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

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

arXiv 最近完成扫描：2026-11-08 13:28 · 全站最近入库：2026-11-08 06:02各信源进度 →

统计口径与数据新鲜度

统计 2026-11-07 13:41 — 2026-11-08 13:41（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：2026-11-08 13:28最近发现新增版本：2026-11-08 06:02库内论文最新更新时间：2026-11-06 02:59

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。

17,100 已收录论文14,661 AI 相关4,057 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 17,100 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-08 13:25（北京时间）
版本更新 · v2arXiv:2506.03234

The Alignment Illusion: How BadReward Corrupts Text-to-Image Generation

Kaiwen Duan, Hongwei Yao, Yufei Chen, Ziyun Li, Tong Qiao, Zhan Qin, Cong Wang

原文摘要

Modern text-to-image generative models increasingly rely on Reinforcement Learning from Human Feedback (RLHF) to align generated content with human preferences and safety standards. This post-training stage is critical for optimizing models toward high-fidelity and value-aligned behaviors using human preference data. However, while previous works have extensively investigated poisoning attacks against the supervised fine-tuning (SFT) stage of text-to-image training, the security vulnerabilities introduced by the RLHF process remain unexplored. In this paper, we present the first exploration into the feasibility of poisoning the RLHF process as a novel attack surface for compromising text-to-image models. We introduce BadReward, an indirect clean-label poisoning attack that manipulates reward signals to embed malicious behaviors. Our approach leverages a feature collision mechanism to produce poison preference image pairs that are visually benign, but aligned with a malicious target concept in the reward feature space. Our experiments demonstrate that this manipulation effectively corrupts the reward signal. Online RLHF further amplifies this corruption through its reward-model feedback loop, whereas offline preference optimization learns directly from the poisoned preferences. Finally, we demonstrate the real-world effectiveness of BadReward on modern text-to-image architectures, showing high attack success rates and visual stealthiness across different RLHF algorithms. Our code is available at https://github.com/ZJUICSR/BadReward.

cs.LG · cs.AI · cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03133

Meet Students Where They Are: Student-Prefix Teacher Adaptation for On-Policy Distillation

Xue Zhang, Songming Zhang, Yunlong Liang, Fandong Meng, Yufeng Chen, Jinan Xu

原文摘要

On-policy distillation (OPD) mitigates the student's exposure bias by optimizing on student-generated trajectories. However, student-generated prefixes may contain reasoning errors or follow patterns unfamiliar to the teacher, undermining the reliability of its token-level supervision. This degradation becomes more pronounced as prefixes grow longer. To address this teacher-side distribution mismatch, we propose Student-Prefix Teacher Adaptation (SPTA), which adapts the teacher to the states on which it is queried during distillation. Specifically, before OPD starts, we collect student rollout prefixes with random lengths and continually train the teacher to continue from the prefixes using reinforcement learning with verifiable rewards (RLVR). By treating each prefix as context and rewarding correct final answers, the teacher learns to recover from student-induced states rather than imitate the student's reasoning. Then, the adapted teacher is used for standard OPD, with no changes to the distillation objective. Experiments on mathematical reasoning benchmarks show that our SPTA improves average student performance over standard OPD and OPD with task-level RLVR-adapted teachers. These results highlight the importance of adapting teachers not only to solve problems independently, but also to provide reliable guidance on the states their students actually visit.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2604.20729

On the regularity index of the minimum distance function in projective nested Cartesian codes

Cicero Carvalho, Maria Vaz Pinto, Rafael H. Villarreal

原文摘要

Let $X$ be a projective nested product of fields and let $δ_X(d)$ be the minimum distance in degree $d\geq 1$ of the projective nested Cartesian code $C_X(d)$. The regularity index ${\rm reg}(δ_X)$ of the minimum distance function $δ_X$ is the minimum integer $d_0\geq 0$ such that $δ_X(d)=1$ for $d\geq d_0$. We give a formula for ${\rm reg}(δ_X)$ by determining an indicator function of least degree for each point of $X$ and using the fact that ${\rm reg}(δ_X)$ is the ${\rm v}$-number of the vanishing ideal $I_X$ of $X$. Then we give an arithmetical criterion that characterizes when $X$ is Cayley--Bacharach.

math.AC · cs.IT · math.AG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03132

Behavioral Affinity Transduction for Zero-Shot Vision-Language Classification

Junxi Huang, Wenshuo Chen, Xiucheng Wang, Meitong Shen, Shaofeng Liang, Yuehan Kou, Yutao Yue, Nan Cheng

原文摘要

An image that is hard to classify on its own may have a helpful neighbor in the same unlabeled batch. Zero-shot vision-language models (VLMs) score images independently and leave this evidence unused. We study small unlabeled query batches with no labeled examples. Visual similarity can reflect pose or background, so it may connect images from different classes. Each image's scores over the class vocabulary offer another way to choose neighbors. We introduce Behavioral Affinity Transduction (BAT), which builds a sparse graph from these class-response profiles. The graph supplies neighbor evidence during prediction. For EM-Dirichlet, it also screens candidate classes and initializes assignments. BAT plugs into direct and iterative readouts of a frozen VLM. Across diverse datasets, BAT improves the mean accuracy of every tested readout. In the main RN50 comparison, soft and hard EM-Dirichlet gain $4.22$ and $2.72$ percentage points.

cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.12074

Non-frontal face recognition using GANs and memristor-based classifiers

Semih Vazgecen, Cristian Sestito, Spyros Stathopoulos, Themis Prodromakis

原文摘要

Face recognition systems have advanced significantly through deep learning techniques, delivering high performance and robustness in complex scenarios. However, these approaches incur substantial computational overhead, limiting their in situ applicability in resource-constrained platforms such as drones, where they can address challenges including non-frontal facial imagery. Memristor-based neuromorphic systems have emerged as a compelling approach for edge AI applications, combining biologically inspired processing with efficient and scalable computation. In this work, we propose a facial recognition framework that addresses non-frontal pose variations by integrating lightweight generative adversarial network (GAN)-based pose frontalisation with memristor-based neuromorphic recognition. The experimental results on two datasets demonstrate the effectiveness of combining adversarial learning with memristive technology, achieving up to 96% identification accuracy. The proposed approach alleviates the computational bottlenecks of conventional AI and offers a scalable, efficient solution for face recognition in dynamic real-world environments.

cs.CV · cs.AI · eess.IV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03130

An Information Hierarchy of Privacy Filters and Odometers

Sophie Taylor, Praneeth Vippathalla, Justin P. Coon

原文摘要

In sequential data analysis, privacy loss accumulates over time, and privacy filters and odometers are used to control and track this cumulative cost. We study how the information used by differential privacy filters and odometers fundamentally limits what they can achieve. We formalise three levels of information: spec-level information, fixed prior to the interaction; mechanism-level information, which includes the mechanisms instantiated during the interaction; and realisation-level information, which further includes the realised interaction. We classify privacy filters and odometers by the least informative level sufficient to determine their behaviour. Our main results establish strict and arbitrarily large separations in achievable performance across this information hierarchy. For privacy filters, moving up the hierarchy can yield an unbounded utility advantage, quantified by a clipped measure of the number of allowed releases. For privacy odometers, it can yield an arbitrarily large improvement in precision, quantified by the expected peak excess of the reported bound over the realised privacy loss. Further, although the information classes are distinct, we show that moving to a richer class need incur essentially no loss in performance. Finally, we show that randomisation provides a resource orthogonal to information: at every information level, it can yield an arbitrarily large utility improvement over deterministic filters. Thus, information access and randomisation can be viewed as fundamental resources governing privacy procedures in adaptive interactions.

cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03128

Belief Propagation-Informed Tesseract Decoding of Quantum LDPC Codes

Sana Javed, Francisco Garcia-Herrero, Bane Vasić, Mark F. Flanagan

原文摘要

In quantum error correction, the decoder must estimate the errors from the measurement data accurately and fast enough to avoid a backlog. Search-based decoders such as Tesseract achieve high accuracy by finding a minimum-weight set of error events with an A* search, but their search effort grows rapidly with the physical error rate and the code size. To reduce this search effort while maintaining a logical error rate close to that of Tesseract, we propose a belief propagation (BP)-guided modification of the Tesseract decoder for quantum low-density parity-check (LDPC) codes. Damped sum-product BP is applied to the triggered detectors before the Tesseract search. If the detectors triggered by the BP hard decision match the triggered detectors, this hard decision is accepted directly, and the Tesseract search is skipped. Otherwise, the BP marginal probability estimates, weighted by a trust coefficient, modify the search priority, while the path cost keeps the channel-prior weights so that the decoding objective is unchanged. The framework also supports BP-based initialization, in which selected BP decisions form the starting point of a residual search whose signed-cost objective allows incorrect initial decisions to be removed. A node-exploration limit bounds this residual search; if the limit is reached, the decoder falls back to the BP-guided search on the original triggered detector set. The method is evaluated on the [[72,12,6]], [[90,8,10]], and [[144,12,12]] bivariate bicycle codes under circuit-level noise. With BP-based initialization and a global trust coefficient, the number of Tesseract search nodes expanded is reduced by up to more than two orders of magnitude at low physical error rates, while the logical error rate remains close to that of Tesseract.

quant-ph · cs.IT

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2604.19009

Guiding Distribution Matching Distillation with Gradient-Based Reinforcement Learning

Linwei Dong, Ruoyu Guo, Ge Bai, Quan Zheng, Yawei Luo, Changqing Zou

原文摘要

Diffusion distillation, exemplified by Distribution Matching Distillation (DMD), has shown great promise in few-step generation but often sacrifices quality for sampling speed. While integrating Reinforcement Learning (RL) into distillation offers potential, a naive fusion of these two objectives relies on suboptimal raw sample evaluation. This sample-based scoring creates inherent conflicts with the distillation trajectory and produces unreliable rewards due to the noisy nature of early-stage generation. To overcome these limitations, we propose GDMD, a novel framework that redefines the reward mechanism by prioritizing distillation gradients over raw pixel outputs as the primary signal for optimization. By reinterpreting the DMD gradients as implicit target tensors, our framework enables existing reward models to directly evaluate the quality of distillation updates. This gradient-level guidance functions as an adaptive weighting that synchronizes the RL policy with the distillation objective, effectively neutralizing optimization divergence. Empirical results show that GDMD sets a new SOTA for few-step generation. Specifically, our 4-step models outperform the quality of their multi-step teacher and substantially exceed previous DMDR results in GenEval and human-preference metrics, exhibiting strong scalability potential.

cs.LG · cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03126

CUA-Master: Scaling Computer-use Agents through Open-World Recursive Improvements

Xinlei Yu, Gen Li, Ji Zhao, Zhangquan Chen, Congcong Wang, Shuai Dong, Zelin Peng, Yuqi Xu 等 18 位作者

