跳到正文
研踪PaperTrail

YOUR RESEARCH, IN SIGHT

让下一篇重要论文，不再错过。

覆盖计算机全学科，以 AI 为快捷入口。关注方向、作者和关键词，把发现变成持续的研究积累。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+2新增论文
+2其中 AI 相关
+0已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-12 23:49 · 全站最近入库：2026-11-12 06:37各信源进度 →

统计口径与数据新鲜度

统计 2026-11-12 00:37 — 2026-11-13 00:37（北京时间） · 点击数字查看对应简报；AI 相关包含在新增论文内。

最近完成检查：2026-11-12 23:49最近发现新增版本：2026-11-12 06:37库内论文最新更新时间：2026-11-11 08:00

每分钟刷新已保存的数据。空检查不代表采集异常；新收录可能包含历史论文。AI 导读按新增版本计数。

21,675 已收录论文18,323 AI 相关6,056 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

时间范围跟随排序：版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常，查看状态 →

找到 4,903 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-12 23:44（北京时间）
版本更新 · v4arXiv:2609.33239

Schrödinger--Föllmer Actor--Critic: Diffusion Policy Improvement with Finite-Sample Analysis

Yuling Jiao, Lican Kang, Jerry Zhijian Yang, Jincheng Ying

原文摘要

Diffusion policies represent multimodal action distributions, but a $Q$-weighted update does not specify how to sample from the resulting target distribution. We propose Schrödinger--Föllmer Actor--Critic (SFAC), an offline-to-online reinforcement learning (RL) method for Kullback--Leibler (KL)-regularized policy improvement through conditional diffusion. A minimax Bellman critic estimates the action-value function, which defines an exponentially tilted target policy. A Doob $h$-transform expresses this update as a correction to the reference diffusion drift. The target drift is a weighted conditional mean of Brownian-bridge labels. We fit it by weighted regression using endpoints from the current policy and pointwise action-value evaluations. In the small-update regime, the KL-regularized update follows the natural policy-gradient direction, and the Doob correction represents the same local change in the space of diffusion drifts. Under suitable conditions, we derive finite-sample bounds that separate the errors from critic estimation, weighted neural drift regression, and diffusion discretization through a reverse-KL actor bound. The average-performance bound depends explicitly on the comparison coefficients over the policy iterations. Across six D4RL offline-to-online tasks, the final scores of the $Q$-guided bridge-regression implementation exceed its offline initialization on every task and Cal-QL on five tasks. We also compare with offline RL and online-only baselines.

stat.ML · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.12998

Linear Probes to Bayesian Manifolds: Geometry of Arithmetic in Large Language Models

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

原文摘要

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

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.28182

Finite-Sample Probabilistic Safety Certification for AI-Based Grid-Edge Coordination

Yihong Zhou, Hanbin Yang, Thomas Morstyn

原文摘要

Coordinating a large population of flexible grid-edge devices can alleviate the need for time-consuming and capital-intensive network upgrades. Artificial intelligence (AI)-based control methods such as multi-agent reinforcement learning or imitation learning are promising to this coordination due to their real-time decision scalability. However, system operators still need an independent and rigorous way to decide whether a given AI system is safe enough for deployment. This paper develops a \emph{finite-sample} probabilistic safety certification framework for AI-based grid-edge coordination. The central idea is to reduce the complete input--AI--decision--evaluation workflow to a binary unsafe outcome under an operator-defined safety specification, and then use exact binomial inference to certify the corresponding unsafe operation probability. Given a set of held-out calibration scenarios, the framework returns the tightest one-sided upper certificate and an accept/reject deployment criterion that controls the probability of false safety certification. Because the certification is for the calibration distribution that may \emph{deviate} from the future operation, we further combine the nominal certificate with physically interpretable sample-space adversarial attacks, a concept widely used in AI models to investigate their fragility. Case studies on grid-edge flexibility coordination with 1{,}000-agent AI models (independent parameters) verify the safety guarantee and the value of integrating adversarial attacks into a rolling-window training-certification-deployment flow.

cs.AI · cs.LG · eess.SY

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11398

Decomposing Earth Embeddings with Sparse Autoencoders

Vitus Benson, Fanny Yang, Markus Reichstein

原文摘要

Earth observation foundation models produce global metre-resolution embeddings that underpin work across agriculture, ecology, hydrology, and beyond. Yet these embeddings are opaque: each dimension entangles unrelated physical concepts, which makes them difficult to interpret and their robustness hard to assess. Here, we propose sparse autoencoders (SAEs) as a principled way to decompose Earth embeddings into sparse, approximately monosemantic features over an overcomplete dictionary. Because LLM judges of interpretations do not transfer to satellite pixels, and spatial autocorrelation matters, we introduce a tailored evaluation protocol. It scores any SAE on reconstruction at matched sparsity, alignment with reference concepts, and spatial scale, each against baselines that include an SAE whose decoder is a frozen random dictionary. On this protocol, SAEs reconstruct AlphaEarth and Tessera embeddings better than every baseline at matched sparsity. Their features align with far fewer reference concepts than raw embedding dimensions (on AlphaEarth, 88% align with at most one of 20 concept groups, most with none, against a median of 12 per raw dimension). They also carry more of their variance at fine spatial scales (23 to 32% within a 640 m patch for BatchTopK and Matryoshka, against 9%). An SAE with a frozen random decoder shares both properties, which therefore follow from sparsity, whereas only a learned dictionary reconstructs the embedding accurately: on AlphaEarth, trained SAEs leave eight times less variance unexplained. We test the sparse code on two applications, discovering land-cover sub-concepts and detecting out-of-distribution pixels by area-of-applicability estimation. Both indicate how sparse features could be used, but neither is yet reliable enough for practical use. Together, these results are a step toward Earth observation models whose features can be inspected.

