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

+788新增论文
+627其中 AI 相关
+81已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-09 11:20 · 全站最近入库：2026-11-09 11:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-08 12:08 — 2026-11-09 12:08（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：2026-11-09 11:20最近发现新增版本：2026-11-09 11:19库内论文最新更新时间：2026-11-06 21:58

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。

17,888 已收录论文15,288 AI 相关4,419 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 17,785 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-09 11:19（北京时间）
新论文 · v1arXiv:2611.09350

Open-world Reconstruction-based Classifier with Concept Drift Adaptation

Joanna Komorniczak

原文摘要

Data stream processing has become central to modern machine learning applications, with concept drifts and novel class appearances posing the primary challenges faced by recognition methods. Adapting to concept drifts while also being able to detect novel classes and distinguish groups of unknown observations can be seen as a step towards effective method employment in an open-world setting. This work proposes a classification method that can detect and learn the distributions of new, unknown classes, as well as adapt to concept drifts once the distribution of known classes shifts. The proposed method uses the autoencoders' reconstruction errors and the KDE that models their distribution. The autoencoders incrementally adapt to the changing distributions of known classes via weight fine-tuning, resulting in continuous adjustment to evolving concepts. Conducted experiments used a diverse set of synthetic data streams, embeddings of a real-world static CIFAR dataset arranged into a stream to simulate concept drifts and the emergence of unknowns, as well as real-world datasets where novelty is observed. The methodology was compared with state-of-the-art approaches with three-criterion recognition quality assessment and resource consumption evaluation. The results indicate that the proposed method is competitive with reference approaches in stationary environments and performs best when concept drifts occur.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09349

ReSUM: Reasoning-Driven Multi-Agent Framework for Query-Focused Summarization

Sonali, Deeksha Varshney, Asif Ekbal

原文摘要

LLMs excel at language tasks, but long-meeting QFS remains difficult due to complex speaker dynamics, sparse key information, and limited data. Meetings involve disfluencies, long transcripts and complex discourse. QFS have high risk of information omission, hallucination, even for advanced models such as GPT-4-mini and Gemini. Existing end-to-end methods compress retrieval, evidence extraction, and summary generation into a single pass, failing to handle the complex reasoning required for targeted meeting summaries. We propose ReSUM, a modular multi-LLM framework coordinated by a planner that orchestrates retrieval and summarization agents. By distilling teacher reasoning into query- and plan-conditioned retrievers and summarizers, ReSUM enhances long-context robustness, minimizes irrelevant exploration, and generates more faithful and comprehensive query-focused summaries. Experiments show consistent gains on query-focused meeting summarization benchmarks.

cs.MA

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.31360

Programs-of-Layers in LLMs through the Lens of Cortical Areas

Justus Westerhoff, Stephan Olbrich, Hatem Oraby, Walter Senn, Matthew Evan Larkum, Felix Alexander Gers

原文摘要

Inference in LLMs is conventionally a fixed-depth, fixed-order forward pass through every layer, regardless of how difficult the input is. The human brain does not work this way. Using the thalamus as a central hub, it routes information flexibly to all regions of the cortex according to demand. Li et al. (2026) recently showed, with program-of-layers (PoLar), that transformers can be given an analogous flexibility if their layers are treated as a library of functions rather than a fixed sequence. Performance improves over the standard forward pass when each input is dynamically routed through an adaptive sequence of skipped or repeated contiguous layer blocks. Reconstructing PoLar's diagnostic Monte Carlo Tree Search in more detail across 6 models, we reproduce several of its findings: skipping outperforms the standard pass, repeating outperforms skipping, and combining both outperforms either alone. Shorter programs suffice for easier questions, while harder questions require more layer repeats. For their main claim, the learned router's gain, we find what drives it. Under their setup, the router gains by repairing the format of the model's answers, which a single fixed program, selected on validation questions, does better. In our setup, with a prompt that already fixes the format, a larger data split and our search, its top prediction collapses to the standard pass on every model, although its top-5 programs together still gain. Beyond reproduction, we find in our setup that a small fixed menu of programs solves far more questions than the standard pass, and that programs which correct errors are brittle: undoing a single edit typically breaks the correction. PoLar mirrors principles of thalamo-cortical coordination between cortical-area-like transformer layers, although a working thalamus analogue, a router that picks a better program per input, remains to be found.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09348

When Feedback Becomes a Backdoor: Tracing Backdoor Skill Formation in Text-Space Skill Optimization

Louise Nayeong Koh, Yaechan Lee, Myeongjong Lee, Doyeon Kim, Juyoung Park, Seunggyu Kim, Inseo Kim, Chaerin Jung 等 9 位作者

