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
+1已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-07 19:34 · 全站最近入库：2026-11-07 17:43各信源进度 →

统计口径与数据新鲜度

统计 2026-11-06 20:14 — 2026-11-07 20:14（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：2026-11-07 19:34最近发现新增版本：2026-11-07 17:43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4,056 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

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

找到 17,100 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-07 19:30（北京时间）
新论文 · v1arXiv:2611.06776

Finding full immersions is fixed-parameter tractable in labeled tour graphs

Hadas Barabash, Jim Geelen, Rose McCarty, Caleb McFarland, Paul Wollan

原文摘要

We give an $O_H(n^5)$-time algorithm to test whether an $n$-vertex labeled tour graph has a full immersion isomorphic to a labeled tour graph $H$. A labeled tour graph is a connected $4$-regular graph whose edges are labeled by binary vectors of fixed length. To obtain a full immersion, a vertex may be removed only by pairing up its four incident edges and "splitting off" both pairs, summing the labels within each pair. Even for one-bit labels, no polynomial-time algorithm was previously known for arbitrary fixed $H$. Our main tool is of independent interest: we give an algorithmic min-max theorem that, in particular, characterizes the maximum number of odd-length circuits in a rooted circuit decomposition of an Eulerian graph. Our results have implications for vertex-minors, whose study is motivated by applications in quantum networking, connections with graph minor theory, and the goal of developing an irrelevant vertex technique for dense graphs. For vertex-minors, circle graphs play the same role as planar graphs do for graph minors. As a corollary of our main theorem, we obtain a fixed-parameter tractable algorithm for apex-vertex-minor testing in $k$-apex circle graphs. We also prove an irrelevant vertex theorem for such graphs of large rank-width. More broadly, our results are essential to forthcoming work of three of the authors proving a structure theorem for graphs with a forbidden vertex-minor. There, our min-max theorem plays the role of both the algorithmic $A$-paths and Two Paths Theorems.

cs.DS · math.CO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06775

Stability and Perfection in Constrained Weighted Network Games

Julien Fixary, Rémi Castera, Rida Laraki

原文摘要

We introduce a general framework of constrained weighted network formation in which agents allocate resources across coalitional interactions while facing feasibility constraints. More precisely, we allow arbitrary coalitions of agents to be assigned a continuous weight, and we incorporate both coalitional capacity constraints and individual budget constraints into the model. This framework extends the weighted network formation model of Bich and Morhaim by combining continuous interaction intensities with endogenous feasibility restrictions. Within this setting, we introduce a new notion of stability -- called coalitionwise stability -- extending pairwise stability to constrained network games. Under standard continuity and componentwise quasiconcavity assumptions on payoff functions, we establish the existence of coalitionwise stable networks. We then investigate refinements of stability inspired by the notion of perfect equilibrium introduced by Selten, and by the notion of perfect pairwise stability developed by Bich and Teteryatnikova for network games. The presence of feasibility constraints fundamentally alters the geometry of the feasible set, which no longer possesses a product structure. As a consequence, trembling-hand type refinements become substantially more delicate, since perturbations of some interactions may render other interactions infeasible. Nevertheless, we introduce constrained analogues of both the trembling-hand and perturbative notions of perfect stability considered by Bich and Teteryatnikova, and study their existence and their relationship. In particular, we show the equivalence between these trembling-hand and perturbative formulations under some componentwise monotonicity assumption.

cs.GT · econ.TH

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2610.20527

JetStream: Generating Verified Query Accelerators for Existing Database Systems

Akhilesh Balasingam, Amadou Ngom, Geoffrey X. Yu, Tim Kraska

原文摘要

Recent work has shown that LLMs can synthesize highly instance-optimized database systems for fixed workloads, but to date these systems only replace the DBMS's native storage and assume static data: users must migrate their data, often give up ad-hoc queries, lose acceleration once the data changes, and lack formal correctness guarantees. We present JetStream, which instead generates query-specific accelerators that extend an existing DBMS. Given a set of parameterized query templates, a staged agentic workflow synthesizes accelerators that run directly over the DBMS's native tables and may maintain custom auxiliary state, while the host DBMS remains responsible for storage, transactions, recovery, and all the other queries. A fixed, engine-neutral substrate keeps auxiliary state transactionally consistent under writes, while generated backend adapters handle engine-specific hooks. JetStream generates two kinds of accelerators: tested accelerators, which may use any optimization and are validated by differential testing, and verified accelerators, which trade some performance for a machine-checked proof that every answer satisfies a trusted specification derived from the SQL. On TPC-H at SF=20, JetStream achieves a $584\times$ geomean read speedup over DuckDB ($89\times$ verified) and retains $139\times$ ($46\times$ verified) under a continuous update stream of 10,000 modified rows per second.