原文摘要

Despite recent advances in CUAs, their development remains constrained by the limited diversity, controllability, and reliability of interaction data. Moreover, conventional optimization pipelines lack iterative feedback, while existing recursive improvement approaches often rely on self-generated experiences, limiting ability expansion and yielding diminishing returns. To address these challenges, we introduce CUA-Master, a nested double-loop recursive improvement framework. At the system level, an outer world-data-model flywheel continuously incorporates open-world corpus and leverages evolving model weaknesses to guide subsequent data construction and optimization iterations. At the data level, an inner flywheel couples open-world seeding and autocatalytic propagation. The former mines authentic workflows as seed tasks from instructional videos, and employs visual-semantic dual-path trajectory reconstruction; The latter transforms these seeds into reusable primitives for controllable synthesis of diverse, complex, and verifiable tasks. To enable efficient and reliable trajectory collection, we further introduce a feedback-guided snapshot-rollback retry mechanism in various real environments. Extensive experiments demonstrate that CUA-Master-27B achieves competitive performance across GUI-only and GUI-CLI hybrid benchmarks, scoring 80.2 on OSWorld Verified, 36.9 on OSWorld 2.0, 43.9 on WeaveBench, and 25.0 on Agents' Last Exam. Our framework also substantially improves the quality and efficiency of data synthesis, and exhibits continuous gains across recursive iterations.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2606.24471

Discrepancy for Random Linear Codes

Dean Doron, Tal Leonov, Jonathan Mosheiff, Henrique Navas, Nicolas Resch, João Ribeiro

原文摘要

We show that random linear codes (RLCs) possess nearly optimal discrepancy-type properties in a broad range of settings. Our main results are two general discrepancy theorems: one controls all translates of a fixed test, and the other controls large families of Fourier-pseudorandom tests. Two motivating examples follow: First, RLCs behave essentially like unstructured random codes for list-decoding from errors above capacity. More precisely, an RLC $C\subseteq \mathbb{F}_q^n$ of rate $1 - \frac{1}{n}\log_q|B_ρ| + \varepsilon$, where $|B_ρ|$ is the volume of a radius-$ρ$ Hamming ball in $\mathbb{F}_q^n$, satisfies $|C \cap B| = (1\pm o(1)) \frac{|C|\cdot |B|}{q^n}$ simultaneously for all radius-$ρ$ Hamming balls $B$ with high probability. This significantly strengthens the best prior results, which follow from work by Blinovsky (1987, 1990) and Levitin (1988) on the covering properties of RLCs. Second, over prime fields, RLCs behave essentially like unstructured random codes for zero-error list-recovery, and list-recovery from erasures, above capacity. More precisely, for a prime $q>2$ and input list size $2\leq \ell\leq q-1$, an RLC $C\subseteq \mathbb{F}_q^n$ of rate $1-\log_q \ell+\varepsilon$ will satisfy $|C \cap S| = (1\pm o(1)) \frac{|C|\cdot \ell^n}{q^n}$ simultaneously for all combinatorial rectangles $S=S_1\times S_2\times\cdots\times S_n$, where $|S_i|=\ell$ for all $i$, with high probability. An analogous result also holds when we can bound $|S_i|$ only for some of the $i$'s. We use this to show the abundance of locally leakage-resilient $n$-party linear ramp secret sharing schemes with any linear reconstruction threshold and sublinear threshold gap $O(n/\log n)$ over fields of polynomial size $q=Θ(n^γ)$ for a constant $γ\in(0,1/5)$. Prior work was stuck at reconstruction thresholds above $n/2$ for both threshold and ramp schemes.

cs.IT · cs.CC · cs.CR · math.CO

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.28501

Improved Approximation for Unsplittable Capacitated Vehicle Routing Using Quota Trees

Zhenyu Wang

原文摘要

In the Unsplittable Capacitated Vehicle Routing Problem (UCVRP), we are given a set $V$ of $n$ clients with normalized demands $d_v \in (0,1]$, a depot $r$, a unit vehicle capacity, and symmetric metric travel costs $c$ on $V \cup \{r\}$. We seek a collection of routes of minimum total cost that start and end at $r$, serve each client's entire demand on a single route, and respect the unit capacity constraint $\sum_{v \in V(R)} d_v \le 1$ on every route $R$, where $V(R)$ denotes the clients served by $R$. UCVRP generalizes the metric traveling salesman problem (TSP) and is therefore NP-hard. Capacity constraints add a packing difficulty: when all clients are colocated away from the depot, minimizing the route cost is equivalent to minimizing the number of bins in bin packing. Consequently, UCVRP admits no polynomial-time approximation ratio below $3/2$ unless $\textbf{P=NP}$ (Grandoni, Mathieu, and Zhou, 2023). The classical $3.5$-approximation has been improved; Zhao and Xiao recently obtained a $(2+\ln 2+\varepsilon)$-approximation. In this paper, we give a deterministic polynomial algorithm with approximation ratio $5/2$ for general capacity and arbitrary positive demands. Our algorithm uses local search to select routes through one client with large demand and clients with small demands. It constructs candidate routes using endpoint quota trees after rounding demands and creating vertex copies. We replace selected routes using sets of two to four clients and a squared-weight local search, keeping their client sets disjoint. The remaining clients are served by partitioning a tour constructed using weighted graphic matroid parity. We also give a deterministic polynomial $5/2$-approximation algorithm for the undirected capacitated arc routing problem (CARP) with arbitrary positive edge demands.

cs.DS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03125

SurgPace: Complexity-Aware Progressive Distillation for Surgical Multimodal Large Language Models