cs.LG · cs.CV · physics.geo-ph

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2609.24504

On Emergent Capabilities and Model Merging

Luca Zhou, Emanuele Rodolà

原文摘要

Model merging combines separately fine-tuned checkpoints by arithmetic on their weights, and is now a routine way to assemble capabilities cheaply. We ask what it does to emergent capabilities: behaviors that arise from fine-tuning but were never an explicit training target. Studying two independent testbeds (activation oracles and emergent-misaligned models) across three model families, we find that the answer is threefold. First, merging preserves an emergent capability that both parents carry: merging two misaligned checkpoints retains most of their broad misalignment across the whole mixing range. Second, merging cannot create an emergent capability that is superadditive in its parents: no weighted merge of two single-task oracles reaches the jointly-trained oracle's auditing ability. Third, when only one parent carries the capability, merging dilutes it faster than the trained capability that accompanies it: the gap is significant in most settings.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11387

Orthogonal Updates for the Win: Towards Accelerated Adaptive Minimax Optimization

Zhiwei Zhai, Xinyu Wang, Wenjing Yan, Lei Ding, Ying-Jun Angela Zhang

原文摘要

Single-loop minimax methods are appealing for modern machine learning, but standard stochastic gradient updates can be unstable and oscillatory under tightly coupled primal-dual dynamics, making performance notoriously sensitive to stepsizes. To fundamentally improve stability, we develop an orthogonal update framework tailored to minimax optimization, which applies updates along orthogonalized directions to reduce update anisotropy. However, orthogonalization reshapes the optimization geometry and makes coordinating primal-dual stepsizes even more delicate. To overcome this challenge, we develop AdaSGDA, which equips orthogonal updates with an adaptive stepsize rule driven by accumulated gradient norms. This mechanism automatically balances coupled primal-dual progress under orthogonalization, yielding an effectively parameter-free method for nonconvex-strongly-concave minimax optimization. We prove that AdaSGDA avoids problem-dependent tuning while attaining the state-of-the-art $\mathcal{O}\left(T^{-1/4}\right)$ convergence rate. We further propose AdaSGDA-VR, which incorporates a carefully designed variance reduction scheme and achieves the faster $\mathcal{O}\left(T^{-1/3}\right)$ rate. Extensive experiments demonstrate that our methods deliver competitive performance compared to existing minimax optimization methods.

cs.LG · math.OC · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11386

What Travels Across Languages? Reading, Transfer and Erasure of Cross-Lingual Concept Subspaces in Bangla Idioms

Syed Rifat Raiyan, Yousuf Ibne Kamal Niloy

原文摘要

Cross-lingual concept analysis requires distinguishing the information that a subspace retains from the classifiers and erasers that transfer through it. We study these operations in six frozen encoders using 10,359 Bangla (Bengali) idiom senses, eight input conditions, and figurative-class and sentiment labels. Class-mean directions provide the most consistent evidence of shared coarse meaning: in post hoc analyses, they outperform random-subspace, permuted-label and covariance-matched nulls in all 12 model-label tests under balanced accuracy. A prospective internal re-split supports idiom-alone sharing in 10 of 12 tests, spanning five encoders. The original confirmatory logistic-regression family rejects both planned nulls in 4 of 12 tests, where English directions gain only a few accuracy points above the majority-class floor (3.0-7.3). Classifier transfer depends on the input map and regularization: English-trained class-mean readouts beat both zero-shot nulls in 8 of 12 discovery tests, and stronger source-selected penalties improve transfer through logistic and LEACE directions. These readouts use Bangla-validation-selected layers and unlabeled target preprocessing. Input controls show greater coarse-class sharing for explicit glosses than for idiom counterparts. Reading, however, does not imply effective erasure: English-fitted LEACE removes less than one percentage point of Bangla figurative-class accuracy, compared with 4-11 points for Bangla-fitted LEACE; INLP shows a similar asymmetry. Pattern-filter analysis, covariance-weighted input maps and a binary Gaussian overlap bound explain how these operations can diverge. The analogous multiclass ordering is an empirical diagnostic, observed in 11 of 12 cells. The results support an estimator-sensitive account of cross-lingual concept sharing and motivate evaluating reading, classifier transfer, erasure and overlap separately.

cs.CL · cs.LG · stat.ML

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.00920

Non-intrusive appliance state detection in unseen homes using classical machine learning