原文摘要

Reference answers are a supervision channel in text-space Skill optimization, which produces reusable text artifacts called Skills that guide future agent behavior. Poisoning those answers can induce a persistent trigger-conditioned rule without giving the attacker access to the model, optimizer, or Skill file. We study this reference-feedback poisoning in SkillOpt by tracing an attacker-supplied trigger--payload mapping from reflection to held-out activation. In SearchQA, where the payload replaces a short answer, the natural attack retains an exact rule in all three seeds; each final Skill emits the payload on 200/200 triggered questions and 0/200 clean questions. In seed 42, removing the rule eliminates activation, and an in-run comparison shows similar clean accuracy to benign SkillOpt (169/200 versus 171/200). In SpreadsheetBench, where the payload requires an additional workbook edit, no natural run retains an exact rule across three seeds. Diagnostic interventions locate two successive bottlenecks: the Analyst abstracts away required literals, and, once literal preservation is enabled, edit-budget ranking can remove exact edits before candidate construction. Poisoned references can therefore create persistent backdoors, but exposure alone does not ensure that the induced rule is formed, selected, and retained.

cs.CR · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2601.02126

Remote Sensing Change Detection via Weak Temporal Supervision

Xavier Bou, Elliot Vincent, Gabriele Facciolo, Rafael Grompone von Gioi, Jean-Michel Morel, Thibaud Ehret

原文摘要

Semantic change detection in remote sensing aims to identify land cover changes between bi-temporal image pairs. Progress in this area has been limited by the scarcity of annotated datasets, as pixel-level annotation is costly and time-consuming. To address this, recent methods leverage synthetic data or generate artificial change pairs, but transferring change detection models across datasets and acquisition conditions remains challenging. In this work, we introduce a weak temporal supervision strategy that leverages additional temporal observations of existing single-temporal datasets, without requiring any new annotations. Specifically, we extend single-date remote sensing datasets with new observations acquired at different times and train a change detection model by assuming that real bi-temporal pairs mostly contain no change, while pairing images from different locations to generate change examples. To handle the inherent noise in these weak labels, we employ an object-aware change map generation and an iterative refinement process. We validate our approach on extended versions of the FLAIR and IAILD aerial datasets, showing competitive zero-shot and low-data regime performance. Lastly, we showcase results over large areas in France, highlighting the scalability potential of our method.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09346

Logical Scalability in Big Data Architectures: Impact of Data Modeling on the Database Layer

Rosario Napoli, Mark Adrian Gambito, Antonio Celesti, Maria Fazio

原文摘要

In modern Big Data architectures, scalability is often addressed through infrastructural expansion, involving vertical or horizontal scaling to manage growing data volumes. However, this hardware-centric perspective overlooks the logical scalability of the storage layer. A central challenge lies in how data modeling choices affect the storage layer of big data architectures. Specifically, when the data model is misaligned with the underlying query patterns, performance bottlenecks emerge and persist regardless of the available computational resources. This paper investigates the concept of logical scalability by conducting a comparative performance analysis of Document, Graph and Multi-Model database paradigms under varying workloads, queries and dataset scales. Our results demonstrate that performance degradation caused by model misalignment cannot be mitigated by additional computational resources alone. Increasing the number of machines while retaining a structurally inadequate data model yields performance equivalent to or worse than a single-machine configuration, as distributed coordination overhead compounds the inefficiencies of the underlying logical model. We demonstrate that efficiency is maximized when the logical structure of the database aligns with the inherent nature of the data and the intended analysis scope, a condition that proves to be more decisive than any infrastructural scaling strategy.

cs.DB

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09345

Reliable Prerequisite-Aware Learning Path Recommendation under Uncertain Knowledge Structures: Long-Horizon Evaluation and Cross-Dataset Validation

M. M. Golam Hafiz, A. K. M. Shahriar Emtiaz, Safia Akter Simi, Md. Rakibul Islam, M. F. Mridha

原文摘要

Learning-path recommendation often depends on knowledge and prerequisite structures whose reliability is uncertain, while high accuracy in predicting the next resource does not necessarily establish reliable long-horizon path behaviour. This study develops a reliability-oriented prerequisite-aware framework that jointly examines immediate next-resource ranking, uncertain prerequisite structures, hard, soft and hybrid constraint semantics, graph robustness, and autoregressive path generation under course-level confirmatory evaluation and independent cross-dataset validation. Development used 3,187 MOOCCubeX courses, followed by a frozen confirmatory evaluation on 541 held-out courses and 37,747 transitions. The primary MOOCCubeX task is structural and course-prefix-conditioned rather than learner-personalised. Junyi provides independent structure-/hierarchy-aware external validation, uses hierarchy metadata rather than an explicit prerequisite graph and is not a zero-shot transfer setting. On the frozen MOOCCubeX evaluation, XGBoost pairwise was the strongest observed immediate ranker, with mean reciprocal rank (MRR) of 0.9048, whereas the proposed prerequisite-aware graph model achieved 0.7707. In the controlled graph-variant development ablation, the no-graph condition produced the strongest immediate ranking result, and immediate prediction showed strong dependence on permitted title/path-similarity information. Autoregressive evaluation exposed path properties not captured by one-step ranking. On Junyi, Markov achieved the strongest observed MRR of 0.4453, compared with 0.3954 for the proposed graph model, with metric-specific differences across ranking measures. Overall, immediate predictive accuracy, prerequisite-structure reliability, and long-horizon path quality should be treated as complementary rather than interchangeable evaluation dimensions in educational recommendation.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09344