Zhicheng He, Xiaohan Xing, Hongqiu Wang, Kun Yuan, Guankun Wang, Yanheng Li, Haoxuan Che, Nassir Navab 等 10 位作者

原文摘要

Multimodal large language models (MLLMs) have shown strong potential in robotic surgery for question answering, scene understanding, and grounded assistance, but their high computational cost limits real clinical deployment. Knowledge distillation provides a practical way to build compact surgical MLLMs, yet direct distillation is not well suited to surgical data because surgical samples vary greatly in difficulty, visual ambiguity, and required clinical context, making uniform teacher-student transfer unstable. In addition, the role of supervision changes during training: early stages require stronger teacher guidance for multimodal alignment, while later stages rely more on task supervision for effective optimization. To address these challenges, we propose SurgPace, a complexity-aware continuous curriculum framework for progressive distillation of surgical vision-language models. SurgPace estimates sample complexity using predictive confidence and Lorentz geometric consistency, and uses the fused score to guide a continuous easy-to-hard training process. We further introduce hierarchical representation alignment, which aligns the student with the teacher at the feature, logit, and probability levels, while gradually shifting the training focus from teacher imitation to task supervision. This design separates sample-level curriculum from objective-level scheduling, leading to more stable and effective distillation. We evaluate SurgPace on Surg-396K, a benchmark built from four public surgical datasets and covering five dialogue paradigms. SurgPace consistently outperforms strong MLLM distillation baselines, demonstrating its effectiveness for building compact, memory-efficient surgical MLLMs with faster decoding under controlled benchmarks. Code is available at https://github.com/hezhicheng2002/SurgPace.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03121

ComplianceAtlas: A Taxonomy-Driven Agentic AI Framework for Automated AI Act Gap Analysis and Remediation

Ilias Driouich, Hongliu Cao, Eoin Thomas, Yannick Devaux, Agnes Brodelle

原文摘要

Assessing regulatory compliance for high-risk AI systems requires systematic identification of technical gaps such as missing controls, in- adequate processes, or shallow evaluation of practices. Compliance officers typically admin- ister structured questionnaires collecting evi- dence of current practices, where organizations respond with descriptions of their existing pro- cesses, technical capabilities, and architectural choices. However, translating these compli- ance questionnaire responses into typed, ac- tionable gap diagnoses remains unsystematic. In this paper, we present a literature-grounded taxonomy of technical compliance gaps cover- ing areas such as accuracy, robustness, logging, and human oversight, and embed it in GapTo- Compliance, a two-stage agentic framework coordinating two specialized agents: a Gap Extraction Agent that performs hierarchical rea- soning from compliance areas to specific gap types, and a Remediation Planning Agent that ranks remediation actions from a remediation catalog via multi-signal scoring combining se- mantic similarity, regulatory alignment, and effort estimation. We evaluate this framework against four automatic compliance gap detec- tion baselines across 9 AI systems and 6 AI families. Our approach achieves a gap-type F1 of 0.831 (vs. 0.533 for the strongest baseline) and remediation P@5 of 0.872 (vs. 0.635 for retrieval methods), demonstrating promising potential for automated compliance workflows. The structured agent coordination reveals a path toward human-in-the-loop compliance systems where automated gap extraction reduces expert burden while maintaining auditability, and the taxonomy-driven design enables extension to emerging regulatory frameworks beyond

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.42138

Approximate Polynomial Satisfiability and Integral Closure Membership in the Counting Hierarchy

Sanyam Agarwal, Archit Hari, Anurag Pandey, Himanshu Shukla

原文摘要

Andrews, Garg, and Schost (FOCS~2026) recently, in a breakthrough result, placed Hilbert's Nullstellensatz (HN) in the Counting Hierarchy ($\CH$) using constant-depth circuits for multivariate resultants. While HN deals with exact polynomial satisfiability, Guo, Saxena, and Sinhababu (CCC~2018; Theory of Computing~2019) introduced approximate polynomial satisfiability (APS) as a natural approximative analogue of polynomial system satisfiability. They proved its $\mathsf{NP}$-hardness and a $\PSPACE$ upper bound, and used APS to construct hitting sets for $\overline{\mathsf{VP}}$ in polynomial space, substantially narrowing the GCT chasm in geometric complexity theory. We place APS in the Counting Hierarchy for rational polynomials given by arithmetic circuits. Our proof combines Noether normalization with a geometric restriction on how points in the image of a polynomial map approach the origin, reducing APS to a test on multivariate resultants which allows solvability in $\CH$. It complements the independent and concurrent proofs of Balaji, Shirmohammadi, Tavenas, and Worrell, and Andrews, Garg, Oliveira, and Schost. Moreover, we also place integral closure membership testing in the Counting Hierarchy. Here, the question is whether a given polynomial $f$ is integral over $\C[f_1,\ldots,f_m]$. Our approach uses a global annihilator criterion and an effective degree bound to reduce non-integrality to APS in polynomial time. In fact, we show that the reduction holds both ways, and Integral Closure Membership Testing is equivalent to APS under polynomial-time reductions.

cs.CC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03120

KAUSTics-Net: Diffusion-Based Through-Water UAV Restoration with a Semi-Synthetic Dataset

Dario Scilla, Omar A. Lopez, Kasper Johansen, Brian O. Nieuwenhuis, Matthew F. McCabe

