Drift Contract：结合谱更新实现深度鲁棒的局部学习算法
局部学习通过各层独立的辅助损失避免了全局反向传播，具备天然的结构并行优势，但长期受困于深度增加时的精度衰减和超参数脆弱性。该研究首次将类似 Muon 的谱更新几何（包含动量正交化与谱步长缩放）应用于逐层局部更新。在 CIFAR-10 MLP 基准测试中，该方法在深度 12 至 48 层、宽度 128 至 2048 的范围内，仅需单一固定步长即可取得最优表现，显著提升了局部学习对网络深度的鲁棒性。
DC AI 热点
全部 AI 动态
9月24日2026-09-24
Signal2Symbol：用于可解释生理时间序列异常检测的神经符号时序推理框架
针对心电图（ECG）和脑电图（EEG）等生理时间序列在复杂时序结构与高决策透明度方面的需求，现有深度学习模型往往缺乏对异常原因及局部关联的可解释性。研究团队提出了一种名为 Signal2Symbol 的神经符号框架，旨在实现可解释的生物信号异常检测。该方法首先将连续的生理信号转化为符号序列，进而结合神经符号时序推理，提升检测决策的透明度与规律挖掘能力。
HARN：面向事件驱动多时间周期预测的分层联想共振网络
针对金融时间序列在多时间尺度预测中重复计算不变表征的难题，研究者提出了分层联想共振网络（HARN）。该模型专为事件驱动的多周期预测设计，通过在不同时间层级维护持久化表征，仅在对应周期的K线数据生成完成后才触发该层更新，显著避免了冗余计算。其架构综合了因果多尺度时间编码、门控联想记忆、跨层级共振以及分层证据聚合机制。
终端基准测试何以称“难”？研究揭示智能体评测中的“真假难度”陷阱
前沿基准测试常依赖当前模型无法解决的任务来衡量能力边界，但零通过率并不等同于真正的高难度。该研究指出，任务全军覆没既可能源于真实的能力代差，也可能是由缺失上下文、损坏的参考方案或基础设施故障等“虚假难度”导致。基于 Terminal-Bench 3 等基准中涵盖 639 个任务、2.8 万余次试验及逾 10 万美元开销的真实记录，研究系统性厘清了导致智能体任务失败的本质原因，为评测体系纠偏。
LWCal：面向含噪标签表格分类器的损失加权概率校准方法
传统事后概率校准通常假设留出校准集标签完全干净，但在实际落地场景中，校准标签往往受弱标注或启发式规则影响而存在噪声。为此，研究人员针对表格分类器提出了一种名为LWCal的轻量级事后校准器。该方法仅依赖CPU运行，通过降低与基模型留出预测概率相矛盾的含噪样本权重，在无需干净验证集的情况下完成有效校准，解决了弱监督或含噪环境下的概率可靠性问题。
面向术中急性肾损伤早期预测的防数据泄露多模态评估框架 SynerT
针对非心脏大手术后急性肾损伤（AKI）早期术中风险分层难的问题，该研究提出了一种名为 SynerT 的纯波形混合时序骨干网络。该网络结合了因果扩张时间卷积网络（TCN）与分层扩张循环层，用于编码术中早期生理轨迹。此外，研究还构建了引入结构化临床背景的多模态扩展变体 SynerT-MM 进行晚期融合预测。注：输入素材尾部存在文本截断，后续具体融合细节未完全展示。
COPE：利用用户嵌入与自评估实现大模型在稀疏反馈下的持续个性化
针对大语言模型在标准对齐后输出同质化、难以满足用户多样化偏好的问题，研究人员提出了COPE优化框架。现有免训练方法多依赖提示工程占用上下文窗口，而微调方法在训练后保持静态，无法持续适应动态偏好。COPE结合个性化用户嵌入与模型自评估机制，有效克服了现实场景中用户反馈稀疏的挑战，实现了大语言模型在部署后的持续个性化优化。
QUARTET：基于四分支交叉注意力与随机游走轨迹增强的关系图Transformer
关系深度学习将多表数据库建模为异构时序图，目前图Transformer在该领域表现优异。然而，现有代表性模型RelGT存在局部采样子图连接松散阻碍消息传递，以及全局注意力依赖单一特征记忆而忽视宏观动态的两大缺陷。为此，研究人员提出了QUARTET架构，通过结合四分支交叉注意力与随机游走轨迹，增强全自注意力机制在关系图表征中的表达能力。
阅读原文 ↗推荐理由：针对关系深度学习中图Transformer的采样稀疏与全局动态建模不足问题，提出了新颖的增强架构。# 图Transformer# 关系深度学习# 图神经网络# 注意力机制# RelBench
智能体训练中工具缓存可能导致策略更新方向反转
工具结果缓存可显著减少智能体训练中的重复计算，但也会耦合采样过程的随机性。最新研究通过一个双动作模型证明，即使独立执行与共享执行在边际上保持了每个采样的条件奖励分布一致，在组内共享随机结果仍可能导致预期的组归一化策略更新方向发生反转。理论推导表明，共享缓存更新反映的是胜负概率之差而非期望奖励之差，揭示了智能体强化学习中常见工程优化策略潜在的算法偏差风险。
PR-Smoother: Simulator-Preserving Non-Gaussian Smoothing for Data Assimilation
arXiv:2609.26890v1 Announce Type: new Abstract: Many physical data assimilation (DA) workflows require smoothing methods that represent non-Gaussian posteriors over physical state variables, scale to high-dimensional simulators, train from observation windows alone, and remain compatible with calibration of the prescribed simulator. We introduce PR-Smoother, a simulator-preserving amortized smoother designed for this prescribed-simulator DA regime. Its key design principle is to keep the prescribed simulator explicit in both the evidence lower bound and the variational family: rather than learning replacement dynamics or a learned trajectory
CORE-STACK+: Meta-Learning for Deep Stacked Generalization
arXiv:2609.26905v1 Announce Type: new Abstract: Stacking heterogeneous vision backbones (CNNs, ViTs, and hybrids) is the de facto recipe for accuracy, calibration, and robustness, yet two coupled pathologies limit its returns. Prediction-space multicollinearity ill-conditions the meta-learner's Gram matrix, inflating weight variance and producing brittle solutions on a thin manifold. Calibration collapse compounds constituent miscalibration through naive linear stacking, so adding more models can hurt expected calibration error (ECE). Existing remedies, ridge regularization, greedy selection, model soups, and SWAG address at most one of these
On Preference Coverage Collapse from Hindsight Relabeling in Multi-Objective Reinforcement Learning
arXiv:2609.26918v1 Announce Type: new Abstract: Hindsight relabeling which retroactively replacing a transition's goal with the outcome the agent actually achieved is an effective tool for improving sample-efficiency in Reinforcement Learning (RL). A natural extension to preference-conditioned multi-objective RL (MORL) relabels transitions with the preference direction the agent achieved rather than the one asked for. We show that this extension is frequently harmful: across four preference-conditioned off-policy algorithms spanning two critic backbones and two preference-sampling schemes on the continuous-control MO-Gymnasium suite, it degra
When Post-Processing Fairness Constraints Help and When They Harm: Evidence from Eight Cross-Domain Evaluations