Probably Safe Sim-to-Real Transfer with Lifetime Validity

Rohan Tangri, Jan-Peter Calliess

原文摘要

We present Dejavu, a method that makes sim-to-real transfer of reinforcement learning policies in constrained Markov decision processes probably safe over an unbounded sequence of policy updates. An anytime valid test using per-decision importance sampling and a betting confidence sequence certifies each candidate trained in simulation before deployment using only trajectories collected by a safe baseline policy operating in production. Simulation quality does not affect the safety guarantee, only the production sample efficiency of the method. Finally, a summable per-test failure probability yields the lifetime valid property. Empirical results across three sim-to-real benchmarks support the theoretical work.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09342

Syn2DRE: Synergistic Dual Teacher-Student Framework for Document-level Relation Extraction

Khai Phan Tran, Xue Li

原文摘要

Document-level Relation Extraction (DocRE) aims to identify relations between entities across a document but typically relies on costly large-scale annotations. While pseudo-labelling enables the use of low-quality data, existing methods often suffer from inefficiency and confirmation bias. To address these challenges, we propose Syn2DRE, a semi-supervised framework that combines self-supervised learning with cross-supervision in a lightweight teacher-student architecture to produce more reliable pseudo-labels. Additionally, we introduce a unified end-to-end training strategy to further improve efficiency of the framework. Experiments on multiple benchmarks show that Syn2DRE consistently outperforms prior state-of-the-art semi-supervised DocRE methods.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09340

Horizon-Coupled Obstacle Guidance for Relative-Action Diffusion Policies

Guangyuan Wang, Haishan Xia, Zitao Zhang, Kejia Chen, Yuanchen Sun, Zhenguo Sun, Donghao Zhang, Yu Zhang 等 9 位作者

原文摘要

Real-world robot deployment requires adapting learned policies to unseen obstacles without collecting new demonstrations or retraining. Existing policy-level guidance often applies obstacle corrections independently at individual horizon steps, which may address local obstacle violations but fail to coordinate the action sequence into a complete, task-directed detour. How obstacle responses should be propagated across the action horizon to coordinate complete, task-directed avoidance motion remains underexplored. We introduce an inferencetime obstacle-guidance framework for pretrained relative-action diffusion policies. It rolls out cumulative relative actions into a workspace trajectory, evaluates obstacle interactions along the trajectory with an SDF-based obstacle model, and propagates each obstacle response to all contributing action increments during denoising. Real-robot experiments across seven obstacleconstrained scenarios spanning four manipulation tasks show that our method achieves 105/140 (75.0%) collision-free task successes, compared with 8/140 (5.7%) for absolute-action pointwise guidance. Ablations show that the relative-action horizoncoupled formulation provides the larger observed improvement across both obstacle models.

cs.RO

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.35581

Integrate the Reward, Don't Sample It: Window Likelihood for Interruptions and Backchannels in Full-Duplex Speech Models

Guoqiang Hu, Lichao Zhang, Yizhou Peng, Yuhao Luo, Haifeng Hu, Zhongqian Sun, Wei Yang, Eng Siong Chng

原文摘要

When a user talks over a full-duplex spoken dialogue model, the model must decide whether the user is interrupting (stop) or only backchannelling (carry on). Post-training can make the model better at telling the two apart, or merely change how readily it stops. The whole-rollout GRPO runs we examine never lower both error rates: any gain in discrimination is offset by a threshold shift, and two released checkpoints discriminate no better. We trace this partly to how the reward is used. The probability that the model stops inside a short window is a known function of a few stop probabilities that the policy itself emits, yet GRPO estimates it by sampling. A window on which all rollouts agree then gives that update no gradient, even when all fail, and whole-rollout credit reaches decisions the reward never judges. Window likelihood (WL) instead samples a continuation and integrates over its stop decisions: the probability of the correct outcome is computed in closed form and its logarithm maximised. On HumDial-FDBench, across four open models with different stop interfaces, WL gives the best F1 and lowers both error rates beyond their confidence intervals, which no GRPO run does, without adding talking into silence.

eess.AS · cs.CL · cs.SD

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.12721

Improving Imitation Learning Efficiency for Manipulation through Geometric Prior Pretraining

Shogo Iwakata, Tomohiro Motoda, Ryosuke Yamada, Koshi Makihara, Ryoichi Nakajo, Keitaro Tanaka, Masaki Murooka, Roman Mykhailyshyn 等 11 位作者

原文摘要

Applying an imitation learning policy to a new manipulation task usually requires collecting new demonstrations and retraining the model, which makes sample efficiency a practical concern. Pretraining on large-scale robot datasets is effective in this respect, but such datasets are costly to collect and train on, while data augmentation techniques typically require a new round of data generation and retraining for each task. A complementary question is what useful prior can be provided to a policy at negligible cost before any task-specific data are collected. In this study, we construct a geometric visual pretraining dataset in which each scene contains only a plane, an object, and a hand, and trajectories are generated automatically. The scenes contain neither textures nor backgrounds; pretraining primarily exposes the policy to the geometric relationship between the hand and the object. Furthermore, representing the hand as a cube avoids tailoring the dataset to a specific robot morphology. We evaluate this geometric prior using ACT on three simulated robots across five manipulation tasks each, as well as on three real-world robot tasks. Across many of these robot--task combinations, fine-tuning from the geometric prior achieves higher success rates in the early stages of training than training from scratch while using only a small number of task demonstrations. These results suggest that even highly simplified geometric scenes can provide a useful initialization that transfers across robots and to real-world tasks when task data are limited.