原文摘要

Accurate mapping of shallow-water environments using unmanned aerial vehicle (UAV) imagery is challenged by water-induced distortions, including refraction, caustics, and wavelength-dependent attenuation. The scarcity of paired distorted and reference imagery further limits the development of supervised restoration methods. To address this challenge, we introduce KAUSTics, a semi-synthetic dataset combining 1,280 reference patches derived from 32 high-resolution UAV images with 12,800 degraded observations generated through Blender-based water-surface rendering and empirical color and blur augmentations. We further propose KAUSTics-Net, a pixel-space conditional diffusion framework integrating a multi-scale conditional encoder, feature-wise linear modulation (FiLM), cross-attention, and a composite reconstruction objective to restore shallow-water imagery. We evaluated the framework against eight image-restoration methods, complemented by architectural ablations and cross-dataset experiments. On KAUSTics, KAUSTics-Net achieved a mean peak signal-to-noise ratio (PSNR) of 31.715 +/- 0.439 dB, structural similarity index (SSIM) of 0.849 +/- 0.002, and learned perceptual image patch similarity (LPIPS) of 0.145 +/- 0.007 across three training seeds. On the independent Sea-Undistort dataset, sampler-aware fine-tuning improved restoration fidelity, reaching 19.469 +/- 0.332 dB PSNR and 0.554 +/- 0.005 SSIM under combined water-surface distortions, although perceptual trade-offs remained. Additional experiments on real UAV imagery examined feature preservation and segmentation behavior. These findings demonstrate the potential of semi-synthetic training and conditional diffusion for shallow-water image restoration while highlighting the importance of evaluating both reconstruction fidelity and downstream remote-sensing utility.

cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.29074

Nested Convex-Body Chasing for Online Optimization with Evolving Feasible Sets

Dhruv Sarkar, Aprameyo Chakrabartty

原文摘要

We study online optimization with nested shrinking feasible regions in two settings: convex optimization with nested evolving feasible sets (CONES) and adversarial constrained online convex optimization (COCO). Our algorithms separate loss control from geometric movement: constrained minimizers and cumulative-loss tests preserve regret guarantees, while a deterministic resettable nested convex-body chaser limits movement. For CONES with a $G$-Lipschitz, $μ$-strongly convex objective on a diameter-$D$ domain, we chase intersections of the current feasible set with adaptive objective sublevel sets. Using the Euclidean chasing ratio $O(\sqrt{d\log(1+d)})$, we obtain nonpositive regret at every prefix and movement $O(\sqrt{d\log(1+d)\,GD\log(eT)/μ})$. The bound adapts to the increase in the constrained optimum value. In dimension two, with all other parameters fixed, every randomized algorithm with terminal expected regret $O(T^β)$, $β<1$, suffers $Ω(\sqrt{\log T})$ expected movement on some deterministic nested sequence, proving optimal horizon dependence. Under linear growth away from the constrained minimizer set, Steiner-point tracking yields movement independent of $T$. For general convex COCO, one-step-delayed chasing with regularized-leader resets gives regret $O(G_fD\sqrt{d\log(1+d)T})$ and cumulative constraint violation $O(G_gD\sqrt{d\log(1+d)T})$. For strongly convex losses, both are $O(d\log(1+d)\log(eT))$ when other parameters are fixed. These reductions replace the $O(d^{d/2})$ projection-path factor in prior analyses by the polynomial dimension dependence of Euclidean nested convex-body chasing.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2604.16923

Alignment Imprint: Zero-Shot AI-Generated Text Detection via Provable Preference Discrepancy

Junxi Wu, Kailin Huang, Dongjian Hu, Bin Chen, Hao Wu, Shu-Tao Xia, Changliang Zou

原文摘要

Detecting AI-generated text is an important but challenging problem. Existing likelihood-based detection methods are often sensitive to content complexity and may exhibit unstable performance. In this paper, our key insight is that modern Large Language Models (LLMs) undergo alignment (including fine-tuning and preference tuning), leaving a measurable distributional imprint. We theoretically derive this imprint by abstracting the alignment process as a sequence of constrained optimization steps, showing that the log-likelihood ratio can naturally decompose into implicit instructional biases and preference rewards. We refer to this quantity as the Alignment Imprint. Furthermore, to mitigate the instability in high-entropy regions, we introduce Log-likelihood Alignment Preference Discrepancy (LAPD), a standardized information-weighted statistic based on alignment imprint. We provide statistical guarantee that alignment-based statistics dominate Fast-DetectGPT in performance. We also theoretically show that LAPD strictly improves the unweighted alignment scores when the aligned and base models are close in distribution. Extensive experiments show that LAPD achieves an improvement 45.82% relative to the strongest existing baselines, yielding large and consistent gains across all settings.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v7arXiv:2412.04970

CMSO-transductions for tree-like graph decompositions

Rutger Campbell, Bruno Guillon, Mamadou Moustapha Kanté, Eun Jung Kim, Noleen Köhler

原文摘要

We give $\operatorname{CMSO}$-transductions that, given a graph $G$, output its modular decomposition, its split decomposition and its bi-join decomposition. This improves results by Courcelle [Logical Methods in Computer Science, 2006] who gave such transductions using order-invariant $\operatorname{MSO}$, a strictly more expressive logic than $\operatorname{CMSO}$. Our methods more generally yield $\operatorname{C}_2 \operatorname{MSO}$-transductions that output the canonical decompositions of weakly-partitive set systems and weakly-bipartitive systems of bipartitions.

cs.LO · cs.CC · cs.DM · cs.FL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03117

Inner optimizers in active-learning Kriging reliability analysis: a pre-specified comparison and the unconservative estimates left by unlearned failure modes