arXiv:2609.26955v1 Announce Type: new Abstract: Fairness audits in production ML typically occur once, at deployment, on a single domain. Both fail in practice: fairness can shift after retraining or a changing user base, and interventions validated on one dataset are rarely tested across the heterogeneous domains an organization deploys. We present FAPE (Fairness Auditing for Production Environments), a four-stage framework evaluating a single post-processing intervention, Fairlearn's ThresholdOptimizer, across eight domain evaluations: criminal justice, income prediction, legal admissions, credit lending, agricultural lending, a multi-domai
Transfer Learning with Conformalized Quantile Regression for Solar PV Forecasting Under Load-Shedding-Driven Data Scarcity
arXiv:2609.26959v1 Announce Type: new Abstract: Solar photovoltaic (PV) forecasting in regions affected by load shedding is challenging because reliable historical observations are scarce. This study proposes a transfer learning framework combined with Conformalized Quantile Regression (CQR) to improve PV power forecasting and provide reliable uncertainty estimates under severe data scarcity. A source-domain PV dataset from Alice Springs, Australia, is used to pretrain a temporal forecasting model, which is then adapted to simulated Bangladesh PV data representing different levels of historical availability. Experimental results show that tra
CRISP: Scalable Importance-Stratified Coresets for Imbalanced Tabular Learning
arXiv:2609.26962v1 Announce Type: new Abstract: Large imbalanced tabular datasets make repeated gradient-boosted tree training expensive. Existing coreset methods often lose accuracy when most majority examples are removed. We present CRISP (Coreset Reduction via Importance-Stratified Pruning), a linear-time method that allocates a negative-class budget across quantile strata of a proxy-model score. Sample weights account for unequal inclusion probabilities. At 95% negative-class reduction on a production fraud dataset, CRISP trains on approximately 1.70M of 25M rows and retains 99.7% of full-data Average Precision. This is a 93.2% reduction
TinyUDE: Solver-Free Universal Differential Equations on Microcontrollers via Lie-Taylor Jet Matching
arXiv:2609.26972v1 Announce Type: new Abstract: Training Universal Differential Equations (UDEs) traditionally relies on backpropagating through numerical ODE solvers, creating memory footprints far exceeding the capabilities of edge microcontrollers. We present Lie-Taylor jet matching, a solver-free training framework that fits a hybrid vector field directly to the first and second time-derivatives of observed system states. These derivatives, the truncated Lie-Taylor jet, are estimated online via Savitzky-Golay filtering, yielding fully analytic gradients without automatic differentiation software. We evaluate whether eliminating the solver
Resource-Efficient Distributed Recursive Gaussian Processes
arXiv:2609.26979v1 Announce Type: new Abstract: Gaussian processes (GPs) provide a flexible framework for learning unknown functions from noisy measurements while quantifying predictive uncertainty, making them well suited for estimation in multi-agent systems. However, when measurements are collected by multiple agents, maintaining a unified GP model without centralized processing requires efficient distributed algorithms that can operate using local measurements and communication with neighboring agents. In this work, we develop two distributed recursive GP (RGP) algorithms for multi-output GP regression: ADMM-RGP and PDMM-RGP. We analyze t
GeoRVQ: Decoder-aware geometry for residual-token prediction in physiological signals
arXiv:2609.27018v1 Announce Type: new Abstract: Residual vector quantization (RVQ) turns physiological waveforms into compact token sequences, but conventional masked modeling treats every incorrect token as equally costly. We propose GeoRVQ, a coarse-to-fine masked token model whose objective reflects the local response of a frozen waveform decoder. Decoder-induced costs define geometry-aware soft targets and expected distortion, while quantizer-causal prediction follows residual dependencies from coarse to fine levels. In a descriptive aggregate over MIMIC-IV Waveform, VitalDB, and CODE-15\%, GeoRVQ increases exact token accuracy from $.133