Zablon Ombiro

原文摘要

Non-intrusive load monitoring (NILM) infers, from the aggregate signal of a single smart meter, when each appliance in a home is ON or OFF, offering significant potential to enhance energy efficiency. Classical machine-learning detectors report F1 scores (the harmonic mean of precision and recall) from 0.53 to 0.98, but the causes of this spread are unseparated, so it is unclear which design choices matter. Real-world deployments, however, require detectors that work in houses never seen during training, whereas most studies test on houses used for training. To address this challenge, we propose a factorial analysis of variance design that separates the effects of classifier, feature set and class-imbalance strategy across five appliances (270 configurations), and a leave-house-out protocol with house-grouped model selection. Unlike studies that vary one factor at a time or pool houses, the approach attributes performance to each factor and measures it on unseen homes. Extensive experiments on REFIT, a public dataset of 20 UK households, show that the appliance explains 38 percent of F1 variance, the feature set 13 percent, the classifier 8 percent and the imbalance strategy 2 percent. Pooled testing overstates support-weighted F1 (0.548 versus 0.447 in unseen houses), class weighting gains grow from 32 percent to 69 percent, and feature engineering gives the same 52 percent gain in both protocols. House-grouped selection raises F1 for three of five appliances, by up to 0.110, and a convolutional neural network scores 0.450 (0.496 calibrated) against 0.447 for classical models. The findings guide appliance-level monitoring in smart-metering and energy-management systems.

cs.LG · eess.SP

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11382

The comparison of density estimation for transfer learning in Banach, Hilbert, Euclidean spaces

Maryam Alsadat Kazemian, Azam Bastanfard

原文摘要

Domain adaptation or DA methods are indispensable when target labels or computational resources are scarce, yet the domain kernel density and subspace alignment approaches completely assume a Hilbertian geometry. They measure distributional discrepancy with symmetric L2 divergences over Euclidean feature spaces or reproducing kernel Hilbert spaces (RKHS). This preprint summarizes the manuscript Non-Hilbertian Domain Adaptation. Total Variation Alignment via L1 Density Regularization on Feature Manifolds, which proposes Banach Manifold Transfer Learning or BMTL. BMTL makes two coupled design decisions. The first one is that features are embedded on a geodesic manifold rather than treated as points of a flat Euclidean space, and the second one is that the density matching regularizer is formulated in an L1 Banach norm, whose minimization is equivalent, by Scheffe's theorem, to contracting the Total Variation or TV distance between domain densities, a bounded metric that weights all regions of the sample space uniformly. The L2 formulation, by contrast, squares deviations and therefore amplifies errors in the tails, over emphasizing density outliers that do not reflect meaningful distributional shift. Experiments on MNIST-USPS, COIL20, and Office+Caltech10 show consistent gains; the L1 variant beats an otherwise identical L2 ablation on every task, for example 92.44% vs. 89.80% on USPS->MNIST and surpasses the strongest baseline DMDA by +3.68% on average over twelve Office+Caltech transfers. This document situates the contribution with an explicit comparison of Euclidean, Hilbert, and Banach geometries.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11381

Does Predictive Performance Translate to Causal Performance? Evaluating tabular foundation models as nuisance learners for causal effect estimation

Andrew R Weckstein, Issa J Dahabreh, Marinka Zitnik, Sonia Hernández-Díaz, Sebastian Schneeweiss

原文摘要

Tabular foundation models based on prior-data fitted networks (PFNs) achieve strong predictive performance, but their value for causal inference is unclear. We evaluated whether PFNs, which predict through in-context learning, could match or outperform conventional nuisance learners in supporting accurate causal estimation and reliable uncertainty quantification (inference). Learners modeled propensity scores and outcome regressions within doubly robust estimators of average causal effects, with and without cross-fitting, across synthetic and semi-synthetic settings with varying confounding complexity and effect heterogeneity. Under simple parametric data-generating processes, the PFNs were competitive with correctly specified parametric regressions and ensembles containing them. In more complex settings, they outperformed all conventional learners evaluated, including richly specified Super Learner ensembles, for both accuracy and uncertainty quantification. As sample size increased, only the two PFN models (TabPFN and Mitra) consistently approached nominal confidence interval coverage, whereas under-coverage persisted for conventional learners. In some settings, PFNs achieved near-nominal coverage and well-calibrated standard errors even without cross-fitting. At the smallest sample size (n=200) and in the higher-dimensional semi-synthetic settings, PFNs and other flexible learners exhibited poor coverage without cross-fitting. For PFNs, this appeared to be driven by propensity score overfitting consistent with in-sample memorization of treatment labels. These findings highlight the promise of prior-data fitted networks as in-context nuisance learners for causal estimation, while underscoring the need for further evaluation in challenging settings.

stat.AP · cs.AI · cs.LG · stat.ME · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11379

Self-Supervised Collapse Prevention with Score-Based Entropy Maximization

Adarsh Kappiyath, Silpa Vadakkeeveetil Sreelatha, Xiatian Zhu, Yann LeCun, Lu Yin, Amir Bar