cs.DB

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.34824

SPARTA: A Sparse Ternary LLM Accelerator with Adaptive Attention and Unified Datapath for Long-Context Edge Inference

Yanzhen Zhou, Zhengyu Han, Qunyou Liu, Zheyu Yan, Weihong Xu, Cheng Zhuo

原文摘要

LLMs on edge devices are constrained by limited memory and bandwidth. Although 1.58-bit ternary weights greatly reduce model storage, long-context inference is increasingly bottlenecked by KV-cache traffic and attention. Existing accelerators also use separate datapaths for ternary linear layers and attention, incurring hardware and execution overhead. We present SPARTA, a sparse ternary LLM accelerator for efficient long-context edge inference. SPARTA integrates: (1) a phase-aware sparse dataflow exploiting heterogeneous sparsity across compute-bound prefill and memory-bound decode; (2) a unified compute fabric mapping ternary SpMM/SpMV and bit-serial attention onto shared processing elements to exploit zero-weight and zero-bit sparsity; and (3) an adaptive two-stage attention engine combining page-level Top-k selection to reduce K-cache accesses with token-level Top-p pruning to reduce V-cache traffic. SPARTA reduces KV traffic by 4.1-4.8x over full attention and 1.3-1.5x over a state-of-the-art sparse-attention scheme, with task-dependent quality loss. Compared with NVIDIA Jetson AGX Orin, SPARTA achieves 3.9x higher generation throughput and 8.3x higher end-to-end energy efficiency on average; versus AMD Ryzen 7 9700X, the gains are 5.0x and 81.7x. It also delivers 1.6-12.0x higher throughput, 2.0-4.2x higher energy efficiency, and 1.3-51.4x higher area efficiency than FIGLUT, TeLLMe, and MCBP.

cs.AR · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06773

Approximation and Hardness for the Data Validation Problem

Antoine Tinguely, G. Wessel van der Heijden

原文摘要

Finnah, Gönsch and Otto [Eur. J. Oper. Res. 2025] recently introduced the Data Validation Problem (DVP): Given a transitive directed acyclic graph $G$ on $n$ vertices (resp. tasks) and $m$ edges (resp. precedences) and a probing budget, a policy can successively pay a probing fee and probe an edge in $G$ as long as the total probing fees do not exceed the budget. If a probe succeeds, the targeted edge $e$ is removed from $G$ and we receive a reward; if the probe fails, the edge stays in $G$ and cannot be probed again. We seek a policy with maximum expected cumulative reward. Crucially, $G$ must stay transitive throughout the probing process, so a policy can only challenge the precedence between consecutive tasks, that is, only probe edges in the transitive reduction of the current precedence graph. We show that DVP is strongly NP-hard and W[1]-hard with the probing budget $T$ as a parameter, even in the deterministic case (when every probe has probability $0$ or $1$ to succeed) and with unit probing costs and rewards. For the same setting, we give an approximation-preserving reduction from the Densest $k$-Subgraph (D$k$S) Problem, with the consequence that there is no polynomial-time $2^{(\log n)^{1-δ}}$-approximation algorithm and no FPT $\sqrt{T}$-approximation algorithm for DVP, assuming the Almost-Exponential Time Hypothesis and the Strongish Planted Clique Hypothesis, respectively. We also present the first known policies with provable worst-case guarantees for DVP, namely an $m$-approximation algorithm for instances with uniform probabilities and probing costs, a $2T/p_{\min}^T$-approximation algorithm for general instances, where $p_{\min}$ is the minimum probing success probability, and an exact FPT algorithm on tournament graphs, in the deterministic setting.