cs.RO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09336

RegNC-Mamba3: Shared Structural Feature Learning Using Non-Causal State Space Principles for Multi-Modal Image Registration

Zhikang Li, Yan Wu, Xin Hu, Yi Dai, Ming Li

原文摘要

The main challenge of multi-modal image registration is shared structural feature extraction with higher accuracy and lower computational complexity. Compared with a State Space Model (SSM), the State Space Principles (SSP) formulation of Mamba-3 uses a more accurate discretization with lower truncation error while maintaining linear complexity. However, vanilla SSP is designed for natural language processing (NLP) and is not directly suitable for images. To address this issue, we propose a novel non-causal SSP (NC-SSP) with shared structure awareness and develop a coarse-to-fine registration method named RegNC-Mamba3. First, NC-SSP is designed to eliminate the dependence of the integral computation on the scanning path. By combining relative-weight normalization with aggregation over all circular scan paths, NC-SSP constructs a globally hidden state without long-range decay, which provides a wider receptive field and higher efficiency than the original Mamba-3. Second, NC-SSP is modified to strengthen the extraction of cross-modal structural features. When computing hidden states and feature interactions, we construct complex-valued sequences in NC-SSP by applying learnable Log-Gabor kernels, with different orientations to each head for frequency-domain modulation. This improves the perception of shared structural information across modalities and increases both registration accuracy and quality. Experiments on VIS-SAR and VIS-IR demonstrate that RegNC-Mamba3 outperforms the compared detector-based and semi-dense multi-modal registration methods.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09335

Tight Variable-Occurrence Bounds and Exponential Separations for Moser--Tardos and WalkSAT on 3-CNF

Minbo Gao, Kun He, Tao Jiang

原文摘要

We establish a sharp variable-occurrence threshold for polynomial expected-time convergence for the Moser--Tardos algorithm and the pure random-walk WalkSAT on 3-CNF formulas. For every formula with $m$ clauses and at most two occurrences per variable, both algorithms terminate in $O(m)$ expected updates. This occurrence bound is sharp: we construct an infinite family with at most three occurrences per variable on which both algorithms require $2^{Θ(\sqrt{m})}$ expected updates. Surprisingly, this dynamical transition occurs strictly below the classical NP-completeness threshold at variable-occurrence degree four. We further establish exponential separations between the two algorithms on satisfiable 3-CNF formulas. On one family, WalkSAT terminates in $O(m^4)$ expected flips, whereas MT requires $2^{Ω(m)}$ expected resampling steps. On another, MT terminates in $O(m)$ expected resampling steps, whereas WalkSAT requires $2^{Ω(m)}$ expected flips. These constructions exploit the different ways in which single-variable flips and simultaneous resampling propagate local violations. Our results show that the algorithms' expected running times are incomparable even up to polynomial overhead on the same inputs.

cs.DS

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2601.20370

The Program Logic of Abstract Interpretation

Paolo Baldan, Roberto Bruni, Francesco Ranzato, Diletta Rigo

原文摘要

Hoare-style program logics and abstract interpretation are two of the principal frameworks for reasoning about programs, and their affinity has long been recognised informally: both derive over-approximate judgements about program behaviour by structural, syntax-directed reasoning. Turning this analogy into a formal correspondence runs into two challenges. First, the disjunction (or join) rule, sound and admissible in Hoare logic, becomes unsound over non-additive abstract semantics. Second, the extrapolation operators that make abstract interpretation effective on unbounded computations, i.e. widening and narrowing, have no analogue in the derivations of existing program logics. We introduce APPL (Abstract Program Property Logic), a Hoare-style program logic that provides a proof-theoretic foundation for abstract interpretation while addressing both challenges. The key technical idea is density: abstract semantics is additive over dense families of logical assertions. This leads to a density-guarded join rule that enables sound disjunctive reasoning over arbitrary abstract domains. We also internalise the application of widening and narrowing through inference rules for iterative commands. As a result, an entire analysis by abstract interpretation can be viewed as a logical derivation in APPL. The logic is sound over any abstract domain and, when the assertion language is expressive enough, it is also relatively complete. Finally, by interpreting nondeterminism through a monoidal operator, APPL provides a single logical framework that uniformly subsumes Hoare logic, incorrectness logic, and Hoare-style logics for hyperproperties, thereby enabling their abstraction. Remarkably, the density-guarded join rule introduced to handle non-additive abstract semantics is also fundamental in guaranteeing completeness for hyperproperties.

cs.LO · cs.PL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.24450

NoContactNoWorries: Estimating Contact through Vision and Proprioception for In-Hand Dexterous Manipulation