原文摘要

Diffusion models have demonstrated the effectiveness of score-based objectives for generative modeling, yet their potential for self-supervised representation learning remains underexplored. We introduce Entropy-Based Regularization (EBReg), a framework that uses latent score estimation to regularize the latent distribution during self-supervised training. EBReg jointly learns an encoder and a latent score model, and utilizes the learned score to provide a diversity-preserving signal that prevents representational collapse. Motivated by a mutual-information lower bound, we formulate the objective as view alignment with smoothed marginal entropy maximization and optimize the marginal entropy without explicit density evaluation. The same score model can additionally be used at test time to refine corrupted representations toward the training latent distribution. Empirically, EBReg learns competitive representations on ImageNet-1K, achieving 75.57% linear-probe accuracy with ViT-L. Across diverse architectures, score-guided test-time adaptation further improves robustness on ImageNet-C, yielding an average relative accuracy gain of approximately 5%.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11372

LLM-based Multi-Criteria Deep Reinforcement Learning for Dynamic Efficient 5G UPF Selection

Abdelrahman Soliman, Ahmed Refaey Hussein, Lei Lei

原文摘要

User plane function (UPF) selection in 5G networks is an active area of research, with various limited approaches proposed to enhance network efficiency and manage resources effectively. In this paper, we address this gap in the related literature by proposing a Deep Reinforcement Learning (DRL) UPF selection method that utilizes a dynamic, multi-objective reward function and validating it on a real-world 5G emulation platform. The proposed method is based on the large language models' (LLMs) translation of network operator intents and the network data analytics function (NWDAF) abnormality alerts. The experimental results validated in the real-world experimental setup demonstrate that our dynamic DRL solution achieves an effective balance between latency, power consumption, and traffic distribution, outperforming the heuristic baselines and fixed DRL by 16.89%. These findings highlight the real-world applicability of our approach for dynamic, intent-driven UPF selection.

cs.LG · cs.AI · cs.CL · cs.NI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11371

D3Tree: Learning to Prune Block-Diffusion Draft Trees for Efficient Verification

Hong Nie, Zhangchi Gao, Jun Yu, Bo Wang, Jianqing Sun, Shengping Liu

原文摘要

Block-diffusion drafters propose a whole block of tokens in one forward pass, and draft trees built from their position-wise marginals increase the expected accepted length of speculative decoding. Every tree node occupies a query row in target verification, so under concurrent serving low-yield branches compete with other requests for verification capacity. Existing tree methods select verified nodes by draft probability under a fixed or cost-selected per-request budget, and thus always retain a prefix of the drafter's ranking. We present D3Tree, which learns which nodes of a block-diffusion draft tree are worth verifying. A tree-size head with 2H+2 parameters predicts each candidate's conditional survival from the draft state, the candidate embedding, and its draft log-probability. Thresholding the product of these scores along parent chains yields an ancestor-closed subtree whose size adapts to each request and decoding round. The head is trained on a frozen drafter with conditional labels on the full candidate tree; an optional joint mode also steers the drafter toward the retained reference path. At serving time, D3Tree verifies whole trees under light load and, under heavy load, allocates one batch-wide node budget by predicted survival, sized to the target's matrix-multiplication tiles. Retained trees share one ancestor-masked target pass, while construction, pruning, and scheduling run on the GPU and overlap with the host. Across six vision-language verifiers and 17 text and multimodal benchmarks, D3Tree is fastest in 70 of 72 load settings, with 3.2-4.0x speedup over autoregressive decoding for one request and 1.3-1.8x at 64 concurrent requests, where fixed-budget trees fall below autoregressive speed. Under a time-varying arrival trace, it has the lowest mean end-to-end latency in every arrival phase.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2604.24549

GradMAP: Gradient-Based Multi-Agent Proximal Learning for Grid-Edge Flexibility

Yihong Zhou, Hongtai Zeng, Thomas Morstyn

原文摘要

Coordinating large populations of grid-edge devices requires learning methods that remain decentralised in deployment while still respecting three-phase AC distribution-network constraints. This paper proposes gradient-based multi-agent proximal learning (GradMAP) to address this challenge. GradMAP trains independent neural network policies for each agent that makes online decisions without communication. During offline training, GradMAP embeds a differentiable power grid environment and uses implicit differentiation to propagate exact AC power flow gradients. To speed up training while maintaining performance, GradMAP reuses expensive environment gradients through a proximal surrogate within a trust region defined in the more direct policy-output space, instead of the probability distribution space used in PPO. Using a single workstation-class NVIDIA RTX PRO 5000 Blackwell GPU with 48~GB of memory, GradMAP trains 1,000 agents on the IEEE123 network in 15 minutes and 100,000 agents on the IEEE8500 network in 40 minutes, without parameter sharing. In both cases, GradMAP achieve the lowest mean daily maximum voltage violation and the lowest energy cost among all evaluated methods, including standard multi-agent reinforcement learning methods and gradient-based learning methods extended to the multi-agent setting by us. The training time to convergence is also the lowest, with a 5x to more than 10x training speedup compared to other methods.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v6arXiv:2509.20789