cs.DS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06772

Energy Efficient Near-Threshold Voltage Massive MIMO Computing

MD Salman Sakib, Mehdi Safarpour, Pejman Ghasemzadeh, Shahriar Shahabuddin

原文摘要

Near-Threshold Computing (NTC) offers a substan- tial energy savings, i.e., up to 10x, for digital circuits. However, due to Process, Voltage, and Temperature (PVT) variations, low voltage operating digital processors face increased susceptibility to timing errors. This work equips conventional massive MIMO algorithms with computational error detection mechanisms, en- abling reliable operation down to near-threshold voltages. While conventional hardware based Timing Error Detection (TED) systems introduce substantial overhead, this work modifies MIMO algorithms to seamlessly integrate low overhead Algorithm Based Fault Tolerance (ABFT). A co-simulation framework combining transistor level SPICE simulations and workload aware fault injection is developed to investigate reliability, Bit Error Rate (BER), energy savings, and overheads of modified algorithms operating at low voltages. The transistor-level characterization was performed in PSpice using generic Level-1 Shichman-Hodges CMOS transistor models with Monte Carlo parameter variations. No specific CMOS fabrication process node was assumed. Conse- quently, the reported energy savings are analytical estimates and are not specific to a particular CMOS fabrication technology. We evaluate Gauss Seidel (GS), ADMM based infinity norm detection (ADMIN), and Conjugate Gradient (CG) detectors using circuit derived fault models. Results demonstrate remarkable algorith- mic resilience. Although undetected computational errors in some trials near the Point of First Failure (PoFF) were observed, BER increases by only 0.7% to 1.4%, while beyond PoFF, ABFT reported nearly all erroneous cases. The proposed checksum based ABFT detects most arithmetic and memory errors, with an estimated overhead below 2% for a 64 x 16 MIMO system, decreasing further for larger configurations.

cs.AR

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.05800

SALT: When More Rollouts Don't Help in Group-Based Policy Optimization and How to Make Them Matter

Powei Chang, Jinpeng Zhang, Chaoqun Sun, MiniWell Tsao, Lianrui Li, Jianxiang Xiang, Chenyu Wang, Yukang Gao 等 9 位作者

原文摘要

Reinforcement learning with verifiable rewards (RLVR) often adopts GRPO-style group-relative updates, sampling multiple rollouts per prompt to construct normalized learning signals. However, merely increasing the number of rollouts does not reliably strengthen learning: under GRPO-style group normalization, per-rollout policy-gradient features can concentrate into a low-rank, signed geometry, causing substantial cancellation during aggregation and weakening the effective update. We address this failure mode with SALT, a Subspace-Adaptive geometry pLug-in componenT that uses sample-wise gradient geometry to reweight the coefficients of group-relative updates. SALT estimates a dominant shared subspace from the mini-batch Gram geometry, decomposes group-relative coefficients into shared and residual channels, and adaptively amplifies the residual channel when signed cancellation is severe. Across diverse reasoning-oriented RLVR benchmarks and model scales, SALT improves effective update geometry and performance without modifying the reward model or the rollout sampling procedure.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06988

Big Models, Small Gains: Frozen Forecasters as Classifiers on MONSTER

Rishit Jain, Rishit Kar, Chetashri Bhadane, Nilesh Patil

原文摘要

Forecasting foundation models are increasingly used as frozen feature extractors for time series classification, but it remains unclear whether their pretrained representations offer an advantage over strong classifiers that do not require pretraining. This raises an important question: do these representations provide useful information beyond what the model architecture alone can capture, particularly when labelled data are scarce? We present a controlled benchmarking of two forecasting models, Reverso and Chronos-Bolt, on six large datasets from the MONSTER benchmark, using 1%, 10%, and 100% of the available training labels. We compare their frozen representations against four established time series classification methods and random-weight versions of the same forecasting models. Our results show that pretraining improves classification accuracy over random-weight counterparts on five of the six datasets, with gains of up to 23 percentage points. However, the frozen representations do not reliably outperform the strongest non-pretrained baseline in any setting: they perform worse in 15 of 17 settings and tie in the remaining two. Further experiments show that restoring amplitude information recovers part of the performance gap, while fine-tuning and forecast-error features do not close it. These findings highlight the importance of strong task specific baselines and random weight controls when evaluating forecasting foundation models for classification. Our code is available at https://github.com/Rachet2006/frozen-forecasters-monster.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06770