WTF?! Simulation-Free Reinforcement Learning with Wasserstein-Tilted Flow Maps
arXiv:2609.27033v1 Announce Type: new Abstract: Reward fine-tuning aims to update a pre-trained flow-based generative model to improve the downstream reward of its generated samples. Existing methods typically formulate this problem as sampling from a reward-tilted distribution, the solution to a KL-regularized reward-maximization problem. Here, we introduce an optimal transport regularizer built directly from the pre-trained drift. Unlike KL reward tilting, the resulting objective transports individual samples toward higher reward rather than reweighting the base distribution. We show that the resulting problem is equivalent to a determinist
An open benchmark for machine learning-based polymer property prediction
arXiv:2609.27036v1 Announce Type: new Abstract: Polymer property prediction lacks open, standardized benchmarks that enable rigorous comparison of machine-learning methods, with existing resources covering only a narrow fraction of polymer architectures, such as homopolymers. We introduce Polymer Benchmark 2026 (PolyBench26), an open dataset comprising nearly 250,000 polymer-property datapoints across eight physical properties, including data from experimental measurements, density functional theory, and molecular dynamics. The benchmark supports four evaluation tasks across homopolymers and alternating, random, and block copolymers: in-distr
ChipMEM: Verification-Grounded Memory for EDA Agents
arXiv:2609.27067v1 Announce Type: new Abstract: Large language model (LLM)-based agents use Electronic Design Automation (EDA) tools to generate and revise register-transfer-level (RTL) designs under synthesis and verification feedback. Recent methods learn from this feedback by distilling reusable skills from execution traces or by training on rewards derived from EDA-tools. Both methods are typically evaluated on the tasks that produced the experience. Repeated access to benchmark feedback on the same task can reward task-specific revision rather than creating reusable knowledge that transfers. We introduce ChipMEM, a verification-grounded
Local Evidence and Geometric Readout Repair in Trained GNNs
arXiv:2609.27092v1 Announce Type: new Abstract: Many node-classification GNNs apply a linear classifier to a nonnegative mixture of local messages. An error can reflect either poor mixture weights or a reachable logit set poorly positioned for the classifier. We separate these causes with an exact-mass linear program and two learned post-hoc repairs. Every reweighted prediction has an equivalent centered logit translation, but only translations in a message-induced displacement set are realizable by reweighting. Across eight datasets, eight GNN backbones, and ten splits, mean accuracy rises from 62.6% for the frozen models to 63.8% with rewei
Learning Risk Scores Robust to Unobserved Confounders
arXiv:2609.27144v1 Announce Type: new Abstract: We consider the problem of learning risk scores to prioritize individuals for scarce resources or interventions, from historical observational data affected by unobserved confounding. Decisions about who receives scarce resources are often guided by risk scores based on recorded characteristics, such as responses to a survey. These risk scores are increasingly being learned directly from observational data: historical records of individuals' characteristics, allocation decisions, and outcomes. Standard methods such as inverse propensity weighting (IPW), which corrects for the bias introduced by
The Linear Representation Hypothesis Needs a Group Action
arXiv:2609.27158v1 Announce Type: new Abstract: To make claims about representations that generalize beyond a particular trained model, we need to specify when two representations should count as equivalent. The Linear Representation Hypothesis is often discussed without making this equivalence explicit. Different notions of equivalence preserve different structures, so metrics, probes, and interventions that appear to study the same representation may in fact correspond to different hypotheses. We therefore argue that the Linear Representation Hypothesis is not one hypothesis but a family of claims distinguished by representation equivalence
Scaling of Capability and Efficiency at Inference Time in Large Reasoning Models