Aligning Inductive Bias for Data-Efficient Generalization in State Space Models

Qiyu Chen, Guozhang Chen

原文摘要

The remarkable success of modern AI has been closely tied to scaling laws, yet the finite supply of high-quality data makes data efficiency--learning more from less--an increasingly important frontier. A model's inductive bias is a critical lever for data efficiency, but foundational sequence models such as State Space Models (SSMs) often rely on fixed, task-agnostic biases. When this fixed prior is misaligned with the underlying structure of a task, the model may require additional samples to overcome its own bias before learning the relevant signal. In this work, we introduce a principled framework for understanding and aligning the inductive bias of linear time-invariant SSMs. We first formalize this bias through an SSM-induced kernel and show theoretically and empirically that its spectrum is governed by the model's frequency response. This characterization motivates Task-Dependent Initialization (TDI), a fast power-spectrum matching method that aligns the initial SSM bias with the task's spectral characteristics before downstream training. Across controlled synthetic experiments, trainable one-layer SSMs, and deep SSMs on diverse real-world benchmarks, TDI can improve data-efficient generalization primarily when task-relevant spectral structure is present and the default SSM bias is spectrally mismatched. Our results provide both a theoretical lens and a practical tool for task-adaptive inductive bias, suggesting a path toward more data-efficient sequence modeling.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11365

MG-Cache: Multi-Granularity Caching for Accelerating Diffusion Transformer Inference

Zesen Chen, Youxuan Xu, Shigang Li

原文摘要

Diffusion Transformer (DiT) models, while having become a powerful backbone for generative synthesis, suffer from high inference cost on generating high-resolution images. Diffusion caching has emerged as a promising acceleration technique to reduce inference cost. However, existing caching strategies fail to accurately identify the varying importance of different computational modules and rarely coordinate cache reusing across the timestep and layer granularities, making it difficult to control error accumulation. To address these limitations, we propose multi-granularity caching (MG-Cache) for DiTs, a novel and fully training-free approach to exploit computational redundancy both on timesteps and Transformer layers. Our core innovation contains a two-stage caching strategy: an offline timestep profiling analysis that identifies redundant timesteps for cache reuse, followed by an adaptive layer-wise caching strategy that combines input stability with importance of each Transformer layer to selectively cache stable Attention and MLP submodules. Extensive experiments on mainstream open-source DiT models validate the effectiveness of our approach, achieving significant acceleration while maintaining high fidelity to the uncached output, with up to a 4.18x acceleration on HunyuanVideo.

cs.LG · cs.PF

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11364

OOM: Offload-Optimized Full-Parameter MoE Training via Memory-Execution Co-Design

Ruijia Yang, Shiyuan Lin, Chengxi Liao, Zeyi Wen

原文摘要

Offloading model state to host memory and streaming layer weights let a single multi-GPU node train large mixture-of-experts (MoE) models. GPU memory then becomes a performance budget, shared by resident expert state, retained activations, and execution worksets. The best allocation depends on the expert execution path, batch size, and sequence length. We present OOM (Offload-Optimized MoE), a full-parameter MoE training system that co-designs expert execution and memory allocation. For expert-parallel execution, OOM updates resident experts on their owning GPUs and jointly plans state residency and MoE activation retention according to pipeline dependencies. For local execution, OOM bounds GPU worksets through dependency-aware partitioning; in hybrid models, it also retains attention results on the host to avoid costly recomputation and selectively regenerates layer boundaries through cheaper layers to make room. Calibrated cost models guide joint configuration within GPU and host budgets to minimize step time; OOM selects among the configured paths through short executions. On four H100s, OOM trains with 1M-token contexts, whereas FSDP2, ZeRO-Offload, and MegaTrain exhaust GPU memory at 512K. It also trains a 122B-parameter MoE model, on which all three exhaust host memory. On four 24 GB RTX 4090s, it trains 512K-token contexts and 1,024 1K-token sequences per step without gradient accumulation. Across matched-global-batch comparisons, OOM delivers 1.58x and 4.08x geometric-mean (up to 2.14x and 7.77x) throughput over CPU-offloaded FSDP2 on the two platforms while using the least host memory among the evaluated systems -- as little as 1/3.0 of FSDP2's.

cs.DC · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11344

Spatial Harness: Grounding Spatial Reasoning in Executable World States with Physics Engines

Chentao Cao, Zhanke Zhou, Sangni Duan, Bo Han, Hang Li

原文摘要

Spatial reasoning requires understanding how objects and observers are arranged in a scene and how their relations change with viewpoint or movement. Vision-language models (VLMs) typically perform both steps through model-internal reasoning, forming an implicit scene interpretation from partial views and imagining the consequences of spatial transformations. A misread scene can lead to a wrong answer, while even a correct interpretation does not ensure that imagined transformations satisfy spatial constraints. We introduce Spatial Harness, which represents scenes as explicit and executable world states to ground spatial reasoning in physics-engine feedback. A Perceiver integrates scene constraints extracted from the question with specialized visual estimates to place objects and cameras in a shared coordinate frame. The resulting state is instantiated in a physics engine, where a Reasoner executes viewpoint changes and uses computed measurements and rendered views to guide reasoning. We evaluate on MindCube, SAT, MMSI-Bench, and SpatialRGPT-Bench with open- and closed-weight VLMs. On MindCube, Spatial Harness improves Qwen3.5-27B from 52.70% to 71.14%, surpassing pySpatial by 7.52 percentage points. Given identical annotated object and camera poses, it reaches 98.25% versus 71.81% for model-internal reasoning. Code is publicly available at: https://github.com/tmlr-group/Spatial-Harness