Benchmarking Fine-Grained Visual Reasoning for Remote Sensing in Urban Management

Jiawei Zhang, Rizhao Cai, Dayan Guan

原文摘要

Fine-grained visual reasoning is essential for remote sensing in urban management, yet remains underexplored in the evaluation of multimodal large language models (MLLMs). We introduce UMBench, a fine-grained urban remote sensing benchmark built from unmanned aerial vehicle (UAV) images for multiple urban management tasks such as traffic monitoring, urban construction, environmental quality, and disaster warning. Evaluations on UMBench reveal that current MLLMs often rely on scene-level priors and overlook fine-grained visual evidence. To address this issue, we further propose a novel evidence-first chain-of-thought method that directs models to inspect fine-grained visual evidence in a training-free manner. Experiments show that our method can improve the performance of both open-source and commercial MLLMs on UMBench.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06768

PyRUA-REAL: Robot Exploration and Adaptive Learning through RSI in the Real World

Qiang Wang, Jianxin Bi, Shulong Jiang, Ruiyang Si, Wenbo Huang, Kaiqi Chen, Jiaxin Li, Liyun Li 等 10 位作者

原文摘要

Coding agents powered by models such as GPT-6-Astra connect spatial reasoning with executable robot behavior, raising the question of how such agents can produce lasting improvements in robot capabilities from physical interaction. In this work, we study embodied recursive self-improvement (RSI) through real-world exploration, code refinement, and deployment using PyRUA-REAL, a framework in which coding agents retain interaction experience as persistent memory and reusable executable skills that inform subsequent exploration and improvement. Our empirical study places physical manipulation at the center of the improvement process, examining how agents collaborate, what kinds of execution feedback support effective code refinement, and how task specifications and previously acquired skills influence robustness and task transfer. We further investigate the efficiency and capability of the resulting programs when deployed without involving an online agent for decision-making. Across eight real-world tasks, PyRUA-REAL achieves an average success rate of 91.1%. In complementary evaluation on 42 tasks from the RoboDojo benchmark, it achieves an average success rate of 84.4%. This study provides empirical guidance for turning physical exploration into reusable robot skills that support continued improvement and robust, efficient deployment.

cs.RO

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2606.16952

Phantoms and Disclosures: A Statistical Framework for Auditing Privacy in Synthetic Data

Kareem Amin, Rudrajit Das, Alessandro Epasto, Adel Javanmard, Dennis Kraft, Mónica Ribero, Sergei Vassilvitskii

原文摘要

The rapid adoption of generative AI and Large Language Models (LLMs) has spurred interest in synthetic data as a privacy-preserving alternative to sensitive real-world datasets. However, generating high-utility synthetic data often carries the risk of memorizing and regurgitating private information from the training corpus. In this work, we present a customizable empirical auditing framework designed to detect and explain such data disclosures. Our framework introduces a mechanism to distinguish between "true disclosures"-where the system directly reproduces a user's information-and "phantom disclosures''-where the system incidentally generates a user's data. By partitioning input data into training and holdout sets and applying rigorous statistical hypothesis testing, we determine if observed disclosures are consistent with strict privacy baselines, such as zero-learning or specific Differential Privacy (DP) bounds. Crucially, this approach requires no model access, no canary insertion, and no reference model training -only the synthetic output and a held-out control set. We demonstrate that this framework effectively functions as a membership inference attack, providing empirical lower bounds on privacy leakage that are tighter than prior data-based auditing methods. Our approach is model-agnostic, applies to any synthetic data generation mechanism, and requires orders of magnitude fewer computational resources than shadow-model or canary-based alternatives.

