$ for the regret of any algorithm dealing with a quantile objective function. Finally, we prove that the exact point-quantile evaluation and exact lower-buffered quantile evaluation are PP-hard under polynomial-time Turing reductions, even for a fixed policy in a two-state, one-action finite-horizon MDP.

AI 总结：该研究针对长 horizon 语言模型任务的信用分配问题，分析策略方差，得出其作为发现预算、受基尼离散度约束及剩余 horizon 决定估计成本等结论，提出自举法消除成本的思路，支持对数值参数化。

链接：https://arxiv.org/abs/2608.22467

作者：Yingru Li

英文摘要：Long-horizon language-model tasks --- multi-step reasoning and tool-using agents alike --- are limited by credit assignment. We analyze it through the policy variance $\sigma_\pi^2(s)=\operatorname{Var}_{a\sim\pi}[Q_{\pi}(s,a)]$, which in a deterministic MDP is the sole source of return variance and is injected in discrete pulses at states we call critical forks. Three results follow. (i) Policy variance is a discovery budget: observing an action of advantage $c$ requires $\Omega(c^2/\sigma_\pi^2(s))$ draws, a bound that is exact on the canonical two-point fork. (ii) Policy variance is bounded by the policy's Gini dispersion, $\sigma_\pi^2(s)\le 1-\|\pi(\cdot|s)\|_2^2$, a rollout-free necessary condition for criticality computable from logits alone. (iii) The remaining horizon sets the estimation cost: at a fork whose downstream success probability is $P$, the Monte Carlo advantage estimate has signal-to-noise ratio of order $\sqrt{P}$, so its sample cost scales as $1/P$ --- a cost that branched sampling shares. Bootstrapping removes it by converting a product of survival probabilities into a sum, provided the value representation is multiplicatively accurate, which argues for log-value parameterization.

AI 总结：该研究开发了稳定无限维线性函数近似的Q学习框架，提出两种SA算法并证明其收敛界，可自动适应几何与光滑性，还展示了其在Q测度学习等场景的应用。

链接：https://arxiv.org/abs/2608.22636

机构：University of Southern California（南加州大学）

作者：Shengbo Wang

英文摘要：Q-learning with linear function approximation can be unstable because an arbitrary approximation architecture need not preserve the Bellman contraction. We develop a stable infinite-dimensional linear function approximation framework for Q-learning from a single Markovian behavior-policy trajectory. The learning variable is a coefficient field $\theta\in C(\mathbb L)$ on a compact latent metric space $(\mathbb L,\rho)$. The framework uses a reconstruction operator that maps $\theta$ to a continuous Q-function and a compression operator that maps Bellman updates back to latent coordinates. Nonexpansiveness of both operators induces a contractive latent Bellman map on $C(\mathbb L)$, with a unique fixed point $\theta^*$ whose reconstruction approximates the optimal Q-function up to representation error. We propose two stochastic approximation (SA) algorithms and establish their sup-norm convergence bounds with a leading term of order $\widetilde O(n^{-1/2})$. The infinite-dimensional formulation provides a powerful abstraction for identifying the structures that govern statistical difficulty. Smoothness of the compression map in $\rho$ is inherited by $\theta^*$ and the SA iterates, allowing uniform estimation errors to be controlled through covering numbers of $(\mathbb L,\rho)$ rather than the dimension of $C(\mathbb L)$. Remarkably, the SA algorithms we propose are agnostic to the choice of $\rho$, and thus can automatically adapt to both the smoothness and the geometry. We further illustrate the framework through Q-measure-learning with linear density approximation and output-layer neural weight training under a frozen pretrained network.

AI 总结：该研究提出SANSAC算法，在传统硬件上验证其性能与SAC近乎相当，证明基于脉冲神经网络的算法可用于复杂连续环境，为神经形态强化学习研究奠定了基础。

链接：https://arxiv.org/abs/2608.22729

机构：New Mexico Institute of Mining and Technology（新墨西哥矿业技术学院）； The University of Texas at El Paso（德克萨斯大学埃尔帕索分校）

作者：Jessica Hunter, Md Maruf Hossain Shuvo, Krishna Roy

英文摘要：Reinforcement learning (RL) algorithms have made strides over the past decade applying them to a wide range of problems and control tasks. However, the deployment of RL on neuromorphic hardware for continuous control tasks remains under-validated. Namely it is unclear whether replacing a conventional actor network with a spiking neural network (SNN) affects the performance of an agent before any hardware-specific benefits manifest. We provide a systematic validation of a minimal, neuromorphically viable spiking actor variant of Soft Actor-Critic (SAC) on conventional hardware, establishing a baseline for future neuromorphic RL research. In this paper, we propose the Spiking Actor Network Soft Actor Critic (SANSAC) to address the use of RL frameworks in continuous environments, designed as a framework that can be implemented on neuromorphic hardware. We compare a traditional Soft Actor Critic (SAC) network to SANSAC in a traditional computer. We demonstrate the near equivalent performance of SANSAC and SAC, while addressing the impact of hidden dimensions. Our results demonstrate the viability of SNN based algorithms in complex continuous environments, as well as competitive performance to traditional neural networks in traditional computers, providing a basis to continue exploring the use of SNNs in continuous RL frameworks.

AI 总结：该研究针对耦合动力学环境的控制问题，提出联合马尔可夫决策过程（JMDP）的最优控制方法，定义分布型贝尔曼最优算子并证明收敛性，推导神经近似采样目标，拓展了JMDP的应用范围。

链接：https://arxiv.org/abs/2608.22765

机构：Elmore Family School of Electrical and Computer Engineering（埃尔莫尔家族电气与计算机工程学院）； Purdue University（普渡大学）

作者：Ege C. Kaya, Aliasghar Pourghani, Mahsa Ghasemi, Vijay Gupta, Abolfazl Hashemi

英文摘要：Coupled-dynamics environments expose the one-step outcomes that would follow from several possible counterfactual actions under a common realization of exogenous randomness. The ordinary Markov decision process formalism allows one to reason about the marginal law of each action but discards dependence across these counterfactual outcomes. The Joint Markov decision process (JMDP) formalism preserves that dependence. Prior work established the formalism and solved the fixed-policy joint moment evaluation problem in JMDPs. This paper develops optimal-control methods. We define a nonparametric distributional Bellman optimality operator for JMDPs, and prove that when the induced marginal MDP has a unique optimal policy, its iterates converge in Wasserstein distance to the optimal joint return law. For the first two moments, we establish convergence under a weaker condition that permits several mean-optimal actions as long as their tie resolutions share a second-moment fixed point. We also derive sampled targets for neural approximation.

AI 总结：本研究提出一种基于强化学习的含噪定位下的宏动作拓扑导航方法，通过以物体为中心的位姿估计替代真实位姿，在Habitat模拟器中实现智能体仅用视觉完成3D公寓内的导航任务。

链接：https://arxiv.org/abs/2608.23055

机构：Institute of Neural Computation, Ruhr University Bochum（鲁尔大学波鸿分校神经计算研究所）

作者：Simon Hakenes, Tobias Glasmachers

英文摘要：Navigating large, photorealistic 3D apartments from raw pixels is widely considered infeasible for plain reinforcement learning. We build an agent that does it anyway, estimating its own pose from the camera alone. The agent has to reach several target objects in sequence, and their positions change between episodes, so it must explore to find them. It builds on our earlier object-centric topological controller, which still read the agent's true pose and its object detections from the simulator. Here we replace that true pose with an onboard, object-centric estimate. For each object we keep a bank of ORB features that, when the object is seen again, yield a rough pose measurement, which a minimal Extended Kalman Filter (EKF) fuses with a motion model. As on a real robot, the executed motions are noisy. The estimate drifts, but the agent and the nearby objects drift together, so a locally consistent pose is enough to follow each short edge and then home in visually on the target, which lets us replace full SLAM with a much smaller model, closer to how biological navigation appears to work. In the photorealistic Habitat simulator, the agent reaches its target objects from vision alone, with a pose that only needs to be locally consistent.

AI 总结：该研究探究线性逆问题扩散模型的后验信息动力学，通过平滑似然力建立I-MMSE恒等式，经理论推导与实验验证条件作用对信息分配的影响及零空间轨迹特性。

链接：https://arxiv.org/abs/2608.21709

机构：ZJU-UIUC Institute, Zhejiang University（浙江大学ZJU-UIUC学院）； Zhejiang University（浙江大学）

作者：Xiangming Meng

英文摘要：Diffusion models are widely used as priors for linear inverse problems, yet endpoint quality does not reveal when measurement information enters reverse denoising or how it is allocated across signal directions. We study this process through the smoothed likelihood force, the difference between exact posterior and prior scores at each noise level. For a fixed measurement, its expected squared norm gives both posterior--prior relative-entropy dissipation and reverse-path relative-entropy growth. Averaging over measurements yields an information--minimum mean-square error (I-MMSE) identity linking information gain to denoising-error reduction. Under finite second moments, the force energy and its ratio to prior-score energy decay quadratically in the noising kernel's signal coefficient at high noise. Solvable models show that conditioning removes class separation already explained by the measurement, reduces a uniform index entropy over \(n\) empirical samples from \(\log n\) to \(H(I\mid r)\), and makes assimilation depend on operator--prior alignment even for identical singular values. Experiments in models with tractable posteriors evaluate these predictions. In a separate illustration with a frozen FFHQ model, masks sharing the same spectrum yield different prior-normalized null-space trajectory statistics.

AI 总结：针对拓扑事件序列因果发现中并发事件因果关系难捕捉、对不完整序列鲁棒性差的问题，提出RCCD方法，经实验验证其在电信网络数据集上优于现有SOTA方法。

链接：https://arxiv.org/abs/2608.21815

机构：School of Software Technology, Dalian University of Technology（大连理工大学软件学院）； School of Computer and Information Engineering, Jiangxi Normal University（江西师范大学计算机与信息工程学院）； School of Computer Science and Technology, Dalian University of Technology（大连理工大学计算机科学与技术学院）； School of Computer Science and Artificial Intelligence, Liaoning Normal University（辽宁师范大学计算机科学与人工智能学院）； University of Natural Resources and Life Sciences (BOKU)（自然资源与生命科学大学（BOKU））； Hangzhou Institute of Technology, Xidian University（西安电子科技大学杭州研究院）； College of Computing & Data Science, Nanyang Technological University（南洋理工大学计算与数据科学学院）

作者：Jiyu Tian, Junhao Dong, Mingchu Li, Lingling Fang, Liming Chen, Andreas Holzinger, Zheng Yan, Yew Soon Ong

英文摘要： Causal discovery on topological event sequences is crucial for ensuring the reliability of networks. However, existing methods struggle to capture the complex causal relationships arising from concurrent events and lack robustness to incomplete event sequences. To address these issues, we propose a resilient concurrent causal discovery method, termed RCCD, enabling robust learning of causal graphs from topological event sequences. Specifically, we first introduce an influence-aware hyperedge causal attention mechanism, which incorporates event duration into the embedding representation, aggregates concurrent event features via hyperedge causal convolution, and injects network prior knowledge to capture the complex many-to-one causal interactions. Furthermore, we design a masked-based alternating causal optimization framework, which forces the model to recover masked event types based on context through self-supervised mask reconstruction, thereby enhancing the resilience of the predictor to missing data. To validate the effectiveness of our method, we conduct extensive experiments on both simulated and real-world telecommunication network datasets. Experimental results demonstrate that the proposed method significantly outperforms existing state-of-the-art methods in both accuracy and robustness, making it more suitable for real-world telecommunication network environments.

AI 总结：本文提出基于变分推断的新方法，可同时推断潜在聚类与因果结构，经合成及真实数据集验证了其在聚类与因果发现上的有效性。

链接：https://arxiv.org/abs/2608.22212

机构：Indian Institute of Technology Hyderabad（印度理工学院海得拉巴分校）； RIKEN Center for AI Project（理化学研究所人工智能项目中心）

作者：Avni Rajpal, Anubhav Kumar, Rishabh Karnad, Mohammad Emtiyaz Khan, P.K. Srijith

英文摘要：Causal discovery aims to understand the relationships between individual random variables. In many applications, such as brain imaging and climate modeling, it is more meaningful to consider interactions among groups of variables. Existing methods assume that knowledge of such groups or clusters is explicitly available when modeling interactions. However, in practice, these clusters as well as the causal relationships among them, are latent. In this paper, we present a novel approach based on variational inference to simultaneously infer both the latent clusters and causal structures. We learn an approximate posterior over clusters and graph-structure by considering variational distributions based on categorical and Bernoulli models respectively. We derive variational lower bounds and estimation techniques to learn variational and model parameters. The effectiveness of our proposed methods for cluster and causal discovery are demonstrated on both synthetic and real data sets.

AI 总结：本研究提出CT-PoreFlow模型，构建QSGS-Transient-7606基准，实现瞬态二维孔隙尺度流动的高效预测，并将其嵌入逆向设计工作流，在多孔介质生成设计中表现优于扩散模型。

链接：https://arxiv.org/abs/2608.22235

机构：School of Civil Engineering, The University of Sydney（悉尼大学土木工程学院）； Zhejiang University（浙江大学）； College of Civil Engineering and Architecture, Zhejiang University（浙江大学建筑工程学院）； School of Civil Engineering, Southeast University（东南大学土木工程学院）； School of Civil and Environmental Engineering, UNSW Sydney（新南威尔士大学悉尼校区土木与环境工程学院）

作者：Yiming Wang, Jiale Zhu, Zhichen Ye, Yandong Lv, Shiqi Wang, Jinlong Liu, Yucheng Fan

英文摘要：Pore-scale flow governs transport and permeability behaviour in porous media engineering applications, yet repeated lattice Boltzmann method (LBM) simulation across many geometries and design queries remains costly for repeated deployment. This study formulates transient pore-scale flow prediction as a geometry-conditioned query-time operator and introduces QSGS-Transient-7606, a benchmark of 7,606 two-dimensional porous structures each paired with 30 logarithmically sampled LBM states. The proposed continuous-time pore-scale flow surrogate model (CT-PoreFlow) integrates topology-aware geometry encoding, compressed spectral mixing, and log-time conditioning with a late-time flux-calibration objective. On unseen test geometries, CT-PoreFlow achieves a velocity relative L2 of 0.2248 and a terminal permeability error of 12.81%. Frozen morphology and computed tomography image audits confirm reasonable cross-geometry robustness without fine-tuning. The surrogate is then embedded in an inverse design workflow, screening 9,216 generative adversarial network and diffusion candidates across 18 property targets prior to LBM verification. Guided GAN sampling attains 98.11% through-connectivity and 72.28% conditional design success, exceeding diffusion-based generation. The framework unifies transient flow prediction, transport-aware screening, and LBM-verified inverse design for porous media.