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11342

Agent Self-Improvement with Frozen Weights: The Added Value and Limits of Failure Traces

Arkid Mitra

原文摘要

We evaluate whether an AI agent's harness, the notes around its model, can learn from the agent's failure traces beyond a matched search that never sees them. It can, but less than it looks. We tested an open 9-billion-parameter model with fixed weights; a larger teacher model wrote its notes. On generated customer-service decisions, where both note writers were told the target rules, notes based on the agent's mistakes improved accuracy by 4.0 percentage points over notes from the same process with the mistakes withheld. That process alone, trying and keeping better-scoring notes, reached about 60% of the total improvement. In full customer-service conversations the same notes showed no demonstrated improvement. Separately, a strict screen of single lessons admitted two customer-service lessons, still awaiting tests on new data, and one tool-use lesson, written without seeing the agent's mistakes, which showed no demonstrated gain on new tool-use tasks. A neutral note with no procedure scored higher than the learned lessons (exploratory). A blind audit judged 146 of 240 tool-use test items defective.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11337

OPSRL-SSP: Optimistic Posterior Sampling for Stochastic Shortest Path with Minimax-Optimal Regret

Yiming Xu, Xian Wei, Cheng Chen

原文摘要

We study online reinforcement learning for the stochastic shortest path (SSP) problem with unknown transition dynamics. Although existing posterior-sampling methods perform well empirically on SSPs, their regret guarantees do not achieve the minimax-optimal order. We address this limitation by proposing OPSRL-SSP, an optimistic posterior-sampling algorithm for the SSP setting. The algorithm operates in epochs and uses a logarithmic, time-dependent schedule for the number of posterior samples. To induce optimism, we augment the transition model with a pseudo-state assigned an optimistic boundary value. We establish a high-probability regret bound of $\tilde{O}(B_\star \sqrt{SAK} + B_\star S^2 A)$, where $B_\star$ is an upper bound on the expected cost of the optimal policy, $S$ and $A$ are the numbers of states and actions, respectively, and $K$ is the number of episodes. In addition, we extend the analysis to general zero-cost SSPs using a universal cost-perturbation argument. The leading term matches the minimax lower bound $Ω(B_\star \sqrt{SAK})$, thereby addressing the open problem raised by Jafarnia-Jahromi et al. [19] for posterior sampling in the SSP setting.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11336

Where Should a Gate Look? Input vs. State Gating in Diagonal State Space Models

Riccardo Zattra, Giacomo Baggio, Leonardo De Marchi, Augusto Ferrante, Francesco Ticozzi, Marco Del Prà, Antonio Orvieto

原文摘要

Modern state-space models owe much of their computational efficiency to their linear, often diagonal, dynamics. Classical recurrent neural networks instead allow nonlinear state-to-state interactions, which can be beneficial when memory updates need to depend on the past context. In this paper, we ask whether state feedback through a simple gate regulating memory updates can enhance the expressivity of diagonal state-space models. To investigate this question, we compare gates that depend on the current input, the recurrent state, or both. We establish a strict expressivity separation: for scalar models, state gating solves induction-head tasks for arbitrary vocabulary size and sequence length, whereas input gating fails on a finite instance under standard stability assumptions on the state transition matrix. Experiments show that state gating retains its advantage on induction-head tasks in higher-dimensional models, whereas input gating performs better on selective copying tasks. Together, our results show that even a simple, diagonal form of state-feedback gating can unlock capabilities unavailable to input-controlled models, while input and state gating offer complementary strengths and achieve the strongest overall performance when combined.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2510.20608

Convergence Analysis of SGD under Expected Smoothness

Yuta Kawamoto, Hideaki Iiduka

原文摘要

Stochastic gradient descent (SGD) is the workhorse of large-scale learning, yet classical analyses rely on assumptions that can be either too strong (bounded variance) or too coarse (uniform noise). The expected smoothness (ES) condition has emerged as a flexible alternative that ties the second moment of stochastic gradients to the objective value and the full gradient. This paper presents a self-contained convergence analysis of SGD under ES. We (i) refine ES with interpretations and sampling-dependent constants; (ii) derive bounds of the expectation of squared full gradient norm; and (iii) prove sublinear convergence rates with explicit residual errors for various step-size schedules. Furthermore, we empirically validate our theoretical framework by training a ResNet-18 model on the CIFAR-100 dataset. We uncover a two-phase dynamic in the gradient noise structure, transitioning from a suboptimality-dominated regime to a geometry-dominated regime. Our treatment unifies and extends recent theoretical threads in the literature.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2512.17473