Wei-Chang Yeh

原文摘要

Active-learning Kriging often replaces candidate pools by continuous metaheuristic search. A pre-specified comparison sharing surrogate, population and stopping rule isolated this inner optimizer: five budget-matched continuous optimizers, including hybrid swarm optimization with L-BFGS-B polishing, against candidate pools on eleven problems (dimension 2 to 100). The primary hypothesis was not supported: against optimizers sharing its criterion the hybrid won 6, tied 41 and lost 19 of 66 Holm-corrected accuracy comparisons, and against pool methods 0, 23 and 21 of 44. With sequential selection on three series systems, multistart L-BFGS-B left a failure mode unlearned in 59 of 60 runs, more often than swarm, genetic-algorithm or AK-MCS. In all 198 runs (22 post-freeze) with an unlearned mode, the learned modes did not compensate and failure probability was underestimated (exploratory). The error criterion can accept substantial underestimation: it accepted 17 runs (13 pool-selected) at 0.38 to 0.89 of the reference; 181 exhausted the budget. Follow-up analyses on single-design-point problems indicate underestimation when selections concentrate near the design point. Kriging-believer pool selection satisfied the surrogate stopping criterion with fewer calls than the best continuous method in all four informative batch cells (two post-freeze). We recommend default pool selection and coverage checks for any selection.

cs.CE · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03116

From Local Skills to Long-Horizon Web Tasks: Progressive Skill Exploration for LLM Web Agents

Xueyang Feng, Xiaohe Bo, Xu Chen, Quanyu Dai, Feng Wen

原文摘要

Large language models (LLMs) are driving autonomous agents toward real-world environments, with the Web emerging as one of the most representative interaction scenarios. However, training Web agents with naive exploration remains challenging: successful trajectories are rare, and the sparse terminal rewards make it difficult for agents to obtain useful feedback from failed trajectories. Inspired by skill acquisition theory, we propose Skill-Composed Progressive Exploration (SCOPE), a short-to-long learning framework for long-horizon Web agents that decomposes complex task exploration into a gradual expansion process from local skills to complete task-solving capability. Specifically, SCOPE iteratively explores around local skills: Skill Discovery extracts reusable local skills and failure experiences from historical rollouts; Skill Orchestration composes discovered skills into high-level skill chains; and Skill Expansion leverages failure experiences aligned with the current skill chain to guide the agent toward exploring new subskills. Through this process, SCOPE progressively constructs task-completing skill chains across rollouts and internalizes them into the policy parameters through agentic on-policy distillation. Extensive experiments on WebArena, WebVoyager, and WebShop show that SCOPE consistently improves exploration dynamics and achieves 3.9% to 18.6% relative gains over the strongest baselines in each setting.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03115

Faster Exact Algorithms for 3- and 4-Coloring: Deterministic Improvements and Further Randomized Speedups

Ryosuke Yamano, Tetsuo Shibuya

原文摘要

We study exact exponential-time algorithms for 3- and 4-coloring. For 3-coloring, the best previously published running-time bound is $O^*(1.3289^n)$, due to Eppstein (SODA 2001). We improve this bound to $O^*(1.3204^n)$ deterministically and $O^*(1.3159^n)$ using randomization. For 4-coloring, the best previously known running-time bound is $O^*(1.7159^n)$, due to Wu, Gu, Jiang, Shao, and Xu (ESA 2024). We improve this bound to $O^*(1.6975^n)$ deterministically and $O^*(1.6933^n)$ using randomization. Our algorithms for both problems combine constraint satisfaction algorithms with dynamic programming over tree and path decompositions, exploiting their complementary strengths on different graph structures. For 6-coloring, the best published randomized running-time bound is exponentially smaller than the best published deterministic bound. For 3- and 4-coloring, however, no such improvement through randomization was previously known. To the best of our knowledge, our results provide the first randomized running-time bounds for these two problems that exponentially improve upon the best previously published deterministic bounds. This raises the question of whether randomization offers an inherent exponential advantage for 3- and 4-coloring.

cs.DS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03113

Preprocessing Choices in Fine-Tuning Data Matter: Utility and Demographic Sensitivity in LLM-Based Recommendation

Furkan Yasir Goksu, Yaren Yilmaz, Sule Gunduz Oguducu

原文摘要

Large language models (LLMs) have introduced a generative approach to recommendation, producing item lists from textual representations of user histories and item information. However, this generative setting raises challenges concerning output validity, prompt consistency, and sensitivity to demographic descriptors. In this work, we investigate how preprocessing choices in fine-tuning data affect recommendation utility and fairness-related behavior in a generative LLM-based recommender system. We first establish a controlled recommendation pipeline that combines supervised fine-tuning, catalog-constrained decoding, and prompt augmentation. We then study three preprocessing choices: item-title representation, demographic balancing, and auxiliary item-knowledge enrichment. Experiments on MovieLens-1M show that the catalog-constrained pipeline eliminates out-of-catalog items and ensures complete ten-item outputs, while prompt augmentation improves prompt consistency and utility. Original MovieLens titles increase HR@10 from 10.67% to 12.47% and Coverage@10 from 41.36% to 47.88%. Fairness-related effects are attribute-dependent: age balancing and knowledge enrichment reduce selected group-level utility gaps while largely preserving utility, whereas gender balancing reduces utility without improving the evaluated gender-related indicators. Overall, preprocessing choices in fine-tuning data should be evaluated jointly in terms of utility, output behavior, and demographic sensitivity.

cs.IR · cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.11813

Automated Reformulation of Robust Optimization via Memory-Augmented Large Language Models