AI 总结：该研究构建StocBench基准，对比测试多种生成模型与少步方法在随机流体流动概率预测中的性能，明确不同模型在随机与确定性任务、不同推理预算下的表现差异。

链接：https://arxiv.org/abs/2608.22309

机构：School of Computation, Information and Technology, Technical University of Munich（慕尼黑工业大学计算、信息与技术学院）

作者：Sebastian Pfister, Benjamin Holzschuh, Nils Thuerey

英文摘要：We benchmark transport-based generative models as well as distillation-based few-step methods for the probabilistic forecasting of stochastic fluid flows, with a particular focus on performance under limited inference budgets. All methods are evaluated on a two-dimensional Kolmogorov flow with stochastic forcing. We measure one-step distributional accuracy against large simulated reference ensembles and assess whether the invariant measure is preserved during autoregressive rollouts via the enstrophy spectrum. On the stochastic task, flow matching achieves the most accurate one-step conditional distribution at high inference budgets, while the second-order exponential integrator DPM-2 is strongest at very low NFE. Few-step distillation methods are competitive with the multi-step methods and preserve the enstrophy spectrum particularly well. A deterministic control task, in which the forcing over the prediction interval is observed, separates aleatoric from epistemic uncertainty. Model performance does not translate between the two settings: the distilled models are competitive on the stochastic task but least accurate on the control task. While stochastic diffusion samplers such as DDPM better preserve the enstrophy spectrum during rollouts in the stochastic setting, deterministic samplers such as DDIM and DPM-2 show better spectral preservation in the deterministic setting.

AI 总结：提出结合流映射优化的高斯过程学习（GPL-FMR）两阶段框架，用于动力系统参数估计，可在观测稀缺有噪时提升估计精度，经多个数值示例验证有效。

链接：https://arxiv.org/abs/2608.22324

机构：Suzhou University of Science and Technology（苏州科技大学）； Xi’an Jiaotong-Liverpool University（西交利物浦大学）

作者：Yue Hao, Dongwei Ye

英文摘要： Parameter estimation is a central task in data-driven learning of dynamical systems. It aims to recover the underlying physical parameters from observed time-series data, thereby providing interpretable insights into the physical mechanisms governing the system. Gradient/derivative matching methods based on Gaussian process provide an efficient way to perform parameter estimation. Those methods avoid repeated numerical integration and enforce local derivative consistency. However, such local matching may result in global inconsistency with the governing flow map, particularly under scarce and noisy observations. To address this limitation, we propose a framework based on Gaussian process learning with flow map refinement (GPL-FMR), a two-stage parameter estimation framework. The first stage is based on Gaussian process learning algorithm and the posterior obtained from which is transferred as an informative prior to the second stage based on flow-map refinement. The second stage further improves the parameter estimation via optimisation based on global dynamical constraints. We demonstrate and analyse its performance on multiple numerical examples, including the Van der Pol oscillator, the Lotka-Volterra model, and the Lorenz-63 system. The results show that the proposed framework consistently improves parameter estimation accuracy, particularly under scarce and noisy observations.

AI 总结：本研究提出一种基于代理模型性能估计的两阶段检索-生成流水线，用于解决汽车引擎盖内板满足KPI要求的逆向设计问题，该流水线基于公开资源构建并作为交互式工具部署，揭示了代理模型误差与类内信号比值对流水线有效性的关键影响。

链接：https://arxiv.org/abs/2608.22457

作者：Sudeep Chavare

英文摘要：An inner hood panel must meet a deflection target, stay below a stress limit, and hit a mass target. Machine-learned surrogates have made the forward direction, geometry to performance, fast and routine. The inverse direction, producing geometry from a stated requirement, remains largely unaddressed for industrial parts whose design space is organized into discrete topology families rather than a continuous parameterization. This work presents a two-stage pipeline for that inverse problem. A reachability stage determines which topology families can satisfy a given requirement vector. A conditional variational autoencoder then generates point-cloud geometry within a selected family, and a neural-operator surrogate estimates the performance of each candidate. The pipeline is built entirely from public data and freely available compute, and is deployed as an interactive tool. The pipeline works, with qualifications that are reported as primary findings rather than caveats. The surrogate is accurate in aggregate, but its error is comparable to the performance differences it is asked to discriminate, which bounds what can be claimed for any individual generated design. That ratio of surrogate error to within-class signal is argued to be the quantity that determines whether a pipeline of this kind can work at all.

AI 总结：该研究针对序列数据分布偏移检测难题，提出基于概率流常微分方程与最大均值差异的在线检测方法，可检测任意分布偏移且无需参数假设。

链接：https://arxiv.org/abs/2608.22807

机构：National Research University Higher School of Economics（国立研究型大学高等经济学院）； The University of Sydney Business School（悉尼大学商学院）； CIREQ（蒙特利尔数量经济学研究中心）； CEBDA（应用经济学与数据分析研究中心）

作者：Artem Kraevskiy, Artem Prokhorov

英文摘要：A rapidly growing range of sequential data tasks, such as identifying trend reversals in financial markets, auto-segmenting video and audio recordings, detecting changes in movement direction from motion sensors cannot be fully addressed without detection of distributional shifts in time-ordered data. We consider a sequential change-point detection problem where the conditional density switches at an unknown time, yet neither the pre- nor post-change distribution admits a closed-form. Classical likelihood-ratio statistics are inapplicable in this settings. A conditional diffusion model, trained on pre-change-point data with a frozen context encoder, defines a deterministic bijection via the probability flow ODE. Pre-change observations are mapped onto standard Gaussian latent variables. Post-change observations, processed through the same frozen map, deviate from this reference. We employ the Maximum Mean Discrepancy as the test statistic, derive closed-form expressions for its components under the Gaussian null, and establish its asymptotic distribution as a degenerate U-statistic. Afterwards we apply an online detection procedure of Shiryaev--Roberts to the resulting statistic with exact threshold calibration. The method detects arbitrary distributional shifts, including covariance rotations and higher-order structural breaks, without parametric assumptions on either regime.

AI 总结：本研究提出理论依据充分、查询高效的DIME框架，可针对扩散模型开展成员推断攻击，在多数据集上的性能优于现有攻击，还提出并评估了对应防御措施。

链接：https://arxiv.org/abs/2608.22824

机构：University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）

作者：Tue Do, Daniel Alabi

英文摘要：Membership inference attacks expose whether individual records were used to train a model, yet existing attacks on diffusion models are largely heuristic and can require substantial query budgets. We introduce DIME (Denoiser Ideal Membership Error), a theoretically grounded and query-efficient framework for membership inference on diffusion models. Our starting point is an exact characterization of the optimal diffusion denoiser for a finite training set, which reveals that membership leakage is governed by the denoiser's implicit reconstruction error. This error decomposes into two complementary signals: a bias term, capturing reconstruction accuracy, and a previously unexplored local crowding term, capturing the geometry of nearby training examples. Both admit efficient estimators using only model queries, yielding a practical attack with as few as two queries. Across CIFAR-10/100, STL10-U, CelebA, and ImageNet, DIME consistently outperforms prior attacks at comparable or substantially lower query cost, improving TPR at 1% FPR by up to $3\times$; remarkably, its two-query variant can outperform existing 30-query baselines. Finally, we suggest, discuss, and evaluate specific defenses to counteract such powerful membership tests.

AI 总结：ProxyFormer通过双流代理架构缓解超长上下文与高分辨率生成的注意力二次增长瓶颈，可扩展可训练序列长度至0.7M，在多针检索任务中保持高准确率，初步验证了其在图像生成中的可行性。

链接：https://arxiv.org/abs/2608.23463

作者：Zhongpan Tang

英文摘要：The quadratic growth of attention computation and key-value (KV) cache with respect to sequence length is a central bottleneck for ultra-long-context language models and high-resolution generative models. We propose ProxyFormer, a general dual-stream architecture built upon proxy tokens. In each layer, fine-grained local features are compressed bottom-up into a small set of proxy states; expensive global interactions are performed only in the compressed proxy space; the globally contextualized proxies are then decompressed and injected top-down back into the local stream. Because the local stream persists across layers, fine-grained information that is not captured by one compression step remains accessible for later refinement, alleviating the irreversible information loss of conventional one-shot compression. We further introduce factorized multi-level compression/decompression, layer-wise dynamic compression ratios, asymmetric dual embeddings, and a proxy-only KV-cache inference scheme. On a 16GB GPU with batch size 1, a standard decoder-only model can train sequences of only about 20K tokens, whereas ProxyFormer with a compression ratio of 64 extends the trainable sequence length to about 0.7M. A model trained with a 64K window retains 92%-95% retrieval accuracy on a multi-needle retrieval task with 1,048,576 tokens, and a model trained with an 8K window exceeds 94% accuracy when extrapolated to 256K tokens. Preliminary image-generation experiments demonstrate the feasibility of ProxyFormer for both pixel-space and latent-space flow matching.

AI 总结：该研究针对均匀与重掩蔽离散扩散模型，提出带留一法去噪器的一阶自适应采样器，证明其采样复杂度由目标分布固有依赖结构决定，而非环境维度，实验验证了维度自适应行为。

链接：https://arxiv.org/abs/2608.23554

机构：University of Pennsylvania（宾夕法尼亚大学）； the Wharton School（沃顿商学院）； Department of Statistics and Data Science（统计与数据科学系）

作者：Daniil Dmitriev, Zhihan Huang, Yuting Wei

英文摘要：Discrete diffusion models offer a promising alternative to autoregressive generation by enabling parallel updates, but their sampling efficiency can depend strongly on the choice of the forward process and the sampler. For the uniform forward process, existing lower bounds for the standard $\tau$-leaping sampler scale linearly with the ambient dimension $d$, raising the question of whether this dependence is intrinsic to the forward process. We answer this question in the negative. We consider a first-order sampler based on the leave-one-out denoiser for uniform and remasking processes whose coordinate updates can be performed in parallel. In both cases, the sampler can correct denoising mistakes during the sampling process, which becomes necessary when many coordinates are updated together. Our main result establishes an adaptive sampling guarantee: up to logarithmic factors, $N = O(\mathrm{DTC}(X_0) / \varepsilon)$ discretization steps suffice to achieve sampling error $O(\varepsilon_{\mathrm{score}}+\varepsilon)$, where $\varepsilon_{\mathrm{score}}$ is the error in score estimation. Thus, the sampling complexity is governed by the intrinsic dependence structure of the target distribution, as measured by its dual total correlation $\mathrm{DTC}(X_0)$, rather than directly by the ambient dimension $d$. Our analysis proceeds through a Bayes-optimal auxiliary sampler that separates discretization error from score-estimation error. We also derive an exact information-theoretic representation of the discretization error in terms of the mutual information between different coordinates of the forward process at different times. This representation applies to general forward processes and, in the uniform and remasking cases, can be controlled by $\mathrm{DTC}(X_0)$. Numerical experiments on structured synthetic distributions illustrate the predicted dimension-adaptive behavior.

AI 总结：研究将Litman提出的边缘耦合扩展到重球和Nesterov动量，证明其临界点可表征梯度下降的不动点与两点轨道，海森矩阵可表征稳定性，且其与对称Verlet作用等同，确立了稳定性边缘与离散力学的联系。

链接：https://arxiv.org/abs/2608.21660

机构：Yale University（耶鲁大学）

作者：Eric Regis

英文摘要：When discrete-time optimizers operate at the edge of stability, they exhibit near-two-periodic behavior. These oscillatory dynamics are reminiscent of conservative systems, such as the dynamics generated by symplectic integrators. However, a precise formulation of the connection between discrete-time optimizers at the edge of stability and discrete mechanics remains underexplored. Recently, Litman introduced the "edge coupling": a functional on consecutive gradient descent iterates whose critical points encode the fixed points and two-point orbits of the gradient descent dynamics. Here we extend the edge coupling to heavy-ball and Nesterov momentum. We show that its critical points characterize the fixed points and two-point orbits, with its Hessian characterizing their stability. We also show that the edge coupling can be identified with the symmetric Verlet action, formalizing the connection between the edge of stability and discrete mechanics.

AI 总结：针对多模态边缘推理的通信开销与延迟问题，提出CST框架，仅传输辅助设备与主设备表征互补的稀疏特征，在低特征传输占比下实现最优任务性能，且端到端推理速度最高提升4.27倍。

链接：https://arxiv.org/abs/2608.22115

机构：University of Arizona（亚利桑那大学）； Old Dominion University（奥多明尼昂大学）； InterDigital（InterDigital公司）

作者：Hai Chi, Junrui Zhang, Rui Ning, Chonggang Wang, Robert Gazda, Huanrui Yang, Hongyi Wu

英文摘要： Collaborative multimodal inference improves edge perception by combining observations from distributed sensing devices, but transmitting high-dimensional helper representations incurs substantial communication overhead and can lead to high end-to-end latency. Existing communication-efficient methods reduce payloads through compression, semantic coding, or feature selection, yet typically optimize compactness or task relevance without explicitly accounting for information already represented at the main device. Consequently, task-relevant but redundant helper features may still consume bandwidth. We present Collaborative Selective Transmission (CST), a main-directed query--response framework that retrieves only helper information complementary to the current main representation. Inspired by Partial Information Decomposition and the Multiview Redundancy Assumption, CST learns sample-adaptive, helper-specific sparse retrieval supports while discouraging retrieval of semantics already covered by the main device or duplicated across helpers. During inference, the main device transmits only support indices, and each helper returns the corresponding latent values, avoiding dense helper-feature exchange. Across three real-world multimodal sensing benchmarks, CST transmits no more than 14.18% of helper feature values while achieving best or near-best task performance among the evaluated methods. Experiments on a five-node NVIDIA Jetson Orin Nano testbed across 5--100 Mbps demonstrate up to a $4.27\times$ speedup over Transmit-All in end-to-end inference, confirming practical end-to-end latency reductions.