Alternating Direction Method of Multipliers for Nonlinear Matrix Decompositions

Atharva Awari, Nicolas Gillis, Arnaud Vandaele

原文摘要

We present an algorithm based on the alternating direction method of multipliers (ADMM) for solving nonlinear matrix decompositions (NMD). Given an input matrix $X \in \mathbb{R}^{m \times n}$ and a factorization rank $r \ll \min(m, n)$, NMD seeks matrices $W \in \mathbb{R}^{m \times r}$ and $H \in \mathbb{R}^{r \times n}$ such that $X \approx f(WH)$, where $f$ is an element-wise nonlinear function. We evaluate our method on several representative nonlinear models: the rectified linear unit activation $f(x) = \max(0, x)$, suitable for nonnegative sparse data approximation, the component-wise square $f(x) = x^2$, applicable to probabilistic circuit representation, and the MinMax transform $f(x) = \min(b, \max(a, x))$, relevant for recommender systems. The proposed framework flexibly supports diverse loss functions, including least squares, $\ell_1$ norm, and the Kullback-Leibler divergence, and can be readily extended to other nonlinearities and metrics. We illustrate the applicability, efficiency, and adaptability of the approach on real-world datasets, highlighting its potential for a broad range of applications.

eess.SP · cs.LG · math.OC · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11332

CVaR Deep Optimal Stopping

Yechan Jeong, Youngjoon Hong

原文摘要

Risk-sensitive optimal stopping is useful when poor stopping outcomes carry substantial costs, as in option exercise, sequential medical treatment, and safety-critical control. A natural objective is Conditional Value at Risk (CVaR) of the stopped reward. We develop a theoretical framework and algorithm for CVaR deep optimal stopping. The framework and algorithm address four difficulties in applying risk-sensitive reinforcement learning to optimal stopping: state augmentation, score-function pathology, blindness to success, and dependence on auxiliary networks. We isolate three structural properties of optimal stopping: single-point reward, action-independent dynamics, and sum-zero gradient. We exploit these properties to establish Markov optimality without state augmentation and derive a pathwise gradient estimator. Under bounded policy derivatives, the pathwise estimator has bounded second moments near deterministic policies, avoiding the divergence that can occur with score-function estimators. Building on this framework, we propose DIOS (Distribution-Informed Optimal Stopping) to address the remaining two difficulties. A per-time-step soft penalty combined with tail-level annealing mitigates blindness to success. Estimating the reward quantile directly from each batch removes the need for an auxiliary critic. For a fixed tail level, we establish a finite-time stationarity guarantee under regularity assumptions along the training iterates. Across the 24 main Bermudan max-call configurations, DIOS achieves the highest mean CVaR and the shortest recorded training time among the tested CVaR baselines.

math.OC · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2510.00666

Blind Image Restoration via Merging the Manifold Hypothesis, Kernel Method and a Diffusion-like Flow

Leah Bar, Liron Mor Yosef, Shai Zucker, Neta Shoham, Inbar Seroussi, Nir Sochen

原文摘要

Blind Image Restoration (BIR) aims to restore degraded images without knowledge of either the type or severity of the degradation. This is a challenging ill-posed inverse problem for which an image prior is essential. We view the prior of an image class as a probability distribution supported on a low-dimensional manifold embedded in the high-dimensional image space. Existing generative approaches primarily model the probability distribution, while the geometric structure is implicit or disregarded. Conversely, geometric approaches do not account for the non-uniform probability distribution on the manifold. We propose a geometric-probabilistic framework for BIR that models these aspects. The probability distribution on the manifold is approximated using a kernel-based method, while a distance from the manifold function induces a geometry-based probability function on the ambient space. This formulation leads to the (Latent) Manifold-Probabilistic Projection Model ((L)MPPM), a restoration model defined in both image and latent spaces. (L)MPPM induces a flow guided by a score field that jointly incorporates geometric and probabilistic information. Evaluated on image restoration tasks and datasets, (L)MPPM consistently outperforms existing approaches, providing empirical evidence for the benefit of incorporating both geometric and probabilistic information into the image prior.

cs.CV · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11330

When Credit Turns Against Tools: Transition-Aware Credit Assignment in Agentic Reinforcement Learning

Xiangyu Lu, Zhanke Zhou, Jiazhe Ning, Chentao Cao, Jiangchao Yao, Bo Han

原文摘要

Agentic reinforcement learning with verifiable rewards trains language models through multi-turn tool interactions, yet their reasoning performance and tool-calling ability can decline during training. We trace this degradation to three coupled mechanisms: (1) tool-token imbalance produces negative overall tool credit when incorrect trajectories contain more tool tokens per turn than correct ones; (2) aggregation amplification can further strengthen this negative signal as token-mean aggregation gives longer incorrect trajectories greater weight in policy updates; and (3) failure-driven retries prolong trajectories as models tend to call tools again after execution errors, extending potential negative credit to recovery attempts. To counter this suppression, we propose Transition-Aware Credit Assignment (TACA), which targets the transition from reasoning to tool use through supervision of both call initiation and generation. Specifically, TACA uses Advantage Projection to enforce a nonnegative mean advantage on tool-call tokens within each prompt group through a minimal uniform shift. Additionally, TACA guides tool-use decisions with Decision Contrast, comparing rewards for tool-assisted and direct answers to the same prompt. Through Supervised Replay, it reinforces tool-call generation using calls from previously sampled correct trajectories. Across models and tasks, TACA improves performance and sustains more reliable tool interaction during training, raising mean accuracy over MT-GRPO on six reasoning benchmarks by 4.61 percentage points for Qwen2.5-1.5B and 3.61 for Qwen2.5-3B. Our code is available at https://github.com/tmlr-group/TACA.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2601.20983