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

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.26826

RefCraft: Benchmarking Visual-Reference-Conditioned Agentic Generation through Web Development

Yuchi Wang, Rongyao Fang, Weihao Yu, Xilin Wei, Tiancheng Gu, Dingkang Yang, Hang Zhang, Dayiheng Liu 等 10 位作者

原文摘要

Beyond basic tool use and coding skills, effectively incorporating visual references into agentic generation is equally essential for applying AI agents to real-world productivity tasks. We study this crucial capability through webpage development, a practical and readily evaluable setting in which existing benchmarks remain largely idealized: they typically focus on reconstructing visual targets or generating webpages from textual instructions alone, overlooking the more common workflow of developing from visual examples. Specifically, we introduce RefCraft, a benchmark comprising 300 realistic multimodal webpage-development tasks that combine natural-language instructions with screenshots from one to five reference webpages. Agents must selectively interpret and utilize these references while rendering, visually inspecting, and iteratively refining their outputs. We evaluate the resulting open-ended artifacts using independently verifiable keypoints and complementary programmatic and VLM judges. Across nine frontier (M)LLMs, the strongest system achieves an overall score of only 52.12 out of 100, revealing that visual-reference-conditioned agentic generation remains a substantial challenge for current models in practical webpage-development settings.

cs.CL · cs.AI · cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2608.14430

Designing Reinforcement Learning for Diffusion Models: A Unified Path-Space View

Yixian Xu, Yuanrui Zhang, Shengjie Luo, Kaiyuan Gao, Liwei Wang, Di He

原文摘要

Reinforcement learning (RL) post-training provides a direct way to align diffusion models with human preferences and task-specific rewards. However, current RL algorithms for diffusion models remain fragmented: reverse-trajectory methods rely on discretized likelihood ratios, whereas forward-matching methods train on reward-labeled noising versions of the rollout samples. This paper shows that these seemingly different losses arise from a single path-space principle. Starting from the regularized diffusion-RL objective, we use importance sampling between sampling SDEs to obtain an explicit policy-gradient estimator on trajectory space. The estimator contains the stochastic Itô integral underlying Flow-GRPO-type updates; we derive an equivalent variance-reduced value-gradient form that recovers the forward-matching structure of AWM and DiffusionNFT. This identifies the empirical gap between these method families as a variance-reduction effect rather than a difference in RL principle. The derivation yields a unified design space organized by value-gradient estimation, weight functions, and sampling choices. Within this space, we propose a multi-sample KDE value-gradient estimator that reuses rollout groups, together with scale-bounded weight families that retain stable existing recipes while excluding singular ones. Experiments on SD3.5-M and Qwen-Image models validate the variance-reduction explanation and show that the resulting recipe improves over prior diffusion-RL baselines.

cs.LG · cs.CV · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06764

Rethinking Access Control for AI Agents: Recent Advances and Open Challenges

Chijin Zhou, Quan Zhang

原文摘要

Large language model agents access data, invoke tools, and act on users' behalf. Keeping these actions within authorized bounds is challenging because agents choose their own execution paths while permissions and operating conditions may change. This paper surveys recent advances in agent access control across four areas: policy specification and derivation, identity and delegation, action and resource enforcement, and information flow control. We compare their control objects, enforcement boundaries, and assumptions to examine what individual mechanisms achieve and how they can work together. Building on this analysis, we rethink four aspects of authorization: plan selection shapes permission requirements; contextual information can inform decisions without itself granting authority; restrictions must retain their meaning across system interfaces; and authorization must be reassessed as workflows evolve. These insights motivate open challenges and future directions, together with concrete research questions and evaluation criteria. In particular, evaluation should examine both observed agent behavior and the unauthorized effects that remain enabled. Our analysis connects existing mechanisms to the requirements for preserving authorization throughout an agent workflow.

cs.SE

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.14836

A Diffusion-Based Kinematics and Dynamics Model for 3D Sign Language Production

Emanuele Colonna, Moises Diaz, Gennaro Vessio, Miguel Angel Ferrer, Giovanna Castellano

原文摘要

Sign Language Production (SLP) generates continuous signing from a linguistic input and drives accessibility technologies such as signing avatars. Current generators, diffusion models included, regress joint coordinates as an unstructured vector, so nothing ensures that the generated body respects anatomy or the mechanics of movement. Physics has entered SLP only through geometric loss terms, whose benefit for intelligibility has not been verified, and the dynamics of the signing body has not been modelled. We present PIDiffSign, a diffusion model informed by physics at two levels. For kinematics, a forward-kinematics head outputs bone directions and lengths and composes them into an anatomically exact skeleton. For dynamics, a rigid-body model of the upper body with anthropometric parameters supports two training priors. A physical-plausibility suite that includes two laws of motor control evaluates the generated signing. Under a controlled protocol, the head lowers back-translation word error rate by about five points on PHOENIX14T and CSL-Daily, whereas geometric penalties have no measurable effect; a control experiment traces the gain to the kinematic-chain representation and not to bone-length conservation. The dynamics priors gain about one further point over matched non-physical controls on PHOENIX14T, but not on CSL-Daily. Generated signing follows the laws of motor control as closely as real signing but carries markedly less kinetic energy. The exact skeletons can drive rigged avatars directly, and the suite offers a physical benchmark for future SLP systems.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06762