AI 总结：BLADE是一种受约束的双层框架，通过钳制熵遗忘损失、非对称增广拉格朗日和LoRA适配器双层结构解决LLM遗忘的鲁棒性问题，在三类基准上性能优于基线，且可应对规模扩大和连续遗忘步骤。

链接：https://arxiv.org/abs/2608.22557

机构：The Pennsylvania State University（宾夕法尼亚州立大学）； American University（美利坚大学）

作者：Md Toufikuzzaman, Ahmad Mousavi, Dongwon Lee

英文摘要：Existing LLM unlearning methods struggle with robustness: unbounded forget losses degrade model coherence, fixed-weight balancing cannot adapt as retain difficulty shifts mid-training, and methods that work on one benchmark falter under scaling or repeated application. We propose BLADE, a constrained bilevel framework whose three mechanisms give smooth, predictable control over the optimization landscape: a clamped-entropy forget loss whose gradient is exactly zero once a token reaches sufficient uncertainty; an asymmetric augmented Lagrangian that permanently ratchets retain protection after any violation; and a bilevel structure confined to LoRA adapters that repairs retain damage before each forgetting step. BLADE dominates across three benchmark families, improving average composite scores over the strongest baselines by $6$% on TOFU, $9$% on MUSE Books, and $7$% on KnowUndo, and it remains stable under $4\times$ scaling and $4$ sequential unlearning steps on MUSE News where the best competing method collapses entirely.

AI 总结：针对产后抑郁症预测的可解释性、稳定性与类别不平衡问题，研究提出SAGE系统，结合局部XAI与GA-ANN，在766人队列中仅用16个特征实现87.96%准确率，性能优于基线方法，可作为医疗资源有限场景的PPD早期识别工具。

链接：https://arxiv.org/abs/2608.22809

机构：Brunel University London（伦敦布鲁内尔大学）； Technical University of Darmstadt（达姆施塔特工业大学）； American International University-Bangladesh（孟加拉国国际大学）； Dhaka University of Engineering and Technology（达卡工程技术大学）； George Mason University（乔治梅森大学）

作者：Md. Rokon Islam Emon, Syed Shariar Alam Shuvo, Shahriar Siddique Ayon, Abdullah Al Mamun, Ahnaf Atef Choudhury

英文摘要：Postpartum depression (PPD) poses a major burden on maternal and child health, especially in low- and middle-income countries where prevalence exceeds 19%. Despite advancements in machine learning for PPD prediction, current approaches are limited by opaque global explanations that lack clinical usefulness at the patient level, unstable feature selection, and poor generalization under class imbalance. We propose SAGE, a Stability-Aware Graph-Based Ensemble feature selection system that incorporates both local explainable AI and a genetically optimized artificial neural network (GA-ANN). Using a primary cohort of 766 postpartum women, SAGE combines information-theoretic relevance, PCA-based structure, and graph-based interactions with bootstrap stability weighting to identify robust and non-redundant predictors. The GA-ANN architecture, optimized using a genetic algorithm and enhanced with GAN based oversampling, achieved strong performance with 87.96% accuracy, 86.32% F1 score, and 0.88 AUC using only 16 features, outperforming baseline and other feature selection methods. Psychological and socioeconomic factors such as EPDS score, PHQ-9 score, feelings about motherhood, and abuse history are the main predictors, while demographic factors have less influence. The LIME-based explanations allow instance-based insight into selected features from the graph, enabling personalized risk assessment. The findings make SAGE a scalable, interpretable, and clinical tool for early identification of PPD in health-care limited resources.

AI 总结：针对联邦多目标优化问题，提出融入动量驱动梯度估计器的方差缩减算法，其收敛速率优于现有方法，经基准实验验证有效。

链接：https://arxiv.org/abs/2608.22945

机构：College of Mathematics and Statistics, Chongqing Jiaotong University（重庆交通大学数学与统计学院）； School of Science, RMIT University（皇家墨尔本理工大学理学院）； School of Mathematics, Yunnan Normal University（云南师范大学数学学院）

作者：Yong Zhao, Chunlin You, Minh N. Dao, Zai-Yun Peng

英文摘要：Federated learning has traditionally been formulated as a single-objective optimization problem, primarily focused on maximizing model utility. In real-world applications, however, machine learning models often need to optimize multiple and potentially conflicting objectives simultaneously. This motivates federated multiobjective optimization (FMOO), which provides a natural framework for jointly handling multiple task-specific objectives in federated learning. In this paper, we propose a momentum-based variance-reduced algorithm for federated multiobjective optimization. The method incorporates a momentum-driven gradient estimator into the local updates to reduce the variance of stochastic updates, leading to an improved convergence rate. We establish theoretical guarantees showing that the expected Pareto stationarity measure of a randomly selected output iterate decays at a rate of $\mathcal{O}(T^{-2/3})$, improving upon the $\mathcal{O}(T^{-1/2})$ rates established for existing methods such as FSMGDA and FedCMOO. Numerical experiments on federated multiobjective optimization benchmarks demonstrate the effectiveness and competitive performance of the proposed algorithm.

AI 总结：本文分析带初始正则化的随机梯度下降（SGDIR），推导其平方损失下的无维数上界，建立匹配的下界，通过实例级比较证明其期望超额风险不大于岭回归，数值实验验证理论结果。

链接：https://arxiv.org/abs/2608.22953

机构：ESCP Business School（ESCP商学院）

作者：Nabil Kahalé

英文摘要：We analyze a variant of stochastic gradient descent with initial regularization (SGDIR) and derive dimension-free upper bounds on its expected excess risk for the squared loss. In the noiseless case, we obtain new bounds for both averaged and non-averaged SGDIR under moment, source, and capacity assumptions. For a particular value of the source parameter, these bounds are of order $m^{-2}\log^{2}m$, where the number of training samples is of order $m$. For another value of the source parameter, we obtain, for any $\epsilon>0$, bounds of order $m^{-3+\epsilon}$, provided that the capacity parameter exceeds $\epsilon^{-1}$. We also establish a lower bound that matches our upper bounds in certain regimes up to a polylogarithmic factor. In the noisy case, we provide an instance-based comparison between SGDIR and ridge regression. Under general assumptions and a mild lower bound on the regularization parameter, we show that the expected excess risk of SGDIR is no larger than that of ridge regression, up to a polylogarithmic factor. Numerical experiments on synthetic and real data are consistent with our theoretical findings.

AI 总结：本文构建训练权重矩阵的物理响应-记忆模型，解释Muon优化器半正交化方向的有效性，提出Bi-Maxwell优化器，其双时间尺度记忆核可减少大型语言模型训练步数。

链接：https://arxiv.org/abs/2608.22994

机构：School of Physics Science and Engineering, Tongji University（同济大学物理科学与工程学院）

作者：Yinze Hu, Hongjun Xiang, Xingao Gong, Hongyu Yu

英文摘要：Training large language models is costly. How low a loss the same compute can ultimately reach depends on how each step's gradient is converted into a weight update; the rule that performs this conversion is the optimizer. From SGD and AdamW to the recent Muon, effective update rules have mostly been shaped by engineering intuition and then selected on benchmarks. Muon semi-orthogonalizes the momentum matrix before applying the update and has kept breaking records on public training benchmarks; yet why the semi-orthogonalized direction works, and over how long a history the momentum should average, are two questions at present answered mainly by experience. Here we treat the weight matrix during training as a responsive medium with memory and build a physical model for it, in which both questions find answers: the semi-orthogonalized direction is the maximally dissipative response under an output-side safety budget, which explains why it works; momentum is the internal stress accumulated by the medium; how long it should average is set by the relaxation of this stress, and a real medium relaxes on more than one timescale, the simplest form being one fast and one slow. On this basis we propose the Bi-Maxwell optimizer. The framework further yields a testable consequence: gradient directions change fast early in training and more slowly later, so the optimal memory length should grow with training stage; step-by-step measurements of a proxy for it by a read-only probe across 8 independent training trajectories are consistent with this consequence. Replacing the memory kernel alone, from a single timescale to two, brings training to the target loss in noticeably fewer steps on a public large-language-model optimizer benchmark.

AI 总结：该研究针对半监督域适应场景中现有方法缺乏不确定性量化的问题，提出将共形风险最小化与最优传输结合的端到端框架，生成紧凑且覆盖有效的预测集。

链接：https://arxiv.org/abs/2608.23153

机构：Duke University（杜克大学）

作者：Manos Giannopoulos, Yi Shen, Michael M. Zavlanos

英文摘要：In high-stakes healthcare applications, machine learning models are frequently trained on data from one patient population and deployed on another, creating a distribution shift that degrades both accuracy and reliability. Semi-Supervised Domain Adaptation (SSDA) addresses this by leveraging labeled data from some source domain to improve model performance on a target domain where labels are scarce. However, existing SSDA methods optimize primarily for point-prediction accuracy and offer no principled uncertainty quantification --- a prerequisite for clinical trust. Conformal Prediction (CP) can address this limitation by providing prediction sets with rigorous, distribution-free coverage guarantees. However, applying CP post-hoc to a pre-trained model can yield prohibitively large prediction sets, as SSDA pre-training methods do not account for the nonconformity score geometry that determines conformal set size. Conformal Risk Minimization (CRM) has been used to resolve this issue in the fully supervised setting by integrating the CP objective directly into model training, but it requires a large labeled dataset to compute nonconformity thresholds during training, precisely the data that is scarce in the SSDA regime. We propose an end-to-end framework that integrates CRM into the SSDA training objective, enabling effective CRM in the limited-labeled-target-data regime. The key idea is to utilize Optimal Transport (OT) to generate pseudolabels for unlabeled target instances, providing the additional training signal needed by CRM to operate using only a small labeled target set. This results in a model jointly optimized for domain invariance and conformal efficiency, producing prediction sets that are compact, coverage-valid, and support domain-specific constraints such as excluding mutually contradictory diagnoses in skin lesion classification.

AI 总结：该研究针对KV缓存压缩的假设缺陷，提出无需训练的压缩器CertKV，在多个长上下文基准测试中表现优异，明确注意力可压缩性的决定因素。

链接：https://arxiv.org/abs/2608.21541

机构：City University of Hong Kong（香港城市大学）； University of New South Wales（新南威尔士大学）

作者：Chiwun Yang, Xiaoyu Li

英文摘要： KV-cache compression is often justified by attention maps with a few large weights. This is incomplete: large weights may not contain most of the mass, omitted values can cancel, and preserving the attention output may not preserve the task. We separate these questions. Global score gaps -- not threshold counts -- determine how many tokens are needed to retain a target mass. For a realized row, the weighted sum of omitted values is the exact missing statistic. A controlled retrieval--aggregation model explains when truncation helps and when it hurts. These results motivate CertKV, a training-free compressor that reserves one tail-summary slot per head and allocates the rest by value dispersion. Under matched budgets, CertKV is top-two in seven of nine LongBench-v2 settings, remains in the leading compressed tier on 128K RULER, and realizes a ten-fold cache budget in a packed Llama prototype. Compressibility depends on the mass, values, future queries, and task -- not on a sparse-looking map alone.

AI 总结：FrugalSOT是面向设备端NLP的资源感知模型选择架构，通过自适应阈值匹配模型复杂度，在树莓派5上可显著降低推理时间与资源消耗，同时保障输出相关性。

链接：https://arxiv.org/abs/2608.21621

机构：Vellore Institute of Technology（韦洛尔理工学院）

作者：Pradheep P, Yuvanesh S, Harish KB, Keerthan Saai Reddy S, Joshva Devadas T, Naveenkumar J, Hemalatha K

英文摘要：In on-device NLP tasks, limited resources of embedded hardware, such as the Raspberry Pi 5, require efficient inference strategies. This paper introduces FrugalSOT (Frugal Search Over The Models), a resource-aware model selection architecture for on-device NLP inference. FrugalSOT estimates each request's complexity by extracting features such as prompt length, named entity density, and syntactic complexity. The request is first made to the least complex model that is likely to pass a relevance threshold. If the output of that model falls short of the threshold, the request is made to a more complex model. It is important to note that the relevance threshold undergoes continuous updates in the background. using past validation outcomes in an adaptation process using a low-pass filtering mechanism, thus imparting adaptation to changing input patterns. Experimental results achieved on a Raspberry Pi 5 show that FrugalSOT reduces average inference time and overall computational resource use to a significant extent compared to a single-model baseline approach, without compromising output relevance to the same extent as the most sophisticated model. These results confirm that adaptive model selection can enable efficient, high-quality natural language processing inference on limited devices.

AI 总结：本文探究注意力基础对学习型KV缓存驱逐软-硬转换的影响，通过2×2×2对比实验发现，Sigmoid注意力经学习型硬驱逐后，其门控模型删除KV条目时PPL变化可忽略，且比H₂O、KeyDiff等方法表现更优。

链接：https://arxiv.org/abs/2608.23296

机构：University of Pittsburgh（匹兹堡大学）

作者：Isaac (Rucheng)Li

英文摘要：Learned KV-cache eviction often faces a soft-to-hard mismatch: during training, differentiable gates typically attenuate token contributions, whereas inference saves memory only when KV entries are physically removed. We ask whether the attention substrate affects this soft-to-hard transition. Using GPT-2-scale Transformers trained on OpenWebText, we run a controlled $2\times2\times2$ comparison over attention type, learned gating, and positional encoding. Although sigmoid attention is worse as a dense language model, learned hard eviction changes the useful operating points: sigmoid-gated models delete KV entries with negligible PPL change relative to their own no-eviction references. Under a matched live-cache protocol on the same dense backbones, learned sigmoid gates obtain lower PPL than our H$_2$O and KeyDiff implementations, whereas softmax gates do not uniformly beat these post-hoc methods. The results suggest that attention normalization can substantially affect whether a training-time soft gate transfers cleanly to hard KV deletion.