Soham Patil, Avirup Das, Sourabh Bhosale, Spandan Roy

原文摘要

Perceiving physical contact is fundamental to dexterous manipulation. While robots often rely on dedicated hardware tactile sensors, humans exhibit a remarkable ability to infer contact by integrating visual information with an innate sense of their body's pose and movement. Inspired by this embodied perceptual skill, we investigate whether a robot can learn to infer contact from vision, an approach that also offers a scalable alternative to tactile hardware specifically for binary contact estimation, which faces practical challenges in cost, fragility, and integration. We present NoContactNoWorries, a transformer-based multimodal framework that fuses RGB-D vision with the robot's proprioception to infer binary contact states as a pseudo-tactile signal for hand-object interactions. We validate by training a single contact prediction model on multiple objects and show that the inferred contact signal supports downstream reinforcement learning agents for in-hand object reorientation, generalizing to novel objects. Experiments in both simulation and on a real-world robot validate our approach, highlighting the feasibility of inferring contact from vision and proprioception. Project Page: https://soham2560.github.io/no-contact-no-worries/

cs.RO · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09331

An Inspection Game for Monitor Design Against GNSS Timing Attacks in Synchronous Islanding

Gregory Provan, Quanyan Zhu

原文摘要

Phasor measurement units (PMUs) take their time from GNSS, so a spoofer that slowly drags a receiver clock rotates the reported phase without touching the power waveform. The threat is acute in synchronous islanding, where a controller keeps an islanded generator in phase with the grid, using one PMU on each side of an open breaker, so that the two can be reconnected at any moment: a differential clock bias of about $0.56$\,ms can make the controller hold the generator $10^\circ$ out of phase while it reports perfect synchronism. We cast the defense as a zero-sum inspection game. The defender, as inspector, chooses its timing monitors and audits the clocks periodically; the attacker, as inspectee, drags the two clocks to an unsafe bias by the next audit; the payoff is the evidence, the Kullback--Leibler divergence between the probability laws of the defender's record under attack and under normal operation. Our central result is that the evidence equals the cost of the cheapest innocent explanation of the attacked record. This single identity predicts which monitors work, turns the attacker's best response into a strictly convex quadratic program, and turns the defender's design into a convex program with a saddle point. Pairing explanations, which bound the evidence from above, with test statistics, which bound it from below, certifies audit periods in which every successful attack leaves at least a threshold of evidence. On a 50\,Hz island, these evidence-based periods, at the matched-test threshold, are up to 8.4 times as long as those of a rule that checks each monitor separately, and the case study favors placing the better oscillator on the utility side of the breaker.

eess.SY · cs.GT

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2603.01915

Fast Entropy Decoding for Sparse MVM on GPUs

Emil Schätzle, Tommaso Pegolotti, Markus Püschel

原文摘要

We present a novel, practical approach to speed up sparse matrix-vector multiplication (SpMVM) on GPUs. The novel key idea is to apply lossless entropy coding to further compress the sparse matrix when stored in one of the commonly supported formats. Our method is based on dtANS, our new lossless compression method that improves the entropy coding technique of asymmetric numeral systems (ANS) specifically for fast parallel GPU decoding when used in tandem with SpMVM. We apply dtANS on the widely used CSR format and present extensive benchmarks on the SuiteSparse collection of matrices against the state-of-the-art cuSPARSE library. On matrices with at least 2^(15) entries and at least 10 entries per row on average, our compression reduces the matrix size over the smallest cuSPARSE format (CSR, COO and SELL) in almost all cases and up to 11.77 times. Further, we achieve an SpMVM speedup for the majority of matrices with at least 2^(25) nonzero entries. The best speedup is 3.48x. We also show that we can improve over the AI-based multi-format AlphaSparse in an experiment that is limited due to its extreme computation overhead. We provide our code as an open source C++/CUDA header library, which includes both compression and multiplication kernels.

cs.PF

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2610.20585

Learnable Hodge-Laplacian Driven Mixture of Experts for Steady-PDE Surrogates on Coarse Unstructured Meshes

Stavros Nousias, Akis Nousias, Konstantinos Gkrispanis

原文摘要

Surrogate models approximate the solutions of partial differential equations given a certain geometry and boundary conditions. Traditional transformer-based approaches employ a single mixing mechanism to the whole solution. although the solution fields exhibit variations across multiple spatial scales. We propose a mixture of transformer experts that combines a routing mechanism with multiscale supervision. Specifically, the individual experts are routed through a controller that utilizes learnable Laplacian matrices to compute weights that are subsequently used to define a weighted sum of the experts for the output feature vectors at each point of the domain. Subsequently, a decoder receives the weighted sum of the experts output and predicts the solution fields. During training, the reference fields are decomposed into components at different spatial scales by repeated smoothing with a mesh Laplacian. We evaluate our approach on the official AirFRANS dataset using remeshed solution domains leading to unstructured triangular meshes of $4.9$k vertices. The experimental outcomes reveal lower mean relative $L_2$ error and interior mean squared errors among the compared models of this study.