Jinbiao Chen, Shuang Jin, Guoyun Zhang, Junyu Zhang, Guanyi Wang, Hanzhang Qin

原文摘要

Robust optimization (RO) provides a principled framework for decision-making under uncertainty, but its practical use is often limited by the need to manually reformulate uncertain optimization models into tractable deterministic counterparts. Recent large language models (LLMs) have been shown promising for automating optimization formulation, yet RO reformulation remains challenging because it requires precise multi-step reasoning and mathematically consistent transformations. To facilitate systematic evaluation of LLM-based reformulation, for which no dedicated benchmark currently exists, we develop AutoRO-Bench, a benchmark featuring an automated data generation pipeline for the core RO reformulation task and a curated dataset for the RO application task. To address the reformulation challenge, we propose Automated Reformulation with Experience Memory (AutoREM), a tuning-free memory-augmented framework that autonomously builds a structured textual experience memory by reflecting on past failed trajectories through a tailored offline adaptation procedure. AutoREM requires neither domain-specific expert knowledge nor parameter updates, and the resulting memory readily transfers across different base LLMs. Experimental results show that AutoREM consistently improves the accuracy and efficiency of RO reformulation across in-distribution datasets, out-of-distribution datasets, and diverse base LLMs.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03111

COVE: A Task-Adaptive, Code-Based Ensemble for Proactive Verification of LLM Outputs

Anjaney Shrivastava, Dharmala Hrudai Aditya

原文摘要

Large Language Models (LLMs) are now embedded in enterprise workflows, yet their outputs remain prone to hallucinations: plausible-sounding but factually incorrect or malformed content. Most verification approaches are reactive, detecting errors only after generation and often spending additional expensive LLM calls to do so. We present COVE, a task-adaptive, code-based ensemble that takes the opposite stance: it prevents errors before generation and validates structure and consistency afterwards using six deterministic, pure-Python innovations that add no extra LLM calls. On an internal evaluation spanning 11 text datasets (1,020 examples) and 4 image datasets (200 examples), COVE lifts text accuracy from a 45.2% unverified baseline to 82.1% (an 81.7% relative improvement), improves image accuracy by 69.5%, and reduces token usage by 48 to 96% (simulated cost model), at a measured verification overhead of 0.26 ms per output (excluding the LLM call). This article describes the engineering design of each innovation (with pseudocode and the exact parameters we use), the measured impact, an ablation of every component, and the practical lessons learned from deploying code-based verification in enterprise pipelines.

cs.SE

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.38496

Accelerating Nonlinear MPC by Behavior Cloning with Separated Timescales: Application to Nuclear Reactor Load-Following

Perceval Beja-Battais, Alain Grosset{ê}te, Nicolas Vayatis

原文摘要

Behavior Cloning (BC) of expert controllers for multiscale dynamical systems is challenging: state variables evolving at widely separated timescales induce highly nonlinear couplings that a shared latent space may fail to adequately represent, degrading closed-loop generalization. We propose Structured Representation Learning for Behavior Cloning (SRLBC), an architecture designed to leverage the physical knowledge of multiscale systems in which slow state variables, fast (stiff) state variables, and predictable exogenous inputs evolve at different timescales. On 500 long-horizon roll-outs of a pressurized water reactor (PWR) load-following benchmark -a high-dimensional, stiff industrial system with challenging safety constraints -and across 20 training seeds, separated embeddings systematically outperform shared embedding baselines on cumulative cost, imitation accuracy and constraint satisfaction. The resulting standalone policies approach NMPC closed-loop performance at a computational cost 15 times lower on average. Since standalone policies still leave a few percent of trajectories infeasible, we embed the learned policies in a safety-augmented approximate NMPC framework, in which the optimizer is only called when the predicted input sequence is not certified as safe. This yields controllers still 6.8 times faster than the NMPC controller on average, with a feasibility (i.e., constraint satisfaction) of at least 99.8% for all architectures, SRLBC retaining the lowest closed-loop cost. Under distribution shift, standalone BC policies degrade sharply: SRLBC degrades the least, and once safety-augmented, it retains a feasibility above 98% with the lowest closed-loop cost. Beyond the quantitative results, Jacobian-based sensitivity analysis shows that the resulting controller relies on a sparse and less redundant set of latent features, each attributable to a single group of physical variables, and aligned with the physical timescale decomposition of the system.

cs.AI · cs.LG · math.OC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03109

Multimodal Prompt Learning with Dual-Granularity Contrastive Alignment for Preoperative Prediction of Lung Adenocarcinoma Invasiveness Using 3D CT

Qing Huang, Xiaotong Tan, Lintong Yao, Wufeng Xue, Ruobing Huang, Dong Ni, Hao Wu, Tian Xiao 等 9 位作者

原文摘要

Accurate preoperative prediction of lung adenocarcinoma invasiveness is important for surgical planning and individualized treatment. However, overlapping CT appearances make it challenging to distinguish invasive adenocarcinoma (IAD) from Pre-IAD lesions. We propose a multimodal prompt-learning framework with dual-granularity image-text alignment for 3D CT-based IAD classification. A 3D ResNet-CBAM encoder extracts global and spatial visual features. A frozen RadBERT backbone is adapted with separate learnable prompts for report-level and keyword-level text encoding. These prompts are initialized using semantic representations derived from real-world clinical reports. Global contrastive alignment connects whole-volume CT representations with reports generated from clinician-annotated radiological attributes to provide overall semantic supervision. A multiple instance learning-based local alignment module associates spatial CT features with vascular-focused radiological keywords, to capture local imaging characteristics potentially associated with tumor invasion. The framework jointly optimizes classification and alignment objectives, while inference requires only CT images. Experiments included 1,593 pathologically confirmed nodules. On the test set, the proposed method achieved an accuracy of 87.92%, F1-score of 87.78%, and AUC of 0.9499. Its AUC significantly exceeded that of the image-only Res-CBAM baseline with the same visual backbone (0.9297; DeLong test, p = 0.0078). Ablation analyses supported the complementary contributions of global and local alignment and the benefit of report-based prompt initialization. Qualitative visualizations showed activation within nodules and adjacent vascular structures. These findings suggest that dual-granularity radiological supervision can enhance CT-based invasiveness prediction while preserving a CT-only inference workflow.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2505.10267