AI 总结：本研究针对供应链市场波动下的预测需求，提出FEF NCL分布式方法，在合成数据集上实现了预测误差降低、延迟风险宏观F1提升及高波动误差减少的效果，为供应链预测提供了可行方案。

链接：https://arxiv.org/abs/2608.21399

作者：Shunmukha Sagar Puppala

英文摘要：Supply chain forecasting systems increasingly operate under market shocks, non-identically distributed regional demand, and limited willingness to centralize commercial data. This work proposes Federated Ensemble Forecasting with Negative-Correlation Learning (FEF NCL), a distributed method that trains specialized forecasting experts across client nodes while discouraging redundant model errors. The framework combines temporal feature encoders, client level drift scoring, reliability-weighted aggregation, and an explain ability layer that exposes the market and supplier variables most responsible for each forecast. A single synthetic dataset is used to evaluate the design. It contains 124,800 weekly SKU region observations from ten regional client nodes, 60 product families, 40 suppliers, five commodity groups, and a 2021-2024 volatility profile with explicit price-shock regimes. Because the dataset is synthetic, the reported results should be interpreted as controlled evidence of internal consistency rather than real-world validation. Across the synthetic test split, FEF NCL reduces weighted mean absolute percentage error from 13.9% for the best federated baseline to 12.4%, improves delay-risk macro-F1 from 0.755 to 0.801, and lowers the high volatility quintile error by 2.1 percentage points relative to SCAFFOLD. The analysis suggests that negative-correlation specialization is useful when clients face different supplier, freight, and commodity conditions, although deployment would require stronger privacy analysis, live drift monitoring, and operational calibration. Index Terms federated learning, ensemble learning, negative correlation learning, supply chain forecasting, market volatility, data drift, demand planning, risk governance

AI 总结：本研究针对拆分式联邦微调（SFF），识别出深度-性能困境，评估多种联邦适配器聚合方法后发现其无法缓解拆分架构缺陷，归因于Transformer拓扑，为分布式LLM微调提供结构基础。

链接：https://arxiv.org/abs/2608.22188

机构：Vellore Institute of Technology（韦洛尔理工学院）； University of Arizona（亚利桑那大学）

作者：Hariharan Ramesh, Someshwaran Murugaiyan, Jyotikrishna Dass

英文摘要：Split Federated Fine-tuning (SFF) is a promising paradigm for scaling Large Language Models (LLMs) by partitioning model depth between resource-constrained clients and a centralized server. While system incentives for throughput and privacy favor deep partitions, the impact of such configurations on model utility remains poorly understood. In this work, we identify and characterize the Depth-Performance Dilemma: the regime that maximizes system efficiency is precisely where fine-tuning quality collapses. Through a comprehensive audit across four model scales (GPT-2 to Llama-3-8B) and diverse benchmarks, we demonstrate that deeper partitions provide monotonic gains in throughput and privacy at the cost of catastrophic performance plateaus. We evaluate a suite of state-of-the-art federated adapter aggregation methods including AVG, STACK, SVD, and FREEZE, revealing that while these techniques are effective in standard Federated Learning, they fail to mitigate the artifacts unique to split architectures. Finally, we provide a mechanistic diagnosis for this failure, tracing the collapse to the near-isometric topology of Transformers, which allows aggregation noise to propagate without attenuation until it triggers Attention Collapse in the server partition. Our findings challenge the prevailing assumption that partition depth is a utility-neutral tuning knob and provide a structural foundation for stable distributed LLM fine-tuning.

AI 总结：该研究针对现有大型语言模型遗忘基准的缺陷，提出扩展WMDP的WMDP++基准，以针对性提取遗忘信息和评估边界问题性能，实现对遗忘算法更严格的压力测试。

链接：https://arxiv.org/abs/2608.22527

作者：Noam Diamant, Ethan Fetaya, Neta Glazer

英文摘要：Recently, machine unlearning, the removal of specific training data influence from a model, has gained increasing attention. In large language models (LLMs), unlearning is particularly challenging due to the ambiguity of inputs and outputs. Con- sequently, rigorous evaluation is critical for assessing both safety and utility, and for driving progress in unlearning meth- ods. We identify two key shortcomings in existing unlearning benchmarks: (1) they do not actively test whether unlearned information can still be forcibly extracted, and (2) they fail to evaluate performance preservation on boundary questions, be- nign queries that are semantically close to the unlearned con- tent. Here we introduce WMDP++, an extension of WMDP that addresses these gaps by incorporating targeted extrac- tion of unlearned information and systematic evaluation on boundary questions. WMDP++ provides a more stringent and informative benchmark for evaluating unlearning in LLMs.

AI 总结：该研究提出GCA协议，通过潜在代码介导的联邦学习，无需传输AE参数或梯度，可防御数据提取攻击，降低通信开销，提升测试准确率。

链接：https://arxiv.org/abs/2608.22593

作者：Jong-Ik Park, Harry Jiang, Logan Blakely, Georgios Fragkos, Shamina Hossain-McKenzie, Carlee Joe-Wong

英文摘要：Autoencoder (AE)-based federated learning (FL) is attractive for anomaly detection when clients have limited local data. However, conventional FL exchanges AE parameters or gradients, incurring substantial communication overhead and potentially exposing input training data information, since AEs are explicitly optimized to reconstruct their inputs. We introduce \emph{Global Centroid Alignment (GCA)}, a latent-code-mediated FL protocol that coordinates clients without transmitting AE parameters or gradients. In each round, (1) clients first train their local AEs using a \emph{reconstruction} update and upload a small subset of encoder latent codes to the FL server. (2) The server pools these codes, fits a clustering model, and broadcasts only \emph{global latent centroids and their support counts}. (3) Each client then updates its encoder by aligning its local latent codes with the \emph{nearest} centroid using \emph{inverse-count} weighting to emphasize globally underrepresented patterns. Steps (1)--(3) repeat over communication rounds. Because GCA exchanges only sampled latent codes and centroid statistics, its communication cost depends on latent dimensionality and the numbers of uploaded codes and returned centroids rather than on AE model size. Across five tabular and two vision benchmarks, GCA yields higher reconstruction error under a server-side client data extraction attack in all 21 comparisons and clearly lower cosine similarity in 20 of 21 comparisons with FedAvg, FedProx, and FedNova, showing its ability to protect training data. It even improves test accuracy over FedAvg by up to $5.76\%$. GCA achieves extraction defense comparable to DP-FedAvg, remains effective when DP-FedAvg does not reduce target resemblance, and lowers per-round communication by up to $99.15\%$.

AI 总结：针对基于蒸馏的联邦学习中标签分布偏差导致的性能下降问题，提出FedCC算法，允许客户端将模糊样本标记为“未知”并结合校准伪标签，在极端标签偏差场景下准确率大幅优于基线方法。

链接：https://arxiv.org/abs/2608.23031

机构：Huawei Heisenberg Research Center, Huawei Technologies Duesseldorf GmbH（华为海森堡研究中心（华为技术有限公司杜塞尔多夫分部））； TUM School of Computation, Information and Technology, Technical University of Munich（慕尼黑工业大学计算、信息与技术学院）

作者：Wenxuan Ye, Onur Ayan, Xueli An, Georg Carle

英文摘要： Federated Learning (FL) enables distributed clients to collaboratively train models without sharing raw data, making it promising for leveraging massive devices in communication networks. In distillation-based FL, each client applies its local model on an unlabeled public dataset, and shares only prediction results with the server. While heterogeneous local data introduces label distribution skew, thus biasing client models toward majority classes and leading to potentially inaccurate predictions. The lack of ground-truth labels in the public dataset hampers the server's ability to calibrate predictions, which ultimately degrades overall performance. To address this, we propose FedCC, a simple and effective algorithm for mitigating client misclassification. Instead of being forced to classify and risking error propagation, clients are allowed to tag ambiguous samples as 'unknown'. This additional class, together with calibrated pseudo-labels on the public data, balances confidence in majority classes against uncertainty in under-represented ones. Extensive experiments demonstrate that FedCC significantly outperforms existing methods, especially under severe label skew. In the extreme scenario where each client holds samples from only one of ten classes, FedCC achieves 67.3% accuracy, while baselines collapse to near-random results.

AI 总结：该研究针对联邦学习中更新掩码的正则化留存问题，证明了客户端掩码设置与联邦平均对正则化代价的影响，在CIFAR-10上开展实验并分析了相关因素的作用。

链接：https://arxiv.org/abs/2608.23286

机构：Faculty of Science & Technology Sophia University（索菲亚大学科学技术学院）； Technical R&D Department Shendian Energy Co., Ltd.（申电能源有限公司技术研发部）

作者：Wenhao Yan, Fu Kuroda, Yucheng Jin, Zhenke Chen

英文摘要：Federated learning on non-IID data seeks flat minima to generalize across clients, and existing methods borrow sharpness-aware minimization from centralized training. There is a second way to reach flat minima, in which the regularization comes for free from noise added to the parameter updates, and it has never been carried over to the federated setting. We show the reason. Masking charges the optimizer for moving in sharp directions. We prove that when each client draws its own mask, federated averaging weakens that charge by exactly the cohort size, and that giving every client the same mask brings it back by a factor equal to the inverse gradient diversity of the cohort. In our experiment setting on CIFAR-10, that factor is 1.19 out of a possible 10. Turning off minibatch sampling raises it to 8.96, while changing data heterogeneity a hundredfold leaves it between 1.17 and 1.50. The configurations keeping the regularization train far too poorly to use.

AI 总结：该研究构建KAN-Robust-Bench基准，探究KAN模型在FGSM等规避攻击下的鲁棒性，给出随机平滑等方法的数学基础，以寻找最优防御策略与架构。

链接：https://arxiv.org/abs/2608.21488

机构：University of New Brunswick（新不伦瑞克大学）

作者：Mohammad Meymani, Roozbeh Razavi-Far

英文摘要：While machine learning models have demonstrated strong performance in many domains, these models have shown profound vulnerabilities when they are exposed to adversarial threats. While adversarial attacks fall into various categories, the most prominent category in research studies is evasion. In evasion attacks, the adversary generates perturbed versions of samples, which might not be observable by human eyes. These samples generally fool the machine learning models with high confidence. This phenomenon poses a significant security violation against machine learning models. In this paper, we investigate the certified and empirical robustness of various Kolmogorov-Arnold network architectures against strong evasion attacks. At first, we provide the mathematical foundations for randomized smoothing and interval bound propagation, and report the $\ell_2$-certified robustness of the models under randomized smoothing. After that, we systematically evaluate the robustness of various defended and undefended KAN models under FGSM, PGD, and C&W attacks in order to find out the optimal defense strategies and architectures.

AI 总结：本文针对长宏观经济序列数据稀缺下的多期罕见事件预测难题，提出基于校准多样性的策略，将六个月覆盖率从67.8%提升至81.4%，并量化了诚实评分下六个月覆盖率能否达90%的待解问题。

链接：https://arxiv.org/abs/2608.21591

机构：University of Moratuwa（莫拉图瓦大学）

作者：Donald Aadithiyan

英文摘要：Multi-horizon rare-event forecasting is hard under long macroeconomic series' data constraints: labeled events are scarce, and standard uncertainty quantification assumes an exchangeability that autocorrelation violates. A controlled ablation shows an apparent rare-event threshold for Adaptive Conformal Inference instead reflects calibration-set size. Across 200 random calibration sets, support width of the nonconformity-score distribution explains up to 85% of coverage variance versus 2% for rare-event count; the same, not the same magnitude, replicates across synthetic conditions and five countries (five-country Spearman $\rho$ 0.45-0.66 vs. 0.02-0.23). A diversity-maximizing selector built on this is the only strategy tested that improves long-horizon coverage (67.8% to 81.4% at six months); Mondrian, shift-robust, and extreme-value alternatives fail to close it. Mondrian even worsens coverage under oracle labels. A compact proposition explains why: coverage deficit reflects how closely the calibration set's upper quantile reaches the test distribution's. Diversity is necessary, not sufficient. Demonstrated on a two-stage U.S. recession-forecasting framework with RegressorChain, whether six-month coverage reaches 90% under honest scoring remains open, a question this paper quantifies rather than resolves.

AI 总结：本研究提出GeoQ框架，用于科学代理模型的非侵入式误差估计，经多类任务验证，其几何感知条件分位数建模可实现感知有效性的误差估计。

链接：https://arxiv.org/abs/2608.21652

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

作者：Khoa Nguyen, Daniel Serino, Aviral Prakash, Marc Klasky

英文摘要： Neural-network surrogate models are increasingly used to accelerate scientific simulations, but their deployment in extrapolative and autoregressive settings requires input-dependent estimates of prediction error. In this work, we introduce GeoQ (Geometry-Aware Conditional Quantile Error Estimation), a non-intrusive calibration framework for estimating surrogate error at individual query points. GeoQ represents the error at a query point as an anchor-averaged calibration error plus a learned nonnegative correction. This correction is modeled as an upper conditional quantile of the anchor-relative error increment, using geometry-based features that encode representation-space displacement and local support density. A cross-fitting procedure generates approximately out-of-sample calibration tuples, while a feature-space k-nearest-neighbor support score identifies regions \textcolor{black}{where the learned error model is supported by calibration data}. We evaluate GeoQ on scalar regression, chaotic dynamics, medium-range weather forecasting, and Richtmyer-Meshkov instability prediction. The results demonstrate that geometry-aware conditional quantile modeling provides a practical and non-intrusive approach for validity-aware error estimation in scientific surrogate models.

AI 总结：该研究针对对抗训练为何减少叠加的问题，基于Ilyas等人的特征分类法，通过实证解释揭示其因果链，为理解对抗样本与叠加的关系提供了新视角。

链接：https://arxiv.org/abs/2608.22155

作者：Adam Elimadi

英文摘要：The study of adversarial examples and their origins remains an open area of research. Mechanistic interpretability, and superposition in particular, offers new avenues for approaching this problem. Gorton & Lewis (2025) demonstrate that adversarial examples arise from superposition and show empirically that adversarial training reduces superposition, yet provide no mechanistic account of why this occurs. We present an empirical explanation inspired by the feature taxonomy of Ilyas et al. (2019), tracing the following chain of causalities: adversarial training abandons non-robust features, leading to fewer total features to represent, resulting in less superposition.