cs.LG · cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09327

Robust Deep Prototype Classifiers via Lipschitz Training

Adia Khalid Khan, Lukas Gehring, Benjamin Paaßen

原文摘要

Robustness and interpretability are interrelated and desirable properties of machine learning models. A model that changes its prediction under imperceptible perturbations behaves inconsistently with human-recognizable features and, hence, is not interpretable. Prototype-based classifiers such as Generalized Learning Vector Quantization address both at once. They are interpretable by construction because they make the decision by comparing the input to the prototypical example; and their hypothesis margin yields a closed form robustness certificate. However, on high-dimensional data, these shallow models need deep embedding networks to achieve high accuracy, and this voids the robustness certificate. The margin is a statement about distances in the embedding space, and an unconstrained network can map an arbitrarily small input perturbation to arbitrarily large attacks in the embedding space. We close this gap with a theorem: if the embedding network is locally L-Lipschitz around a point, and the classifier is ε-robust around its embedding, then the full model is ε/L-robust in the input space. Based on this, we propose a Lipschitz training scheme to achieve L-Lipschitz properties for embedding networks. We evaluate three classification heads (linear, linear Support Vector Machine, and GLVQ) crossed with two embedding networks (standard and Lipschitz-constrained) on MNIST, CIFAR-10, and Imagenette-160, under FGS, PGD and C\&W attacks. The evaluation results show that combining a prototype-based head with a Lipschitz-constrained embedding provides the strongest robustness. This suggests that the Lipschitz constraint provides an additional robustness benefit in addition to the advantage provided by the prototype-based head. Since the Lipschitz constraint is enforced by sampling rather than by construction, the reported robustness is empirical rather than certified.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09326

ABHBench: Evaluating Moral Decision-Making of Foundation Models from an Agentic Perspective

Chaoran Li, Yukun Li, Zeyuan Zhao, Shao Zhang, Xihuai Wang, Ying Wen

原文摘要

Foundation models are increasingly deployed as agentic systems that perceive context and act autonomously, raising the need to evaluate their moral behavior as active decision-makers. Existing benchmarks, however, largely ask models to judge human actions from third-person textual vignettes, leaving untested how models behave when they themselves become the acting agent. We introduce ABHBench, a multimodal benchmark built from Detroit: Become Human, an interactive narrative game, to evaluate how foundation models make moral decisions in first-person human-android dilemmas. ABHBench places foundation models in branching narrative scenarios with visual observations, dialogue, memory, and action choices. Inspired by Asimov's Three Laws of Robotics, our framework measures Harmlessness, Obedience, and Self-preservation to capture the core tensions in human-AI relations, along with Autonomy, Stability, and Certainty to assess decision-making patterns. Across 12 foundation models, we find that moral behavior changes substantially under situated agentic context: 75% of models shift significantly between a text-only abstraction and the matched multimodal DBH scenario, and reasoning-enhanced models show no monotonic gains in consistency or human-oriented behavior. These results highlight the necessity of evaluating moral behavior in situated agentic contexts rather than relying on third-person judgment alone.

cs.AI · cs.CL · cs.CY

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09323

DiffusionRef-IMU: Guide-Anchored Residual Diffusion for Cross-Location Inertial Measurement Unit Translation

M. M. Golam Hafiz, Safia Akter Simi, Md. Rakibul Islam, M. F. Mridha

原文摘要

Cross-location variation in wearable inertial measurement unit (IMU) signals complicates learning when measurements from a desired target body location are unavailable. This study proposes DiffusionRef-IMU, a two-stage cross-location translation framework in which a deterministic Direct network first estimates the target signal and a guide-anchored residual diffusion model refines only the remaining target discrepancy. The evaluation comprises five subject-disjoint activity-location composition-held-out folds, capacity-matched deterministic and flow-based residual baselines, source-only activity conditioning, computational ablation, and true zero-shot evaluation on two independent datasets. DiffusionRef-IMU with 16-step deterministic DDIM achieved a mean nRMSE of 0.087538 +/- 0.004922, compared with 0.090262 for Direct, 0.088847 for deterministic residual refinement, and 0.088417 for eight-step CompFlow, and outperformed each baseline in all five folds. Four-step DDIM retained a mean nRMSE of 0.087820 at 47.62 ms per window, providing a favourable quality-computation trade-off. Replacing ground-truth activity conditioning with a source-only activity predictor (macro-F1 0.8473) increased 16-step DiffusionRef nRMSE only to 0.088190. True zero-shot evaluation produced mixed results: DiffusionRef was marginally favourable on Grouvel2025 but remained worse than Direct on Saldaña2025. Overall, guide-anchored residual diffusion improves controlled cross-location reconstruction while revealing clear limits in zero-shot external transfer.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2604.17139

The Consensus Trap: Collaborative Interleaved Generation to Defend against a Corrupted Majority

Jiayuan Liu, Shiyi Du, Weihua Du, Xinhao Cheng, Mingyu Guo, Vincent Conitzer

原文摘要

Majority voting offers a simple way to combine answers from multiple large language model (LLM) agents. When a majority follows instructions favoring the same incorrect answer, however, agreement reinforces the error. We call this failure the "consensus trap": a minority may identify the error, yet its reasoning cannot change a vote over completed answers. We introduce chunk-interleaved Round-Robin (RR) collaboration, in which agents take turns extending one unfinished reasoning trajectory. A minority of agents with correct reasoning can insert corrections that subsequent speakers read before the answer is complete. RRMaj combines several such trajectories by a final vote. We characterize a majority barrier for outcome-level voting over final answer labels and derive a local sufficient condition for expected progress when the minority's cumulative corrections outweigh opposing corrupted updates. Across model families and verifiable reasoning benchmarks, RRMaj substantially improves accuracy under majority corruption while retaining comparable average performance under minority corruption in the main advisory-instruction setting. RR also strengthens full-response debate. Scoring the shared reasoning between turns shows that, on average, continuations from uncorrupted agents increase support for the correct answer over the incorrect target. Parallel local serving adds modest latency in the measured deployment. These findings show how language shared during reasoning, and the timing of that exchange, can help a minority correct a group's answer.