arXiv:2609.27166v1 Announce Type: new Abstract: Capability and efficiency are two key dimensions of reasoning in large language models (LLMs). Capability refers to the ability to solve a given problem correctly, whereas efficiency refers to the ability to do so with limited resources. When LLMs use Chain-of-Thought (CoT) reasoning to solve problems of controlled hardness, both the number of problems solved correctly and the number of tokens required to reach a correct answer depend on problem hardness and model size. However, how these factors jointly shape capability and efficiency remains poorly understood. Here, we use hierarchical Bayesia
Data-driven discrete-time deep recurrent neural network-based modeling for dissipative systems
arXiv:2609.27186v1 Announce Type: new Abstract: Physical AI has gained increasing attention for its role in developing AI systems that better understand, predict, and control real-world dynamics. Achieving this requires AI models that not only achieve high prediction accuracy but also preserve fundamental physical properties of dynamical systems. In this paper, we propose a deep discrete-time dissipative recurrent neural network (DissipNet) that explicitly enforces dissipativity, a key property related to stability and energy dissipation, through structural weight constraints and a dedicated training algorithm. By construction, the proposed n
ZO-COSMO: Index-Free One-Hop Mixing for Decentralized Zeroth-Order Optimization
arXiv:2609.27199v1 Announce Type: new Abstract: Sparse communication in decentralized zeroth-order learning requires compatible peer-state coordinates. We characterize this one-hop condition and develop \textsf{ZO-COSMO}, coupling two-query estimation with average-preserving masked consensus using $q$ values per active link. Global supports serve all-neighbor mixing; matching updates require agreement only within each pair. We derive a sharp contraction-per-scalar bound within the matching class and convergence guarantees for the core and sparse-momentum updates. At fixed matching, exact moment identities characterize how shared directions pr
A Systematic Benchmark of Explainable Methods for Temporal Attribution in Sequential Recommendation Systems
arXiv:2609.27201v1 Announce Type: new Abstract: Sequential RecSys are central to modern personalization, exploiting user's historical interaction sequences to drive next-step decisions. Deep learning models, particularly CNN and Transformer-based architectures, have proven highly effective at capturing temporal dependencies in these histories. For transparency and trust, understanding which past interactions drive a given recommendation is increasingly important --- both for developers auditing model behavior and for users seeking a rationale. However, the non-linearities that give these models their predictive power also render them black bo
Scalable Subgraph Sampling via Resistance Curvature
arXiv:2609.27209v1 Announce Type: new Abstract: Subgraph sampling reduces the training cost of large-scale graph neural networks, but sampling criteria may overlook the geometric roles of edges. We propose a resistance-curvature-guided sampling framework built on ERC-LG, a curvature approximation method for large-scale graphs. ERC-LG combines Johnson-Lindenstrauss projections with regularized multi-GPU batched conjugate gradient solvers, avoiding explicit Laplacian pseudoinverse computation and full embedding storage. The resulting curvature informs node- and edge-sampling probabilities for constructing GNN training subgraphs. Experiments sho
Tail-Aware Geometry Learning for Conformal Ellipsoids
arXiv:2609.27221v1 Announce Type: new Abstract: This paper studies multivariate conformal prediction (CP), a distribution-free uncertainty quantification framework with finite-sample coverage guarantees. The efficiency of multivariate prediction sets hinges critically on the residual geometry encoded by the nonconformity score, while existing minimum-volume methods rely on quantile thresholds that ignore tail residual severity and implicitly bind geometry learning to coverage level. We propose a tail-aware geometry learning framework for conformal ellipsoids that decouples tail sensitivity in geometry learning from the final coverage guarante
1 / 2下一页 →