AI 总结：DeMixPert通过分解扰动响应为系统、扰动特异性及种群变化成分，结合高斯原型可逆网络建模，在未见过扰动的OOD单细胞预测任务中取得优异性能。

链接：https://arxiv.org/abs/2608.23114

作者：Jiawen Liu, Xuechenxiao Cao, Yutong Li, Bing Liu, Jiaming Liang, Tinghe Zhang, Xiaoqi Sheng, Hongmin Cai

英文摘要：Predicting transcriptome-wide responses to unseen genetic perturbations remains a major computational challenge because accurate prediction requires recovering both perturbation-specific transcriptional shifts and heterogeneous cellular responses. Existing methods often entangle deterministic response structure with stochastic population-level variation, causing dominant shared patterns to mask weaker perturbation-specific signals and impair distributional modeling. To address these challenges, we propose \textbf{DeMixPert}, an approach for Decomposed response Modeling with Gaussian Mixtures for Out-Of-Distribution (OOD) single-cell Perturbation prediction. DeMixPert decomposes perturbation-induced changes into a basal-state-dependent systematic response, a perturbation-specific response, and population-level variation. The systematic component is derived from the basal state encoded from control-cell expression, whereas the perturbation-specific component is inferred from pretrained target embeddings for unseen-target generalization. DeMixPert models population-level variation using a Gaussian prototype Invertible Network and adaptively combines reusable Gaussian prototypes according to the basal state and perturbation condition. The resulting mixture is mapped to a condition-specific variation distribution. Sampled variations are integrated with the systematic and perturbation-specific components, followed by joint decoding with the basal state to reconstruct perturbed-cell gene expression. Experimental results show that DeMixPert effectively captures heterogeneous single-cell perturbation responses and achieves superior performance across unseen-perturbation settings. The source code is made publicly available upon publication.

AI 总结：本研究针对大规模5G网络PCI分配难题，提出带神经块求解器的同余分解框架，将问题解耦为子问题并结合图神经网络优化，实验显示其性能优于现有基线方法。

链接：https://arxiv.org/abs/2608.21485

机构：The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））； Shenzhen Research Institute of Big Data（深圳大数据研究院）； Columbia University（哥伦比亚大学）； Sun Yat-sen University（中山大学）； Tongji University（同济大学）； University of Chinese Academy of Sciences（中国科学院大学）； Huawei Technologies（华为技术有限公司）

作者：Yeqing Qiu, Chengpiao Huang, Ye Xue, Akang Wang, Fan Xu, Zhipeng Jiang, Dong Zhang, Ruoyu Sun, Qingjiang Shi, Zhi-Quan Luo

英文摘要：Physical Cell Identity (PCI) assignment is essential for interference management in dense 5G networks. As cellular networks scale, PCI reuse becomes unavoidable, which may cause collisions, confusions, and multiple forms of modular interference. Jointly mitigating these effects gives rise to a large-scale, multi-objective combinatorial optimization problem that is difficult to solve efficiently at practical network scales. In this work, we propose a congruence decomposition framework with neural block solvers for large-scale PCI assignment. The proposed decomposition exploits the arithmetic structure of PCI values to decouple multiple modular interference objectives into a collection of blockwise Min-$k$-Partition subproblems, followed by a graph coloring procedure to resolve PCI conflicts. For the resulting NP-hard Min-$k$-Partition subproblems, we develop neural block solvers by parameterizing their relaxed quadratic formulations with graph neural networks, enabling efficient optimization at large scales. Discrete assignments are recovered through conditional expectation rounding with theoretical guarantees. Experiments on synthetic cellular graphs and real-world 5G networks show that the proposed method consistently outperforms existing modular-interference-aware baselines in modular interference reduction, conflict elimination, and computational efficiency.

AI 总结：针对文本属性图小样本节点分类中现有方法LLM信息利用不均、成本高的问题，提出置信感知双教师学习框架CoTeach，动态为节点选更可靠教师，提升性能并降低LLM成本。

链接：https://arxiv.org/abs/2608.22127

机构：Chungbuk National University（忠北国立大学）； Chungnam National University（忠南国立大学）； The Pennsylvania State University（宾夕法尼亚州立大学）

作者：Hojin Kim, Sujin Yoon, Sungsu Lim, Dongwon Lee, David Yoon Suk Kang

英文摘要：Text-Attributed Graphs (TAGs) integrate graph structures and node-associated textual attributes, and recent studies have increasingly leveraged Large Language Models (LLMs) to improve TAG learning in few-shot settings. However, existing approaches typically utilize LLM-derived information uniformly across all nodes, despite substantial variations in its reliability, while also incurring considerable monetary costs. We argue that the most appropriate source of supervision may differ across nodes, as Graph Neural Networks (GNNs) and LLMs exhibit complementary strengths in exploiting structural and semantic information, respectively. To this end, we propose CoTeach, a Confidence-aware dual-teacher learning framework that dynamically selects the more reliable teacher for each node. Experimental results demonstrate that CoTeach consistently improves few-shot node classification performance while reducing unnecessary LLM utilization and associated monetary costs.

AI 总结：针对野外可穿戴与智能手机情感识别的标签难题，该研究采用自监督学习结合多任务归纳式图神经网络的子图训练方法，在K-EmoPhone数据集上取得唤醒度、效价任务的准确率提升。

链接：https://arxiv.org/abs/2608.22387

机构：Khalifa University of Science and Technology（哈利法科技大学）； Aristotle University of Thessaloniki（亚里士多德大学塞萨洛尼基分校）

作者：Ioannis N. Ziogas, Leontios J. Hadjileontiadis, Ahsan H. Khandoker, Aamna Al Shehhi

英文摘要：Wearable and smartphone-based emotion recognition (WER) remains a challenging setting in affective computing, due to the notorious difficulty and bias associated with in-the-wild label collection. The high inter-and intra-subject emotional variability motivates us to explore WER modeling through graph node classification in a limited resources learning scheme powered by Self-Supervised Learning (SSL) graph masking augmentation tasks. We employ a subgraph sampling approach during training, utilizing labeled and unlabeled data, along with supervised, semi-supervised, and SSL mechanisms in a multi-task inductive graph neural network architecture. Our evaluations on K-EmoPhone through leave-one-group-out cross-validation in the binary arousal and valence tasks yield average accuracy gains of 4.3% and 7.8%, compared to the full resource setting, utilizing only 20% and 25% of the labels, respectively. Our model analysis sheds light on the relation of SSL graph augmentations to emotional arousal and valence and justifies the approach of SSL-driven subgraph training for in-the-wild WER.

AI 总结：ReCoG是将图结构学习与多模态表示学习紧密耦合的新范式，在节点分类和链接预测基准中优于多模态图结构学习基线，证明了结构与语义协同演化的重要性。

链接：https://arxiv.org/abs/2608.22786

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

作者：Rui Xue, Tianfu Wu

英文摘要：Multimodal graph learning requires jointly training over graph structure and heterogeneous node attributes, yet existing methods largely decouple these processes: prior multimodal graph neural networks (GNNs) focus on aligning modalities in a shared embedding space while operating on fixed or weakly adapted graph structures, and graph structure learning approaches infer topology from unimodal node representations without accounting for multimodal interactions. This separation fundamentally limits the ability of GNNs to capture semantically meaningful relationships in multimodal settings, where observed edges are often noisy, incomplete, or misaligned with underlying semantics. We propose ReCoG (Reciprocal Co-Evolution for Multimodal Graph Learning), a new learning paradigm that tightly couples graph structure learning and multimodal representation learning through end-to-end reciprocal interaction. Concretely, ReCoG integrates (i) a multimodal graph refiner that infers and corrects edges using cross-modal semantic evidence, and (ii) a coupled cross-modal message passing mechanism that performs joint intra- and inter-modality propagation over the refined graph. This unified design yields greater expressiveness than decoupled or two-stage formulations and allows dynamic interaction between topology and representation learning. Across diverse benchmarks for node classification and link prediction, ReCoG consistently outperforms strong multimodal graph structure learning baselines, including graph foundation models. Our results demonstrate that reciprocal co-evolution of structure and semantics is important for effective multimodal graph learning, challenging the prevailing separation between topology and representation learning.

AI 总结：该研究在AIFB、CoDEx、BioKG数据集上，结合GNN编码器与盒嵌入语义损失，验证了层级感知语义损失能显著提升知识图谱链接预测的MRR，且效果优于额外子类边的方法。

链接：https://arxiv.org/abs/2608.22981

机构：Chalmers University of Technology（查尔姆斯理工大学）； University of Gothenburg（哥德堡大学）； University of Cambridge（剑桥大学）

作者：Filip Kronström, Ross D. King

英文摘要： Knowledge graphs are often accompanied by ontological class hierarchies that encode valuable semantic information, yet many link prediction methods either ignore such hierarchies or incorporate them indirectly through additional graph edges. Recent work introduced hierarchy-aware graph neural networks (GNNs), which use semantic losses derived from box embeddings to encourage satisfaction of subclass relationships during GNN-based representation learning. While this approach has shown promise for biological regression tasks, its effectiveness for knowledge graph link prediction has not been investigated. In this paper we evaluate hierarchy-aware semantic losses on link prediction across three benchmark datasets: AIFB, CoDEx, and BioKG. We combine graph neural network encoders with box-embedding-based semantic losses that encourage learned representations to better satisfy ontology-derived class hierarchies, and compare this approach to both standard link prediction models and models incorporating subclass relations as graph edges. Across all datasets, hierarchy-aware semantic losses significantly improve mean reciprocal rank (MRR) and consistently outperform models that incorporate hierarchy information through additional subclass edges. Relative to the baseline GNN models, MRR improved by 7.6%, 2.4%, and 15.5% on AIFB, CoDEx, and BioKG, respectively. Furthermore, semantic losses consistently outperform the alternative of augmenting the graph with subclass edges. These results are consistent with ontology-derived class hierarchies providing complementary information to graph structure, and suggest that encouraging hierarchical consistency through semantic losses is an effective and comparatively parameter-efficient mechanism for improving knowledge graph link prediction.

AI 总结：针对SIMON、SIMECK等轻量级IoT密码的差分聚类识别空白，采用ML引导的GRL方法，提取pDDT特征构建有向图，KNN模型表现最优，框架可推广至其他相关LCAs家族。

链接：https://arxiv.org/abs/2608.23054

机构：School of Computing, Mathematics and Engineering, Charles Sturt University（查尔斯斯特大学计算、数学与工程学院）

作者：Jonathan Cook, Sabih ur Rehman, M. Arif Khan

英文摘要：SIMON and SIMECK belong to a family of Lightweight Cryptographic Algorithms (LCAs) based on the Feistel block cipher, designed for Internet of Things (IoT) devices. As with all Feistel ciphers, they are susceptible to differential cryptanalysis, necessitating rigorous resilience evaluations. While state-of-the-art techniques leverage heuristics and sampling to improve efficiency, little work has applied Machine Learning (ML) guided Graph Representation Learning (GRL) to efficiently identify and visualise high-probability differential clusters. We address this gap by introducing an efficient feature engineering strategy that extracts four differential attributes from a partial Difference Distribution Table (pDDT), revealing structural information concealed in raw differential data. Utilising the enriched features, we construct and compare three ML-guided directed graphs for SIMON$32$ and SIMECK$32$ using K-Nearest Neighbour (KNN), Decision Trees (DT), and Random Forests (RF). To the best of our knowledge, our framework produces the first graph-based visualisation of the differential clustering effect, in which high-probability single-bit differentials form geometrically close clusters in the learned embedding. All three models achieve a precision of $1.0$ in identifying high-probability differentials, confirming zero false positives. KNN achieves the strongest cluster separation, the highest F1 score and the lowest graph construction time of approximately $2.3$ seconds, while DT and RF produce optimal paths with near-perfect regression. The results are consistent across both LCAs, demonstrating the applicability of the framework to other AND-rotation LCA families.

AI 总结：本研究将联邦持续学习建模为分布式漂移加惩罚控制问题，提出FedQCL框架，通过虚拟队列调控遗忘，在多基准上优于现有方法并降低异质数据下的遗忘。

链接：https://arxiv.org/abs/2608.21539

机构：IIIT Delhi（德里印度信息技术学院）； IIT Dharwad（达尔瓦德印度理工学院）

作者：Nazreen Shah, Naveen Kumar Reddy Somireddy, Zubair Shaban, Ranjitha Prasad, B. N. Bharath

英文摘要：Federated Continual Learning (FCL) is fundamental to real-world distributed learning systems, requiring models to adapt to sequential, non-IID data across clients while mitigating catastrophic forgetting and client drift. Existing approaches formulate continual learning (CL) as a sequence of per-task optimization problems, applied locally at each client and coupled through aggregation, using heuristic mechanisms such as replay, regularization, or projection-based constraints. However, forgetting in FCL is inherently a long-term, distributed phenomenon, arising from the interaction of temporal task evolution and cross-client heterogeneity, which is not explicitly regulated. In this work, we cast FCL as a stochastic control problem and propose Federated Queue-regulated Continual Learning (FedQCL), a framework based on Lyapunov drift-plus-penalty (DPP) optimization. FedQCL introduces virtual queues to track the accumulation of forgetting across tasks and clients, enabling explicit control of the stability-plasticity trade-off. By optimizing a DPP objective, the method jointly improves current-task performance while the queue-based formulation provides an interpretable and tunable mechanism to balance adaptation and retention through a single parameter, without requiring gradient projection or additional communication overhead. Empirical evaluations on standard benchmarks, including Split-CIFAR-10, Split-CIFAR-100, and Split-TinyImageNet, demonstrate that FedQCL outperforms state-of-the-art baselines with respect to accuracy while significantly reducing forgetting under heterogeneous data distributions.