HandReader: Advanced Techniques for Efficient Fingerspelling Recognition

Pavel Korotaev, Petr Surovtsev, Alexander Kapitanov, Karina Kvanchiani, Alexander Nagaev

原文摘要

Fingerspelling is a significant component of Sign Language (SL), allowing the interpretation of proper names, characterized by fast hand movements during signing. Although previous works on fingerspelling recognition have focused on processing the temporal dimension of videos, there remains room for improving the accuracy of these approaches. This paper introduces HandReader, a group of three architectures designed to address the fingerspelling recognition task. HandReader$_{RGB}$ employs the novel Temporal Shift-Adaptive Module (TSAM) to process RGB features from videos of varying lengths while preserving important sequential information. HandReader$_{KP}$ is built on the proposed Temporal Pose Encoder (TPE) operated on keypoints as tensors. Such keypoints composition in a batch allows the encoder to pass them through 2D and 3D convolution layers, utilizing temporal and spatial information and accumulating keypoints coordinates. We also introduce HandReader_RGB+KP - architecture with a joint encoder to benefit from RGB and keypoint modalities. Each HandReader model possesses distinct advantages and achieves state-of-the-art results on the ChicagoFSWild and ChicagoFSWild+ datasets. Moreover, the models demonstrate high performance on the first open dataset for Russian fingerspelling, Znaki, presented in this paper. The Znaki dataset and HandReader pre-trained models are publicly available.

cs.CV · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2603.13663

PDE-SSM: A Spectral State Space Approach to Spatial Mixing in Diffusion Transformers

Eshed Gal, Moshe Eliasof, Eldad Haber

原文摘要

The success of vision transformers-especially for generative modeling-is limited by the quadratic cost and weak spatial inductive bias of self-attention. We propose PDE-SSM, a spatial state-space block that replaces attention with a learnable convection-diffusion-reaction partial differential equation. This operator encodes a strong spatial prior by modeling information flow via physically grounded dynamics rather than all-to-all token interactions. Solving the PDE in the Fourier domain yields global coupling with near-linear complexity of $O(N \log N)$, delivering a principled and scalable alternative to attention. We integrate PDE-SSM into a flow-matching generative model to obtain the PDE-based Diffusion Transformer PDE-SSM-DiT. Empirically, PDE-SSM-DiT matches or exceeds the performance of state-of-the-art Diffusion Transformers while substantially reducing compute. Our results show that, analogous to 1D settings where SSMs supplant attention, multi-dimensional PDE operators provide an efficient, inductive-bias-rich foundation for next-generation vision models.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.22535

Robot-Factored World Models via Robot Rendering

Byungjun Kim, Taeksoo Kim, Hyunsoo Cha, Hanbyul Joo

原文摘要

Action-conditioned video world models predict future observations from an initial observation and an action signal. We propose robot-factored world models (RoFacto), which move two robot-specific factors outside the world model to separate action realization from scene-response prediction. First, action realization: each command sequence is rolled through the robot's own controller and kinematics without scene contact into a deployment-available nominal trajectory that does not require future realized states. Second, robot rendering: this nominal trajectory is rendered through the robot URDF, factoring the robot's geometry, kinematics, and appearance out of the model and into explicit rendered robot geometry. Paired end-effector and scene depth provide geometric cues for contact and occlusion, while camera-aware static RGB context and rendered robot geometry form a shared visual interface across viewpoints and embodiments through which the model learns scene responses. Our experiments show that the rendered interface outperforms vector-conditioned baselines, supports transfer to unseen robot embodiments, and generates robot manipulation videos from retargeted human demonstrations.

cs.RO · cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03106

A Tightly Coupled Magnetic-Inertial SLAM System for Field Robots

Chuan Huang, Tongxing Peng

原文摘要

Field robots operating in mines, tunnels, and disaster areas require reliable localization despite GNSS denial and degraded camera and LiDAR measurements caused by dust, smoke, or poor illumination. Missions involving both ground and aerial robots further motivate complementary sensing modalities that do not depend on visibility, preinstalled infrastructure, or wheel contact. This letter presents MISLAM, a tightly coupled magnetic-inertial simultaneous localization and mapping system using an IMU and a nine-magnetometer array, without relying on external odometry or platform-specific motion assumptions. The system combines the inertial navigation model with local polynomial magnetic-field models for short-term motion estimation and a global Gaussian process model for drift correction using an error-state Kalman filter. Offline evaluation on three indoor datasets collected with a handheld rig shows reductions in total endpoint closure error of 45.7% - 78.0% relative to magnetic-inertial odometry. Under poor illumination, MISLAM maintains tracking after visual-inertial odometry fails, achieving a horizontal closure error of 0.332 m, although vertical drift remains a limitation. These results support magnetic-inertial SLAM as a complementary localization modality for robotic sensor suites operating under degraded visual conditions.

cs.RO · eess.SP

查看论文
arXiv ↗PDF ↗