U.S. State Guidance on AI in K-12 Education: A Comparative Content and Citation Analysis

Shahan Ali Memon, Anna-Maria Gueorguieva, Anukriti Kumar, Lucy Lu Wang, Jevin D. West, Nic Weber

原文摘要

In the years following the public release of ChatGPT in late 2022, more than half of U.S. states have published formal guidelines around AI usage in K-12 education, creating a record of largely independent state education departments responding to a shared technological disruption. In this work, we analyze publicly available AI usage guidance documents from 35 states in the U.S. and Puerto Rico. We compare the guidance in these documents across various dimensions, including pedagogy and instruction, audience, actionability, workforce and career readiness, responsible use and equity, and alignment with state-level demographics and administrative capacity. Finally, we investigate who authored these documents and their cited sources, identifying epistemic foundations underlying their claims. We find that guidance priorities converge on topics anchored in laws such as data privacy and academic integrity, each addressed in 97% of documents. While most documents mention equity of access (86%) and algorithmic bias (97%), fewer discuss specific marginalized populations. Across documents, there is a common framing of urgency around AI integration to prepare students for workforce readiness (86%) and promote loosely-defined personalized learning (86%), which is not matched by specific guidance for capacity-building, actionability, or accountability. State priorities show no relationship to state demographics, capacity, or partisanship. Instead, we identify notable interdependencies between states with shared citations to non-profit organizations like TeachAI and Code.org, with some states reproducing TeachAI's model guidance near-verbatim, and peer citations concentrating on a few early states such as Oregon and Washington. No state includes parents as contributors on these documents, and only five include students.

cs.CY · cs.DL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06760

Route First, See Later: Enabling Vision Agents to Reason Scientifically by Separating Routing Policy from Action Spaces

Shubhabrata Mukherjee, Jeffrey Donatelli, Daniela Ushizima

原文摘要

Scientific image-analysis workflows require agents to select among diverse capabilities like description, detection, counting, and segmentation. Current multi-tool architectures conflate intent interpretation with visual execution, masking whether a failure arises from flawed reasoning (misrouting) or failed perception (invalid tool execution). This lack of auditability renders agents unreliable for high-stakes scientific automation. We introduce ZETA, an extensible architecture that operationalizes the decoupling of routing policy from action execution. ZETA resolves scientific intent before invoking specialized visual tools, treating measurement as a deterministic operation grounded in explicit geometry rather than generative output. This architecture exposes the "black box" of agent decision-making, allowing routing correctness and downstream execution to be evaluated as independent failure modes. To our knowledge, we provide the first controlled study in scientific vision-agent routing that isolates explicit routing-policy specification while holding the legal action interface fixed. Evaluation on a zero-shot benchmark of 51 scientific-intent queries shows that explicit routing semantics can substantially improve routing accuracy, particularly at semantically adjacent action boundaries, although improvements vary across models. For Qwen 3.8, explicit routing-policy specification improves segmentation routing accuracy from 6.2% to 75.0% and detection routing accuracy from 60.0% to 100%. ZETA's execution traces make routing decisions and downstream artifacts separately observable, providing a basis for distinguishing intent-selection errors from visual execution failures. ZETA establishes the necessary infrastructure to isolate execution errors from interpretive ones, a critical requirement for deploying reliable scientific vision agents.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06759

Continual Classification Beyond the Kernel Regime: Harnessing the Sparsity Bias of L2 Regularization in Diagonal Linear Networks