AI 总结：针对持续学习下的多模态大语言模型，研究人员提出持久公平性后门攻击，通过两种机制注入持久群体歧视，该攻击能规避标准防御且在多轮持续学习中留存。

链接：https://arxiv.org/abs/2608.21577

机构：Emory University（埃默里大学）

作者：Yuyang Luo, Kai Shu

英文摘要：Multimodal Large Language Models (MLLMs) are increasingly deployed in high-stakes domains where fairness is a critical safety requirement. In practice, these models are continually updated through continual learning (CL) to adapt to evolving tasks and data distributions. Prior work has shown that backdoor attacks can manipulate MLLM responses through hidden triggers, but naively implanted backdoors degrade as models undergo subsequent updates of CL. Although fairness has emerged as a central concern for MLLM deployment, whether backdoor-induced fairness violations can survive CL remains unexplored, leaving two critical questions unanswered: (1) whether a backdoor can reliably induce fairness violations in MLLMs, and (2) whether such fairness-targeted backdoors can persist through continual learning. We bridge this gap by proposing Persistent Fairness Backdoor Attack (PFBA) to inject persistent and group-specific discrimination into MLLMs. Specifically, PFBA achieves this through two novel mechanisms. The Latent Space Fairness Reinforcement reshapes the model's deep feature geometry by anchoring privileged-group representations to preserve utility while repelling and clustering targeted-group representations to sustain discrimination, and the Continual Learning Simulation iteratively optimizes the trigger against simulated parameter drift to ensure backdoor persistence across future updates. Extensive experiments demonstrate that PFBA induces severe fairness disparities that persist across continual learning rounds, evading standard backdoor defenses. The data and code are publicly available at this https URL.

AI 总结：本研究提出SynEHR框架，通过两项创新模块优化纵向EHR合成，在真实数据集的多维度评估中优于现有最优模型，生成更具临床一致性与时间真实性的EHR数据。

链接：https://arxiv.org/abs/2608.21673

机构：Florida State University（佛罗里达州立大学）

作者：Ximiao Li, Lin Jiang, Rongchao Xu, Dahai Yu, Zhe He, Guang Wang

英文摘要：Longitudinal electronic health records (EHRs) document patients' sequences of clinical visits over time, preserving the temporal evolution of disease progression and care delivery. However, real longitudinal EHRs are difficult to access because they contain large amounts of fine-grained, patient-specific information. Synthetic EHR generation therefore provides a valuable approach for preserving the statistical patterns and clinical structure of patient visit trajectories, enabling broader modeling and analysis when real records are limited. Although recent generative models have made progress in producing future visit sequences, they remain limited in explicitly integrating inter-visit irregular temporal evolution and intra-visit clinical event structures in EHRs, leading to clinically inconsistent and temporally unrealistic visit sequences. In this work, we propose SynEHR, a lightweight adaptive LLM-based framework for longitudinal EHR synthesis. There are two novel designs in SynEHR, i.e., a Temporal State Conditioning Module captures irregular temporal states across visits and a Temporal-Relational Adaptation Module combines these states with patient history to dynamically construct patient-specific relational representations. SynEHR then builds on a parameter-efficient LoRA-adapted language-model generator with next-visit generation capability to train the two modules for temporally and clinically informed generation. Extensive experiments on real-world EHR datasets across fidelity, privacy, and downstream utility evaluations demonstrate that SynEHR outperforms state-of-the-art models by generating more clinically coherent and temporally faithful longitudinal EHR data.

AI 总结：该研究针对临床视觉大语言模型的缺陷，提出BioMed-Agent-RL医学智能体，整合多模态元学习与强化学习等技术，在多基准实验中准确率达约73%，较现有模型提升约5%，为临床智能体系统建立新标准。

链接：https://arxiv.org/abs/2608.21864

作者：Md Asaduzzaman Jabin, Zihao Wu, Tianming Liu

英文摘要：The current progress of Clinical Vision Large Language Models (C-VLLMs) has substantially improved digital diagnostics, still these frameworks often endure lesion noises, modality misalignment, hallucination, and missed contextual grounding in complex clinical cases. Moreover, prevailing agent systems usually depend on static and non-adaptable pipelines and lack the versatility necessary for complex medical reasoning. To resolve these difficulties, we present BioMed-Agent-RL, a unified medical agent that incorporates adaptive orchestration, policy, and reward-based reinforcement learning (RL) models for biomedical applications. To ensure reliability, it invokes clinical context-aware preference optimization (CPO), direct preference optimization (DPO), and group relative policy optimization (GRPO) with dynamic entropy regulation. This pipeline utilizes a multimodal meta-learning approach that operates as a field-specific expert and human judgment synthesizer. The agent adaptively utilizes a set of model-level expertise, such as clinical grounding and reasoner, lesion segmenter, and field-specific synthesizer, across various clinical modalities (e.g., X-ray) by utilizing an iterative and adaptive RL approach. The agent learns to seriously synthesize misleading, conflicting vision cues and trust in inherent reasoning, while specialist advice is faulty. An intensive ablation study is conducted across multiple benchmarks, and the agent significantly outperforms existing state of the art models, such as GPT-5, attaining up to ~73% accuracy (gain of ~5%) over contemporary baselines. As a result, the framework suggests a new standard for building factual, reliable, robust, and expert-like intelligent agent systems for independent clinical reasoning.

AI 总结：该研究针对多任务LoRA方法存在的训练-推理不一致问题，提出无路由的CD-LoRA，通过一致性驱动对齐机制提升多任务微调的稳定性与性能，优于现有多适配器基线。

链接：https://arxiv.org/abs/2608.21909

机构：School of Artificial Intelligence, Jilin University（吉林大学人工智能学院）； International Center of Future Science, Jilin University（吉林大学未来科学国际中心）

作者：Qian Zha, Jinda Liu, Yuan Wu, Yi Chang

AI 总结：针对时间域泛化中复杂流场景的局限，提出FreKoo++框架，统一连续Koopman模态动力学与自适应谱解耦，在TDG基准上实现最先进性能。

链接：https://arxiv.org/abs/2608.22224

机构：Australian Artificial Intelligence Institute (AAII)（澳大利亚人工智能研究所（AAII））； Faculty of Engineering and Information Technology（工程与信息技术学院）； University of Technology Sydney (UTS)（悉尼科技大学（UTS））

作者：En Yu, Xiaoyu Yang, Wei Duan, Guangquan Zhang, Jie Lu

AI 总结：该研究针对CIFAR-10-C开展条件级分析，对比三种测试时自适应策略与源模型的性能，发现自适应存在失效场景，揭示可靠性过滤会影响EATA的表现，强调需开展条件级评估。

链接：https://arxiv.org/abs/2608.22233

机构：Heritage Institute of Technology（遗产技术学院）

作者：Sreeja Guha Majumdar, Aratrika Saha

AI 总结：本文提出MASH-Bench基准，通过实验发现大规模枪击风险分类的跨源泛化受特征完整性和标签流行度限制，DANN可部分提升GVA的极高风险召回率，为相关研究提供受控环境。

链接：https://arxiv.org/abs/2608.22460

机构：Virginia Commonwealth University（弗吉尼亚联邦大学）

作者：Neha Sharma, Ritesh Sharma

AI 总结：本文针对设备端LLM联邦微调的高通信成本问题，提出SplitLite方法，利用残差的低秩结构压缩传输量，在保持性能的同时大幅降低了通信成本。

链接：https://arxiv.org/abs/2608.23018

机构：The University of Hong Kong（香港大学）； California Institute of Technology（加州理工学院）

作者：Tao Li, Yulin Tang, Qi Guo, Xianhao Chen

AI 总结：针对现有测试时适应（TTA）方法在ECG分类任务中忽略节拍-节律结构、易受伪影影响的问题，提出BeatRhythm-TTA框架，通过SQI门控方案与双层次一致性提升性能，在多域ECG诊断中获2.70%的Macro-F1相对提升。

链接：https://arxiv.org/abs/2608.23347

机构：Westlake University（西湖大学）； Hokkaido University（北海道大学）； RIKEN（理化学研究所）

作者：Wenhan Jiang, Zhipeng Deng, Jiale Zhou, Haolin Wang, Yafei Ou, Yefeng Zheng

AI 总结：针对资源受限场景下轻量级时间序列预测器小样本学习的困境，提出MetaCaster多智能体框架，可高效训练专用预测器，在18个数据集等实验中兼顾数据、计算效率与预测性能。

链接：https://arxiv.org/abs/2608.23473

机构：University of Houston（休斯顿大学）； NEC Labs（NEC实验室）； University of Waterloo（滑铁卢大学）； University of Connecticut（康涅狄格大学）； University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）； Singapore Management University（新加坡管理大学）

作者：ChengAo Shen, Wenchao Yu, Fangyu Wu, Dongjin Song, Hanghang Tong, Dongsheng Luo, Wei Cheng, Haifeng Chen, Jingchao Ni

AI 总结：本研究针对现有化学LLM评估基准的局限性，提出ChemDIRT基准，涵盖八大类化学任务，评估模型在指令和分子表示变化下的准确性与一致性，测试发现LLM存在提示敏感性、表示依赖及任务间性能不均问题。

链接：https://arxiv.org/abs/2608.21504

机构：University of Notre Dame（圣母大学）

作者：Eric Inae, Tim Gunn, Chris Bond, Meng Jiang

AI 总结：该研究提出固定半径距离带Shepard ρ基准，对比8种DR方法，发现现有排名指标有偏，固定半径指标可揭示聚类压缩等问题，还能解决单点及少数总体的评估难题。

链接：https://arxiv.org/abs/2608.21779

机构：toor Inc.（toor公司）

作者：Yoshio Takaeda

AI 总结：本研究定义了判断Adam停滞的损失景观相关指标，通过$2\times2$示例和FINER图像拟合案例说明Adam的对角预条件无法消除交叉耦合病态性，为分析Adam的收敛特性提供了基于预条件海森的视角。

链接：https://arxiv.org/abs/2608.22145

机构：Texas State University（德克萨斯州立大学）

作者：Rodion Podorozhny

AI 总结：本研究通过可复现、计算匹配的实验评估现代手写体热身对历史阿拉伯文OCR的影响，发现其并非普遍有效，发布了可独立验证的SaudiHeritage-OCR包。

链接：https://arxiv.org/abs/2608.22316

机构：Dal Research Team（达尔研究团队）

作者：Sumaih Almarshad, Maram Alamri, Dona Aloraini, Fares Altuwaim, AlJawharh AlOtaibi, Reem Alyabis, Rayah Aldawsari

AI 总结：CatchBench是首个在同一任务-方法接口下对智能体运行前、运行中、完成后三种信息状态进行评分的基准，涵盖多类模型与配置，揭示基准分数需结合标签过程才可解释。

链接：https://arxiv.org/abs/2608.22808

机构：University of Southern California（南加州大学）

作者：Yue Zhao

AI 总结：本研究构建了基于真实手术室录制的扩展多模态手术团队互动数据集，含多级标注与反事实标注，支持团队协作建模及AI辅助协作系统设计，为高风险领域协作行为分析提供统一资源。

链接：https://arxiv.org/abs/2608.23344

机构：University of Trento（特伦托大学）； UiT the Arctic University of Norway（挪威北极大学）

作者：Vincenzo Marco De Luca, Antonio Longa, Andrea Passerini

AI 总结：针对非工作时段商业存款的支票欺诈风险，ChequeMark多视图集成ML框架结合XGBoost、Isolation Forest与GraphSAGE生成客户风险评分，在分布偏移场景下性能优于单一模型，提升了鲁棒性与可解释性。

链接：https://arxiv.org/abs/2608.21629

机构：Royal Bank of Canada（加拿大皇家银行）

作者：Ann Youduo Xu, Emily Yu, Justin Leski, William Lam

AI 总结：本研究开发PiNN框架预测薄壁截锥壳临界屈曲载荷，结合RBD公式，经133次实验验证，其精度与物理一致性优于DNN，可用于不确定性感知的壳体结构设计。

链接：https://arxiv.org/abs/2608.21818

机构：University of California Los Angeles（加州大学洛杉矶分校）； Adani University（阿达尼大学）； Indian Institute of Technology Kanpur（印度理工学院坎普尔分校）

作者：Devasmit Dutta, Budhaditya De, Rohan Majumder, Sudip Kumar Mishra

AI 总结：针对混沌系统深度学习预测的长期误差累积问题，提出DAW框架，利用动力系统局部维度重塑损失，在KS方程上显著降低了长期自回归误差。

链接：https://arxiv.org/abs/2608.22277

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

作者：Zhou Fang, Gianmarco Mengaldo

AI 总结：针对金融欺诈检测中联邦学习结合TreeSHAP解释易受成员推理攻击的问题，提出客户端级差分隐私的DP-FedSHAP架构，权衡解释保真度、隐私与AUPRC性能。

链接：https://arxiv.org/abs/2608.22607

机构：London Metropolitan University（伦敦都市大学）； School of Computing and Digital Media（计算与数字媒体学院）

作者：Muhammad Waleed Gul, Elaheh Homayounvala

AI 总结：针对分子基础模型的化学无效增强等局限，提出Mol-JEPA多模态框架，利用模态掩码融入生化上下文，在基准测试中展现出优异性能。

链接： https://arxiv.org/abs/2608.22642

机构：University of Tübingen（蒂宾根大学）； Boehringer Ingelheim（勃林格殷格翰）； The University of Texas at Austin（德克萨斯大学奥斯汀分校）； Brown University（布朗大学）

作者：Florian Rottach, Sebastian Schieferdecker, William Rudman, Randall Balestriero, Carsten Eickhoff

AI 总结：本研究采用局部神经算子（LocalNO），基于太阳30个太阳半径处的径向速度与径向磁场，重构太阳磁流体动力学的多场多尺度内边界状态，以支撑日球层建模与太阳风预测。

链接：https://arxiv.org/abs/2608.22782

机构：Georgia State University（佐治亚州立大学）； Predictive Sciences Inc.（预测科学公司）

作者：Vignesh Kumar Pandian Sathia, Reza Mansouri, Dustin J. Kempton, Pete Riley, Rafal A. Angryk

AI 总结：本研究针对Ariel任务的光谱数据归因与误差评估需求，提出基于影响函数的可追溯谱推断方法，通过极限学习机实现高效计算，其误差代理可准确关联光谱误差，为科学机器学习提供运行框架。

链接：https://arxiv.org/abs/2608.23458

机构：ML Analytics； Delft University of Technology（代尔夫特理工大学）； King’s College London（伦敦国王学院）； European Space Agency（欧洲空间局）

作者：Nikki Grens, Luís F. Simões, Kai Hou Yip, Theresa Lueftinger

AI 总结：研究人员针对关系异常检测中展平数据丢失结构及难融入符号证据的问题，提出RAD方法，结合图表示学习与规则信号，在多基准任务中较现有方法取得更优异常检测性能。