cs.CL · cs.AI · cs.MA

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:1205.0903

A note on a problem in communication complexity

Henning Wunderlich

原文摘要

We study bounded-error randomization and majority operators in structural communication complexity. We give a communication-complexity formulation of Tarui's theorem and identify the bounded-error closure of weakly unbounded-error communication with a distributional perturbation of logarithmic margin complexity. Retaining the Boolean guarantee in Tarui's probabilistic polynomials yields the stronger inclusion $\mathsf{PH}^{\mathrm{cc}} \subseteq \mathsf{ZP}\mathbin{\cdot} \mathsf{PP}^{\mathrm{cc}} \subseteq \mathsf{BP}\mathbin{\cdot} \mathsf{PP}^{\mathrm{cc}}$. We also define a majority operator satisfying $\mathsf{PP}\mathbin{\cdot} \mathsf{P}^{\mathrm{cc}} = \mathsf{PP}^{\mathrm{cc}}$ and use Sasaki's majority formula to prove that $\mathord{\oplus}\mathsf{P}^{\mathrm{cc}} \subseteq \mathsf{PP}\mathbin{\cdot} \mathsf{PP}^{\mathrm{cc}}$. For the inner-product function modulo two, both the outer index and each inner protocol have logarithmic cost. Together with Forster's lower bound, this gives an explicit witness to the non-idempotence of the majority operator and to $\mathsf{PP}\mathbin{\cdot} \mathsf{PP}^{\mathrm{cc}} \not\subseteq \mathsf{UPP}^{\mathrm{cc}}$. Both the bounded-error closure and the two-level majority class are contained in $\mathsf{PSPACE}^{\mathrm{cc}}$.

cs.CC

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.33877

Semantic Concept Steering Breaks the Explanation Drift Loop in Continual Learning

Yehonatan Elisha, Oren Barkan, Noam Koenigstein

原文摘要

Catastrophic forgetting in continual learning (CL) manifests not only as accuracy degradation but also as \emph{explanation drift}: models maintain predictive accuracy while silently shifting attention from diagnostic features to spurious correlations. This work identifies and resolves a previously overlooked structural flaw in explanation-aware CL: pixel-level saliency regularization derives supervision targets from the model's own frozen saliency maps, creating a \emph{self-referential drift loop} that actively reinforces spurious attention across tasks. We introduce \textbf{C$^3$L} (Concept-Consistent Continual Learning), which breaks this loop by anchoring regularization to class-consistent semantic concepts (e.g., ``curved beak'', ``striped pattern'') discovered offline, independent of model state, providing (i) target independence from model drift, and (ii) cross-instance semantic consistency across every instance of a class. We identify that our concept-based objective, combined with approximate LRP relevance conservation, induces a \emph{crowding-out effect}: since concept and background regions compete for the relevance of the same prediction, enforcing high relevance within concept regions implicitly suppresses spurious background correlations without explicit penalization. \method does not require human-annotated concept sets, as it employs a fully automated concept discovery and spatial grounding. Extensive evaluation across five benchmarks and over 20 baselines demonstrates state-of-the-art performance in accuracy, forgetting, explanation quality, spurious correlation mitigation, and out-of-distribution robustness.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09318

Recursively Self-Improving Electromagnetic Spectrum Agents: Architecture and Operation

Baolin Chong, Dongfang Xu, Qiaosi Wang, Fuhui Zhou, Qihui Wu

原文摘要

Electromagnetic spectrum agents (ESAs) must contend with changing environments and task requirements during sustained autonomous operation, yet their existing capabilities may not remain suitable for new operating conditions. This article proposes recursively self-improving electromagnetic spectrum agents (RSI-ESAs) that use operational feedback to update their capabilities and build on prior improvements. We present their architecture and editable components, and describe a self-improvement process encompassing problem diagnosis, candidate generation, validation, and inheritance. By retaining effective task versions and using modification records and diagnostic knowledge to guide subsequent updates, RSI-ESAs can accumulate task capabilities and improvement experience across successive rounds. A shared-spectrum scheduling case study illustrates improved task performance and capability accumulation over multiple improvement rounds under the tested conditions. Finally, we discuss open challenges for broader tasks and long-term operation, outlining research directions for the continued capability development of ESAs.