Gilad Karpel, Itay Evron, Ron Meir, Daniel Soudry

原文摘要

We study logistic regression for continual learning with linear predictors in a high-dimensional statistical setting. Existing theoretical work has largely focused on continual linear regression with variants of L2 regularization, motivated by their analytical simplicity and connection to the kernel regime of neural networks. We first derive the optimal L2 regularization strength for continual linear classification and characterize its generalization performance, recovering behavior analogous to prior results for continual linear regression. Next, we examine the effect of L2 regularization in deeper models using a minimal two-layer diagonal linear network. In the single-task setting, L2 regularization on the weights of such models is known to promote sparsity in the resulting predictor, yielding a rich, sparse feature-learning regime beyond the kernel regime (a similar effect arises in two-layer ReLU networks). However, in continual learning, where L2 regularization is typically centered on previous-task weights, this sparsity bias is no longer preserved. We therefore propose a weighted variant of L2 regularization that extends this beneficial sparsity bias to continual learning. We show theoretically and empirically that, for a sufficiently sparse teacher, our variant converges substantially faster than directly learning a linear predictor with conventional continual L2 or L1 regularization, trading some long-horizon bias for markedly improved performance over moderate task horizons.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2511.17747

AEGIS: Preserving Privacy of 3D Facial Avatars with Adversarial Perturbations

Dawid Wolkiewicz, Anastasiya Pechko, Przemysław Spurek, Piotr Syga

原文摘要

The growing adoption of photorealistic 3D facial avatars, particularly those utilizing efficient 3D Gaussian Splatting representations, introduces new risks of online identity theft, especially in systems that rely on biometric authentication. While effective adversarial masking methods have been developed for 2D images, a significant gap remains in achieving robust, viewpoint-consistent identity protection for dynamic 3D avatars. To address this, we present AEGIS, the first privacy-preserving identity masking framework for 3D Gaussian avatars, concealing identity from automated face recognition while preserving the avatar's perceptual realism and animation usability. AEGIS applies adversarial perturbations to the base (DC) color coefficients of the Gaussian primitives, guided by an ensemble of pre-trained face verification networks, embedding the protection directly in the 3D representation so that it applies across viewpoints and animations, without retraining or modifying the avatar's geometry. Evaluated against black-box recognizers, AEGIS reduces face retrieval and verification accuracy to near-zero while maintaining high perceptual fidelity. It also preserves key facial attributes such as age, gender, and emotion, demonstrating strong privacy protection with minimal visual distortion. We further introduce a viewpoint-robust leakage metric that quantifies residual identity leakage across an avatar's rendered viewpoints, which single-view evaluation cannot capture.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2601.06641

Leveraging Soft Prompts for Privacy Attacks in Federated Prompt Tuning

Quan Minh Nguyen, Min-Seon Kim, Hoang M. Ngo, Trong Nghia Hoang, Hyuk-Yoon Kwon, My T. Thai

原文摘要

Membership inference attacks (MIAs) pose a serious privacy threat in federated learning (FL). While MIAs have been extensively studied in standard FL, the recent shift toward federated fine-tuning introduces new and largely unexplored attack surfaces. In this work, we show that federated prompt-tuning, which adapts pre-trained foundation models using lightweight input prefixes, exposes a novel and effective vector for membership inference. We propose PromptMIA, a membership inference attack tailored to federated prompt-tuning, in which a malicious server introduces adversarially crafted prompts and exploits their updates during collaborative training to determine whether a target data point belongs to a client's private dataset. We formalize this threat via a security game and demonstrate that PromptMIA achieves consistently high attack advantage across diverse benchmark datasets, substantially outperforming current SOTA federated MIAs. We also provide a theoretical lower bound on the attack advantage that explains the observed empirical behavior. Finally, we show that existing MIA defenses are often ineffective against PromptMIA, highlighting the need for defense mechanisms specifically tailored to prompt-tuning in federated settings.

cs.LG · cs.CR

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.10795

Factual recall in linear associative memories: sharp asymptotics and mechanistic insights