链接：https://arxiv.org/abs/2608.23468

机构：Vanderbilt University（范德堡大学）

作者：Noah Dahle, Anne Tumlin, Ngoc Tran, Xenofon Koutsoukos, Tyler Derr

AI 总结：本文针对不平衡时间序列量化问题，提出类条件高斯混合量化器 CC-GMNet-TS，结合 Transformer 特征提取器与类专属混合模型，在三个基准上取得优于传统方法的低误差。

链接：https://arxiv.org/abs/2608.21473

机构：Texas State University（德克萨斯州立大学）

作者：Md Shahriar Kabir, Mayesha Maliha R. Mithila, Anne H. H. Ngu, Mylène C. Q. Farias, Byron Gao

AI 总结：该研究针对独立同分布N选最优搜索推导了精确认证极限，提出双门审计规则，其在数学推理和代码选择的回顾性分析中可降低保留误差。

链接：https://arxiv.org/abs/2608.21496

机构：Technical University of Darmstadt（达姆施塔特工业大学）

作者：Ricardo Fitas

AI 总结：本研究提出心音段选择算法与多输入CNN，同步分析多通道心音，在735例患者数据上实现96.5%准确率，优于单通道及异步多通道方法。

链接：https://arxiv.org/abs/2608.21499

机构：INESC TEC； Faculdade de Ciências, Universidade do Porto（波尔图大学理学院）； LIACC, Universidade da Maia（马亚大学LIACC）； Instituto Superior de Engenharia do Porto（波尔图工程高等学院）

作者：Marcelo Nogueira, Jorge H. Oliveira, Carlos F. Ferreira, Miguel T. Coimbra, Alípio M. Jorge

AI 总结：本研究针对网球领域多模态损伤风险与表现预测方法不足的问题，提出PART多模态加权集成学习框架，整合多类数据提取专属特征，采用自适应权重策略，经9名大学网球运动员数据验证，可监测健康、估算损伤风险，对休闲球员也有应用潜力。

链接：https://arxiv.org/abs/2608.21530

机构：Monmouth University（蒙茅斯大学）

作者：Weihao Qu, Dongyang Wang, Ling Zheng, Francisco E. Alvarez, Shobharani Polasa, Jiacun Wang

AI 总结：研究未知线性顺序学习问题，反例可能不真实且k未知，确定最优查询复杂度，还研究低维几何表示的目标排序情况，给出上下界并指出无噪声项的差距。

链接：https://arxiv.org/abs/2608.21579

机构：Princeton University（普林斯顿大学）； Tel Aviv University（特拉维夫大学）； Technion – Israel Institute of Technology（以色列理工学院）； Google Research（谷歌研究院）

作者：Noga Alon, Shay Moran, Shlomo Moran

AI 总结：该研究发现RSSM架构世界模型（如DreamerV3）的隐藏状态$h_t$含隐式困惑信号，经线性探针、编辑验证其因果性，该信号可在多数控制任务中泛化。

链接：https://arxiv.org/abs/2608.21582

机构：University of Moratuwa（莫拉图瓦大学）

作者：Donald Aadithiyan

AI 总结：本研究基于All of Us计划的大规模数据，用LightGBM模型结合12种生物标志物的纵向轨迹，实现了对老年人临床前功能衰退的早期预测，性能优于静态指标，可零负担整合至电子健康记录。

链接：https://arxiv.org/abs/2608.21589

机构：Indiana University（印第安纳大学）

作者：Rashmita Kudamala, Aravind V. Kuruvikkattil, Lalitha Pranathi Pulavarthy, Saptarshi Purkayastha

AI 总结：本文针对稀疏非负矩阵补全问题，提出可扩展的随机交替最小二乘算法，结合稀疏优化与CUDA内核加速，成功应用于果蝇连接组突触权重矩阵分析，揭示其潜在低秩结构的预测能力。

链接：https://arxiv.org/abs/2608.21607

机构：Flatiron Institute（弗莱堡研究所）； Courant Institute, NYU（纽约大学库朗数学研究所）； Google Research（谷歌研究院）

作者：Lawrence K. Saul, Ningyuan Huang, Dennis Bollweg, Jeff Soules, Diana C. Halikias

AI 总结：本文针对语义通信评估碎片化问题，从评估视角综述文本与图像语义通信的关键性能指标，分类梳理各类指标，分析现存挑战并展望未来研究方向。

链接：https://arxiv.org/abs/2608.21626

机构：Ericsson Canada（爱立信加拿大公司）； Concordia University（康考迪亚大学）

作者：Niloofar Tavakolian, Hakimeh Purmehdi, Jungyeon Baek

AI 总结：本文针对北马其顿17428台商用智能电表的两年用电数据，评估OWA、SoftImpute、形状建模自动编码器三种缺失值插补算法，发现OWA精度最高且稳定，SoftImpute稳定但精度低，自动编码器方差大，建议按需选算法并探索混合架构。

链接：https://arxiv.org/abs/2608.21638

机构：FEEIT Ss. Cyril and Methodius University in Skopje（斯科普里圣西里尔与美多德大学FEEIT学院）； Elektrodistribucija DOOEL EVN Group（EVN集团下属Elektrodistribucija DOOEL公司）

作者：Daniela Stojcheska, Marija Markovska, Dimitar Taskovski, Branislav Gerazov, Boris Nikolov

AI 总结：针对多任务学习中损失尺度不匹配导致同方差加权性能下降的问题，提出BPGS方法，经实验验证其在合成测试和三个真实基准上均表现出良好的尺度鲁棒性与性能。

链接：https://arxiv.org/abs/2608.21653

机构：Neryva（内里瓦）

作者：Krishna Subedi

AI 总结：针对工业模拟中神经PDE代理无法处理复杂网格的问题，提出RWR模型，通过交织潜在注意力与消息传递松弛，在基准测试中精度最高且数据效率优，可处理大规模问题。

链接：https://arxiv.org/abs/2608.21677

作者：Anuj Kumar, Heiko Zimmermann, Josiah Bjorgaard, Jacan Chaplais, Nikolaos Bouklas, Matteo Salvador, Alexander Lavin

AI 总结：针对现有扭曲序贯蒙特卡洛稀有事件估计依赖稀有正样本导致不可靠的问题，提出自适应多级扭曲SMC，通过逐步稀有中间事件学习扭曲，提升语言模型稀有不安全行为概率估计准确性，助力模型安全评估与对齐。

链接：https://arxiv.org/abs/2608.21736

作者：Zixuan Liu, Fangzheng Wu, Brian Summa, Zizhan Zheng

AI 总结：ChainPrune 是一种推理路径语义结构优化方法，通过整合推理路径为树结构、选择主导路径及结合 DPO 与监督损失，减少长思维链冗余，在保精度的同时降低了步骤长度与计算开销。

链接：https://arxiv.org/abs/2608.21860

机构：Alibaba Group（阿里巴巴集团）

作者：Weihang Pan, Zhengxu Yu, Yuxiang Zhang, Wenzhi Li, Zhongming Jin, Binbin Lin, Xiaofei He, Jieping Ye

AI 总结：针对气候降尺度中动态一致性缺失的挑战，提出DySCo框架，通过数据驱动的张弛法重构训练动态配对轨迹，训练两阶段算子，在保持统计性能的同时提升动态一致性，优于现有方法。

链接：https://arxiv.org/abs/2608.21998

机构：Massachusetts Institute of Technology（麻省理工学院）； City University of Hong Kong（香港城市大学）； Google Research（谷歌研究院）

作者：S. Stamatelopoulos, M. Wang, I. Lopez-Gomez, L. Zepeda-Nunez, Z. Y. Wan, R. Carver, F. Sha, T. P. Sapsis

AI 总结：该研究提出ARCHER模型，通过离散旋转网格建模姿态后验，实现冷冻电镜跨结构通用的零样本姿态估计，在测试结构和实验颗粒上均达到高精度，且保留下游构象信号。

链接：https://arxiv.org/abs/2608.22029

机构：Pritzker School of Molecular Engineering, University of Chicago（芝加哥大学普利兹克分子工程学院）； Department of Computer Science, Rensselaer Polytechnic Institute（伦斯勒理工学院计算机科学系）

作者：Nhan D. Nguyen, Bao Pham

AI 总结：本研究针对海上石油平台能源低效问题，采用机器学习构建柴油消耗预测模型，结合搜索算法优化发电机负载，实现日均27%的柴油消耗降低，为平台能效提升提供可行方案。

链接：https://arxiv.org/abs/2608.22076

作者：Khivishta Boodhoo, Josh Plumbly, Nicholas Watson

AI 总结：针对会话语音智能体流式ASR因每轮重置状态导致的话语起始性能下降问题，提出两种保留跨轮次话语上下文的状态管理策略，在实验中实现15-21%的相对WER降低。

链接：https://arxiv.org/abs/2608.22101

机构：University of Washington（华盛顿大学）； SRI International（国际 SRI 研究所）

作者：Sameep Chattopadhyay, Alexander Erdmann, Mari Ostendorf

AI 总结：本研究通过三项测量探究语言模型在Apple神经引擎的部署与解码速度，发现部署由计算表达而非内容决定，权重编码影响加速器使用，提出先选编码再分配参数的设计流程。

链接：https://arxiv.org/abs/2608.22110

作者：Shahir M A

AI 总结：该研究提出TANGO和WANGO两种模型，替换Transformer的自注意力与前馈子层，经对比实验，TANGO在多数据集上验证负对数似然最优，WANGO在线性复杂度架构中表现最佳且优于Recurrent Transformer++。

链接：https://arxiv.org/abs/2608.22117

机构：Luddy School of Informatics, Computing, and Engineering（勒迪信息学、计算与工程学院）； Indiana University Bloomington（印第安纳大学伯明顿分校）

作者：Joshua Nunley

AI 总结：本文提出分块稳定自适应三次正则化优化器ARC-φ₁，将参数按张量分块求解，在大参数INR任务上收敛速度与性能显著优于Adam。

链接：https://arxiv.org/abs/2608.22129

机构：Texas State University（德克萨斯州立大学）

作者：Rodion Podorozhny

AI 总结：针对工业降阶建模中的自相交轨迹问题，提出LA-NODEs框架，经理论分析与IPMSM、DES两类工业模型验证，可提升建模精度与保真度，为复杂工业系统高精度数据驱动建模提供有效方法。

链接：https://arxiv.org/abs/2608.22142

机构：Lanzhou University of Technology（兰州理工大学）； Hiroshima University（广岛大学）； Nanjing University - Suzhou Campus（南京大学苏州校区）； Princeton University（普林斯顿大学）

作者：Xiaorui Wang, Yu Zhou, Wenjie Mei, Dongzhe Zheng, Yang Bai, Masaaki Nagahara

AI 总结：该研究提出带行为调优的贝叶斯蒸馏（BBT）混合模型，结合贝叶斯与神经网络模型优势，在人类概念学习案例中更准确预测人类行为并揭示心理学洞见。

链接：https://arxiv.org/abs/2608.22154

机构：Princeton University（普林斯顿大学）； Princeton AI Lab（普林斯顿人工智能实验室）； Massachusetts Institute of Technology（麻省理工学院）

作者：Brenden M. Lake, Akshay K. Jagadish, Guangyuan Jiang

AI 总结：该研究探讨时间观测协议的反事实评估，提出特定价值的识别理论，结合模拟和真实数据集分析，为未部署协议的观测设计提供支撑。

链接：https://arxiv.org/abs/2608.22221

作者：Xizhe Zhang

AI 总结：该研究将语言模型头部与softmax作为整体推导更新几何，基于Muon的奇异值条件提出最大化最小行间距并约束直径的方法，为V≫d时的近似等距问题提供了思路。

链接：https://arxiv.org/abs/2608.22253

机构：Columbia University（哥伦比亚大学）

作者：Aditya Somasundaram

AI 总结：该研究针对内存高效优化器的状态异质性提出自适应对数空间量化，在AdamW、CAME等模型的语言模型预训练实验中，可显著降低优化器存储并保持良好困惑度表现。

链接：https://arxiv.org/abs/2608.22322

作者：Yan Wang

AI 总结： 针对Delta规则循环模型在极端上下文外推下的局部范数爆炸问题，提出SANE方法，在安全阈值下可稳定长序列推理且性能不下降，过松阈值会导致推理能力丧失。

链接：https://arxiv.org/abs/2608.22354

机构：College of Science, Shantou University（汕头大学理学院）； Yuanshi Intelligence（元氏智能）

作者：Qingwen Lin, Boyan Xu, Xiao Liu, Zhifeng Hao, Ruichu Cai

AI 总结：该研究提出跨变量代理学习方法，基于M-JEPA在无降水监督下预训练，实现了优于ECMWF集合的季风降水预测，为大气表示迁移提供诊断框架。

链接：https://arxiv.org/abs/2608.22358

作者：K M Anirudh, S Sandeep, Hariprasad Kodamana

AI 总结：该研究提出基于和乐的图曲率离散化框架，引入两种聚合机制实现曲率相关响应的规范等变更新，经单位球校准验证了方法的有效性。

链接：https://arxiv.org/abs/2608.22453

机构：National University of Singapore（新加坡国立大学）； Nanyang Technological University（南洋理工大学）； Nankai University（南开大学）

作者：Hao Li, Yuhan Peng, Junwen Dong

AI 总结：本研究通过受控消融实验，探究将PQCs作为残差边载模块集成到ResNet-50中用于神经解码的效果，发现骨干梯度训练投影的量子边载模型可提升准确率，测量引导训练能改善表征几何，且未实现量子计算优势。

链接：https://arxiv.org/abs/2608.22475

机构：Reichman University（赖赫曼大学）； Dartmouth College（达特茅斯学院）； The Hebrew University of Jerusalem（耶路撒冷希伯来大学）

作者：Diana Legziel Levy, Menachem Finkelstein, Peter Chin, Eilon Vaadia, Sarel Cohen

AI 总结：该研究针对神经网络在有限域运算等价表示间的迁移问题，提出训练模型预测伽罗瓦作用以构造轨道规范代表的方法，实现了乘法运算向未见过的基表示的迁移。

链接：https://arxiv.org/abs/2608.22513

机构：Towson University（陶森大学）

作者：Zheng Zhang, Na Zhang

AI 总结：本文提出DIVE框架，通过在优化层面反转有害样本梯度方向，将有害数据转化为有益贡献，可提升分类性能、最大化数据效率并稳定优化，还能泛化到大语言模型微调。