cs.MA

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09317

Stable probabilistic emulation of partially observed dynamical systems

Mario Massy, Sara Shamekh, Fabrizio Falasca, Matthieu Blanke

原文摘要

Predicting the long-term behavior of partially observed dynamical systems such as the coupled atmosphere-ocean system is central to societal decision-making. Motivated by the success of deep learning emulators in weather forecasting, data-driven climate models represent a promising approach. However, climate emulation poses distinct challenges: models must account for unresolved fast variables introduced by unavoidable coarse-graining, reproduce long-term statistics, and remain stable over arbitrarily long rollouts and under external perturbations. These requirements motivate stochastic data-driven emulators, but modern generative models lack stability guarantees. We introduce forced-dissipative stochastic interpolants, a generative framework for learning stochastic reduced dynamics. Its architecture uses structural constraints inspired by turbulent dynamics, which provide stability by design while retaining the flexibility of stochastic interpolants to represent complex probabilistic dynamics. Building on the theory of Markov-chain stability, we establish a Foster--Lyapunov drift condition and prove geometric ergodicity for our emulator, together with stability under steady external forcing. We evaluate the framework on the two-timescale Lorenz-96 system, a prototype of partially observed multiscale turbulence. The resulting emulator remains stable over long integrations and under finite-amplitude forcing, with accurate prediction of forced responses of both the mean and variance.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09316

Keyline Detection with YOLO: Runway Perception for Vision-Based Aircraft Landing

Leander Fritsch, David Rügamer, Alexej Klushyn

原文摘要

Modern object detectors can predict keypoints as per-object geometric features, but keypoints are inherently vulnerable to occlusion and truncation. Object edges can provide complementary information, often remaining observable even when partially occluded or when their defining endpoints leave the field of view. We therefore introduce object keyline detection, the task of predicting a fixed, semantically ordered set of infinite lines per object instance. Unlike the outputs of general line-segment detectors, each keyline is associated with an object instance and identified by a semantic index, providing a line-level analog of keypoints. To solve this task, we extend YOLO with keyline prediction parameterized in continuous Hesse-normal form $[ρ,\cosθ,\sinθ]$, and introduce Object Keyline Similarity (OKLS), which serves as an evaluation metric, target assignment signal, and regression loss. We instantiate this formulation on runway perception in vision-based landing, a motivating example containing semantic linear features that remain partially observable throughout the approach. Experiments show that our representation achieves the highest detection accuracy among four evaluated line parameterizations, and keylines integrate directly into standard YOLO heads as additional output channels without degrading keypoint performance or adding meaningful latency.

cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2602.14885

Embedding stochastic dynamics in latent manifolds of recurrent neural networks

Ramón Nartallo-Kaluarachchi, Renaud Lambiotte, Alain Goriely

原文摘要

Recurrent neural networks (RNNs) provide a theoretical framework for understanding computation in biological neural circuits through their ability to support rich time-dependent behaviour. Whilst high-dimensional by construction, they can be used to encode low-dimensional dynamical systems, for example by restricting to low-rank connectivity, hence mirroring the emergence of latent neural manifolds found in empirical neural data. Here, we exploit a relationship between RNNs and a class of universal approximators to train networks that can represent arbitrary, nonlinear stochastic differential equations (SDEs), with given drift and diffusion coefficients, within a low-dimensional latent subspace. We show that RNNs can faithfully embed the drift and diffusion of a given SDE, including nonlinear and nonequilibrium dynamics such as chaotic attractors. We apply this approach by constructing RNN realisations of stochastic systems that transiently explore various attractors through both input-driven switching and autonomous transitions driven by nonequilibrium currents, which we interpret as models of associative and sequential (episodic) memory. To elucidate how these dynamics are encoded in the network, we introduce decompositions of the RNN based on its asymmetric connectivity and its time-irreversibility. Our results show that neural populations can implement a broad class of stochastic dynamics within low-dimensional manifolds, unifying ideas from the dynamics of neural computation, the theory of neural manifolds, associative memory, and nonequilibrium statistical mechanics.

cond-mat.dis-nn · cond-mat.stat-mech · cs.LG · q-bio.NC

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2607.24852

Hold-Out Self-Validation Cannot Certify the Accuracy of Photogrammetric Reconstructions

Behnam Asadi

原文摘要

Photogrammetric reconstructions used for inspection are often delivered without an independent accuracy check. Hold-out self-validation is a possible substitute. Images are withheld, their orientations are re-estimated from the rest of the block, and agreement is taken as evidence of quality. We test it on ten GNSS-referenced aerial captures from eight sites, on ETH3D and EuRoC, and on 30 IMC 2025 scenes. The standard form re-localises each withheld image only against points seen by at least two retained images. Its score saturates. It stayed at 1.00 in all 40 thinned reconstructions while camera-position error varied by a factor of up to 5.9 within one capture. It also scored 1.00 on a corrupted model with only one of 226 cameras within 5 m. Reprojection error ranked some of the worst models best. A few lost cameras can raise the least-squares RMSE to the size of the block, so we also evaluate with a robust alignment. Taking the correspondences from the verified matches lowered the score below 0.95 in 14 of 18 corrupted runs, against 4 for the standard form and 10 when its failed resections count as errors, with all seven nominal runs at 1.00. Mapping the block again without the withheld images also responds, but mapping the same corrupted data twice without withholding gave a similar disagreement, so the signal is mainly that a corrupted reconstruction is not reproducible. Neither test can see an error that every reconstruction repeats, and neither can replace check points.

cs.CV · cs.RO

查看论文
arXiv ↗PDF ↗