Monotone Optimisation with Learned Projections

Ahmed Rashwan, Keith Briggs, Chris Budd, Lisa Kreusser

原文摘要

Optimisation algorithms typically assume oracle access to objectives and constraints, yet in many applications these functions are available only through data. To recover them, a common approach learns a surrogate for each function and optimises over it, but this can be unnecessarily expensive and may fail to preserve the structure required by the optimisation algorithm. We study this problem for monotone optimisation, a class of non-convex problems which can be solved globally using the Polyblock Outer Approximation (POA) algorithm. Our key observation is that POA does not require the constraint functions themselves, but only a scalar quantity we call the radial inverse. Learning this quantity directly replaces POA's inner bisection loop over a learned constraint with a single radial inverse evaluation. We characterise exactly which maps arise as radial inverses of monotone constraints, and relate their approximation error to perturbations of the feasible set and optimal objective value. Building on these results, we introduce Homogeneous-Monotone Radial Inverse (HM-RI) networks, structured min-max models that enforce the monotonicity and positive-homogeneity of radial inverses by construction. Across synthetic monotone optimisation problems and a cellular transmit-power application, HM-RI improves solution quality over direct constraint surrogates and local optimisation baselines by up to 58.6 percentage points relative to the Oracle reference objective, while achieving a 3.0-4.7x speed-up over bisection-based POA with learned constraint surrogates, with the largest solution-quality gains occurring in data-limited and piecewise-constant settings.

cs.LG · math.OC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11324

When Sensors Fail: Coverage Auditing with Incomplete Sensor Records

Chung-I Huang, Wen-Yi Chang

原文摘要

Sensor outages can conceal prediction-interval coverage deficits behind acceptable marginal performance, while unavailable future readings leave some issued forecasts unscorable. We present a measurement-record audit that links issue-time outage duration, original target availability, and time-uniform statistical evidence. The audit retains complete forecast keys, batches all sensors and overlapping horizons at each measurement tick, and projects simultaneous covered-count and scored-count bounds onto a cumulative conditional expected coverage ratio. This target is distinguished from empirical coverage and outage-conditional population coverage. Selected traffic and air-quality results illustrate the gap between pooled and long-outage coverage; in METR-LA, only 7.6% of issued long-gap forecasts have originally observed targets. Of 30 primary sequential outage-state audits, 26 are inconclusive and four lack scored support. This concise presentation gives the core construction and selected aggregate evidence without the full experimental supplement.

cs.LG · stat.AP

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11318

PROLA: Principal-Orthogonal Low-rank Adaptation for Predictive Spatiotemporal Weather Downscaling

Minseo Yoon, Minseong Bae, Sojin Lee, Hyunwoo J. Kim

原文摘要

High-resolution weather prediction is important for resolving local atmospheric patterns missed by coarse global forecasts. Large pretrained weather backbones have improved global forecasting, but adapting them to produce future high-resolution fields from recent coarse states remains challenging and costly. We formulate this setting as Predictive Spatiotemporal Weather Downscaling (PSWD), a task that jointly refines spatial and temporal resolution while predicting future high-resolution trajectories. To this end, we propose a framework that decomposes prediction into spatial and temporal stages using a shared pretrained backbone, fixed numerical scaffolds, and lightweight residual heads. For backbone adaptation, we introduce PROLA (PRincipal-Orthogonal Low-rank Adaptation), which allocates a fixed low-rank budget between pretrained weights' principal directions and their orthogonal complement. PROLA further rescales rank-1 optimizer updates using gradient signals, improving adaptation without increasing trainable rank. Experiments in the PSWD setting show that PROLA outperforms representative low-rank adaptation baselines. Further experiments across diverse datasets and tasks support the broader applicability of PROLA, showing that its benefits extend beyond the primary weather downscaling setting.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.11313

Shape-Adaptive Subspace-Constrained Quadratic Matrix Factorization

Zheng Zhai, Zhuan Liang, Sujun Guo

原文摘要

Quadratic manifold fitting provides a flexible framework for modeling nonlinear low-dimensional structures embedded in high-dimensional spaces. However, existing methods typically employ prescribed noise models or fixed loss functions, limiting their adaptability to varying noise characteristics. To address this limitation, we propose an alternating maximum likelihood estimation method for subspace-constrained quadratic matrix factorization, termed AMLE-SQMF, which jointly estimates the quadratic manifold and the underlying noise distribution. Specifically, we model the residuals using a generalized Gaussian distribution (GGD), whose shape parameter controls the tail behavior and determines the associated fitting loss. By exploiting the closed-form maximum likelihood estimator of the scale parameter, we eliminate the scale variable and derive a profile negative log-likelihood. Based on this formulation, we develop an alternating optimization algorithm that iteratively updates the manifold parameters and estimates the noise shape parameter. This enables adaptive loss selection within a unified likelihood-based framework, without requiring prior specification of the noise characteristics. Experiments on synthetic and real-world datasets demonstrate the effectiveness of AMLE-SQMF in quadratic manifold fitting and its robustness to diverse noise distributions.

cs.LG

查看论文
arXiv ↗PDF ↗