链接：https://arxiv.org/abs/2608.22522

机构：Brandeis University（布兰迪斯大学）

作者：Adrian Nyakairu, Hongfu Liu

AI 总结：该研究针对临床知识图谱构建难题，提出结合多智能体关系提议等技术的Clinical Graph-JEPA框架，在MIMIC-IV数据集上验证其提升了留一边缘恢复的MRR。

链接：https://arxiv.org/abs/2608.22583

机构：New York University（纽约大学）； University of Chicago（芝加哥大学）

作者：Kushagra Yadav, Nalin Prabhath, Amit Lamba, Goeun Han, Yining Mao

AI 总结：该研究探究表格基础模型TabPFN v3是否可用于非表格任务，将MNIST手写数字识别、法德语言识别、Tiny ImageNet图像分类转为表格形式，其在部分任务上达到了对应专用模型的相当准确率。

链接：https://arxiv.org/abs/2608.22594

机构：Technische Universität Dresden（德累斯顿工业大学）； Maynooth University（梅努斯大学）； Universität Würzburg（维尔茨堡大学）

作者：Goran Nakerst, John Brennan, Wouter Beugeling, Masudul Haque

AI 总结：该研究构建力学驱动的对抗智能体评估拓扑优化中深度学习代理模型的脆弱性，发现初始噪声会引发机械失效，物理梯度调节未必提升鲁棒性，SIMP优化器可部分恢复性能，为鲁棒生成式设计智能体训练提供基础。

链接：https://arxiv.org/abs/2608.22606

作者：Hoang Anh Nguyen, Yuan Hong, Hongyi Xu

AI 总结：该研究利用基于星系图像训练的类LLM模型AstroPT作为测试平台，探究星系属性的出现顺序，为校准应用于LLMs的机械可解释性方法提供了受控沙盒。

链接：https://arxiv.org/abs/2608.22614

机构：IUCAA（印度大学间天文和天体物理中心）； AstroAI； Washington University in St. Louis（圣路易斯华盛顿大学）

作者：UniverseTBD: Kshitij Duraphe, Aman Kumar, Michael J. Smith, Shashwat Sourav

AI 总结：KMGen是首个端到端框架，可全自动提取KM曲线并生成合成患者不良事件轨迹，在多肿瘤试验中精度达标，相关流水线已开源。

链接：https://arxiv.org/abs/2608.22618

机构：Mayo Clinic Alix School of Medicine（梅奥诊所阿利克斯医学院）

作者：Jalen Jiang, Chufan Gao, Ethan Rasmussen, Stephen Z. Xie, Jimeng Sun

AI 总结：该研究提出MOSH-WM模型，通过掩码支撑构建软哈密顿动力学，在OBJ3D、CLEVRER数据集的视频预测任务中，相比基线显著降低误差，且误差积累更慢。

链接：https://arxiv.org/abs/2608.22750

作者：Zhekai Wang, Haoxiang Huang, Xiang Liu, Zhikang Chen, Yueqing Sun, Qi Gu, Shiji Zhou, Miao Liu, Sen Cui

AI 总结：本文提出LpWM模型，以稀疏表示替代密集表示建模动作条件潜在动力学，在PushT任务上规划成功率优于密集模型，且能揭示可解释的动力学结构。

链接：https://arxiv.org/abs/2608.22764

机构：NYU（纽约大学）； Duke University（杜克大学）； Mila（米拉）； Brown University（布朗大学）

作者：Yilun Kuang, Yash Dagade, Quentin Le Lidec, Lucas Maes, Randall Balestriero, Yann LeCun

AI 总结：针对黑盒LLM在生物多样性监测分类推理中置信度估计难的问题，提出层级感知监督估计器，提升了微AUROC，验证了层级结构对弃权规则的重要性。

链接：https://arxiv.org/abs/2608.22839

机构：Vector Institute for AI（向量人工智能研究所）； University of Guelph（圭尔夫大学）

作者：Shuting Xie, Nathaniel Lesperance, Graham W. Taylor

AI 总结：该研究提出ADAPT框架，可单次蒸馏生成跨后训练大语言模型的多规模多变体模型，实现平滑规模插值与自适应规模选择，优化长文本推理的计算-准确率权衡。

链接：https://arxiv.org/abs/2608.22854

机构：Harvard University（哈佛大学）； IST Austria（奥地利科学技术研究所）

作者：Yan Zhou, Sara Kangaslahti, Jonathan Geuter, Nihal V. Nayak, Marco Fumero, Francesco Locatello, David Alvarez-Melis

AI 总结：该研究提出Mapping the Concept Landscape框架，以样本级图刻画语义概念全局分布，开发贪心概念覆盖最大化算法，实现更优数据剪枝效率并提供可解释的审计轨迹。

链接：https://arxiv.org/abs/2608.22858

机构：Huazhong University of Science and Technology（华中科技大学）； Ant Group（蚂蚁集团）； Chinese Academy of Sciences（中国科学院）

作者：Dongyue Wu, Tao Ma

AI 总结：本研究针对注意力、状态空间及混合序列模型，提出无需训练的轻量级前缀不变性审计方法，可精准定位因果关系破坏点，且在测试中成功发现Zamba2等模型的缺陷。

链接：https://arxiv.org/abs/2608.22876

机构：VIDRAFT AI Research（VIDRAFT AI研究机构）

作者：Taebong Kim, Youngsik Hong, Minsik Kim, Sunyoung Choi, Jaewon Jang, Minseo Kim

AI 总结：该研究指出约束型N选1推理时缩放中存在安全黑客行为问题，即不安全但可行的输出会随采样数增加而被优先选择，通过理论推导和实验验证了该问题的存在及固有难度。

链接：https://arxiv.org/abs/2608.22915

机构：LY Corporation（LY公司）； University of Tsukuba（筑波大学）； RIKEN AIP（理化学研究所人工智能项目）

作者：Akifumi Wachi, Takumi Tanabe, Youhei Akimoto

AI 总结：本研究通过三类工业监测场景对比评估时间序列基础模型与轻量基线的性能及成本，发现TSFMs并非拟合轻量模型的默认替代，而是任务依赖的部署选项。

链接：https://arxiv.org/abs/2608.22968

机构：National Taiwan University of Science and Technology（台湾科技大学）

作者：Guan-Hua Wen, Kuan-Yu Chen

AI 总结：本研究提出ST$^2$U方法，将测试时遗忘转化为全轨迹范围的受限知识边界控制，在三个基准和三个模型系列上，该方法比测试时基线大幅减少受限知识重新进入，实现了保留与遗忘的良好平衡。

链接：https://arxiv.org/abs/2608.23034

作者：Xunlei Chen, Qinghui Gong, Ruini Xue, Yaodong Hu, Tian Lan, Wenhong Tian

AI 总结：本研究提出轻量级半结构化剪枝框架RoI，通过可微子集采样学习稀疏掩码，减少参数与内存开销，在Qwen2.5 LLM家族上实现高效且具竞争力的性能，为LLM高效部署提供实用路径。

链接：https://arxiv.org/abs/2608.23048

机构：Viettel AI（越南电信人工智能公司）； Viettel Group（越南电信集团）； Hanoi University of Science and Technology（河内科学技术大学）

作者：Ha Dinh, Xuan Duy Ta, Khoat Than, Khac-Hoai Nam Bui

AI 总结：针对低比特大语言模型量化的误差问题，提出AWSRC编解码器修复权重，在Qwen2.5-3B-Instruct上可显著缩小与BF16模型的质量差距，且侧载体积小、性能优于同类编解码器。

链接：https://arxiv.org/abs/2608.23144

机构：Huawei Boole RC（华为布尔研究中心）

作者：Zehao Liu, Chuangchuang Fang, Yang Ren

AI 总结：该研究针对重型车辆发动机诊断，提出训练时随机禁用模态的模态丢弃方法，在融合文本、传感器和故障码的三类模态数据上，使发动机部件分类准确率较仅用文本提升3.5个百分点，为该领域首次应用三类模态融合。

链接：https://arxiv.org/abs/2608.23161

机构：Halmstad University（哈尔姆斯塔德大学）

作者：Adeel Zafar, Slawomir Nowaczyk, Hamid Sarmadi, Saeed Gholami Shahbandi

AI 总结：该研究提出反事实转移图（CGT）以评估时间序列分类器的跨类转移质量，发现反事实可达性与分类器准确率相悖，且可对比不同反事实解释器的性能。

链接：https://arxiv.org/abs/2608.23164

机构：Otto-von-Guericke University Magdeburg（马格德堡奥托·冯·格里克大学）； Jagiellonian University（雅盖隆大学）； Halmstad University（哈尔姆斯塔德大学）； Center for Applied Intelligence Systems Research（应用智能系统研究中心）

作者：Syed Muhammad Hamza Zaidi, Szymon Bobek, Grzegorz J. Nalepa, Myra Spiliopoulou

AI 总结：本研究对比分析深度学习中无标签表示质量度量，按构造分类并建立关联，经合成实验、260种视觉模型及6个数据集验证，发现本征维度（ID）最可靠，且所有度量可靠性受架构类别和训练目标调节，为相关度量的理解与应用提供了清晰依据。

链接：https://arxiv.org/abs/2608.23182

机构：Linköping University（林雪平大学）

作者：Daniel Richards Arputharaj, Daniel Jönsson, Gabriel Eilertsen

AI 总结：该研究以伦敦监测站点为对象，通过场景式实验评估四种树基ML模型，发现纳入远程交通数据可提升NO₂估算精度，交通变量在交通主导环境中对NO₂的贡献可与相邻监测站污染物测量相当。

链接：https://arxiv.org/abs/2608.23219

机构：Queen Mary University of London（伦敦玛丽女王大学）； Instituto Politecnico Nacional（国家理工学院）

作者：Valeria Legaria-Santiago, Amadeo Arguelles, Magdalena Saldana-Perez, Jocelyn Richardson, Marcella Bona

AI 总结：该研究针对RAG的证据利用测量与上下文预算分配瓶颈，提出因果留一法探针与闭环次模调度器，实现组合召回率提升16.7-20.5个百分点，确立顺序反馈驱动编排为生成式搜索的范式。

链接：https://arxiv.org/abs/2608.23252

机构：National Engineering Research Center for Software Engineering, Peking University（北京大学国家软件工程研究中心）； Peking University（北京大学）； Tencent（腾讯）

作者：Peiyang Liu, Xi Wang, Di Liang, Wei Ye

AI 总结：针对现有外科AI忽略团队互动建模的问题，提出tempo-relational框架结合Time-Expanded图，通过反事实程序生成可解释建议，在模拟手术实验中提升了多目标预测性能。

链接：https://arxiv.org/abs/2608.23254

机构：University of Trento（特伦托大学）； UiT the Arctic University of Norway（挪威北极大学）

作者：Vincenzo Marco De Luca, Antonio Longa, Giovanna Varni, Andrea Passerini

AI 总结：该研究提出混合Transformer框架，结合多源时间编码器与动态条件融合解码器，在MIMIC-III数据库上实现无创连续血压预测，性能优于基线方法，符合AAMI和BHS A级阈值。

链接：https://arxiv.org/abs/2608.23276

机构：College of Computer Science, Beijing University of Technology（北京工业大学计算机学院）

作者：Yuexin Ma, Jingqi Hou, Yuxuan Kang, Zhaoying Liu

AI 总结：针对通用算法无法适配计数数据分布特性的问题，提出基于泊松等统计模型的3CPO聚类算法，可识别相关列子集提升可解释性，经多领域实验验证其聚类质量与鲁棒性。

链接：https://arxiv.org/abs/2608.23287

机构：Aalto University（阿尔托大学）； University of Helsinki（赫尔辛基大学）

作者：Collin Leiber, Kai Puolamäki, Heikki Mannila

AI 总结：该研究针对非洲高分辨率贫困数据不足的问题，提出不确定性感知的EO-ML贫困制图方法，生成统计可靠的预测区间，还开发了风险可控的援助分配程序，为政策制定提供可靠补充。

链接：https://arxiv.org/abs/2608.23322

机构：Chalmers University of Technology（查尔姆斯理工大学）； University of Gothenburg（哥德堡大学）； Linköping University（林雪平大学）； Karlstad University（卡尔斯塔德大学）； Massachusetts Institute of Technology（麻省理工学院）

作者：Markus B. Pettersson, James Bailie, Mohammad Kakooei, Eagon Meng, Adel Daoud

AI 总结：本文针对隐私增强实例编码的可逆性，提出考虑编码器频谱结构的新理论界，其更紧致且适用于确定性编码器，可扩展至其他范数度量，经多类实验验证优于现有界。

链接：https://arxiv.org/abs/2608.23382

机构：The Pennsylvania State University（宾夕法尼亚州立大学）

作者：Seokjin Hwang, Yuting (Ray)Li, Kiwan Maeng

AI 总结：该研究提出了公理交易者框架，将系统化交易建模为潜在状态驱动的时不变机制，确定了构建量化投资系统所需的五个关键常数，为该系统的架构提供了近乎确定的标准形式。

链接：https://arxiv.org/abs/2608.23416

作者：Jiayu Li

AI 总结：本文提出无训练外推框架ChebBooster，基于切比雪夫多项式理论解耦为离线预计算与轻量在线阶段，在三种DiT模型上实现最高3.68倍延迟加速，优于现有无训练基线。

链接：https://arxiv.org/abs/2608.23429

机构：College of Electronics and Information Engineering, Shenzhen University（深圳大学电子与信息工程学院）； School of Artificial Intelligence, Shenzhen University（深圳大学人工智能学院）

作者：Chengjie Lu, Tianchi Deng, Zhengqi He, Chengwen Luo, Xueliang Li

AI 总结：该研究评估贪心K中心在多种度量空间中用于主动学习选择的性能，发现将未标注实例映射到带熵加权的预测概率空间时效果最优。

链接：https://arxiv.org/abs/2608.23461

机构：University of California, Berkeley（加州大学伯克利分校）

作者：Siddharth Chilamkur, Dorit S. Hochbaum