Alessio Giorlandino, Sebastian Goldt, Antoine Maillard

原文摘要

Large language models excel at factual recall, yet the fundamental limits of storing input--output associations with neural networks remain unclear. We study these limits in a minimal setting: a linear associative memory that maps $p$ input embeddings in $\mathbb{R}^d$ to their $d$-dimensional targets via a linear layer, requiring each mapped input to score its own target above all other outputs. Since all inputs compete for the same outputs, these constraints are strongly correlated, which makes the storage capacity $p_c$ --- the largest number of associations that can be stored --- hard to characterise. We address this in three steps. First, we introduce a decoupled model, in which each input competes with its own independent set of outputs, and provide numerical and analytical evidence that it is equivalent to the original model, both in capacity and in the structure of the learnt solution. Second, using tools from statistical physics, we show that the decoupled model has capacity $p_c \log p_c / d^2 = 1/2$, extend this computation to linear two-layer networks, and go beyond the capacity by predicting the spectrum of the learnt weights and the storage mechanism. The latter explains how the optimal solution outperforms a naïve Hebbian rule: rather than boosting input--output alignments with broad fluctuations, it raises the correct scores just above the extreme-value threshold set by the competing outputs. Third, we prove the decoupled model's capacity thresholds rigorously, including in the non-convex two-layer case; combined with a recent proof for the original model, this shows that both models have the same capacity for full-rank weights. These findings give a sharp characterisation of factual storage in linear networks, and suggest decoupling as a general tool for memorisation and learning problems in which each input must be matched to one of an extensive number of outputs.

stat.ML · cond-mat.dis-nn · cond-mat.stat-mech · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06754

A Unified Framework for Extending Centrality Measures to Groups

Piotr Kępczyński, Oskar Skibski

原文摘要

We propose a unified framework for extending centrality measures from individual nodes to groups. For each centrality measure, we introduce four group variants---maximum, minimum, sum, and proportional---which differ in how individual contributions are aggregated at the group level. The framework builds on a new unified perspective on major classes of centrality measures, according to which the score of a node can be expressed as a sum of assessments made by individual nodes, pairs of nodes, or groups of nodes. Each of the four variants then corresponds to a different rule for aggregating the assessments of the members of a group.

cs.SI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.14106

Token Time Continuous Diffusion for Language Modeling

Parikshit Bansal, Sujay Sanghavi

原文摘要

In this paper we introduce token time continuous diffusion (TTCD), a new diffusion language model which (a) operates in continuous space, deterministically mapping Gaussian noise to a final token canvas with no further sampling, and crucially (b) incorporates a new notion of per-token times, with some tokens proceeding from noise to token at a faster rate than others. Continuous space modeling helps TTCD avoid the parallel sampling of multiple tokens, which is a key source of inaccuracy at high speedups for models that iterate purely in discrete space. The notion of per-token times helps TTCD to better model conditional generation, allows for more sure tokens to proceed at a faster rate, and allows for differentiated inter-token influences during refinement. TTCD outperforms discrete models at high speedups. We train a 160M parameter TTCD model on OpenWebText, and then self-distill it; we find that at high speedups we are comparable in unconditional generation quality, and outperform in conditional generation, several existing models of similar size trained, on the same data, and self-distilled. We achieve similar gains in Sudoku solving as well.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.03633

Pretraining Data Scale Reverses the Muon-AdamW Fine-Tuning Gap

Catherine Li, Jingchu Gai, Andrej Risteski, Jacob Mitchell Springer, Aditi Raghunathan

原文摘要

Muon has recently emerged as a promising optimizer for pretraining large language models, offering a speedup over AdamW in minimizing pretraining loss. However, it remains unclear whether Muon's advantage persists after fine-tuning. In this paper, we study the downstream behavior of Muon and AdamW and uncover a surprising trend: beyond a sufficiently large pretraining data budget, AdamW outperforms Muon after fine-tuning. Moreover, as model size increases, AdamW overtakes Muon at progressively smaller pretraining token-parameter ratios. To explain this reversal, we view fine-tuning as a perturbation of the outputs and introduce the margin-to-spread ratio, which measures how far the correct logit lies above the wrong ones relative to how widely the wrong logits are spread. In language models, Muon has the larger margin-to-spread ratio at small data budgets and AdamW at large ones, mirroring the fine-tuning trend. In a linear associative memory, we prove that after one update the margin-to-spread ratio of gradient descent gains on that of Muon as the data budget grows, and simulations show that this trend persists after extended training.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2605.10832

Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents

Shijue Huang, Hangyu Guo, Guanting Dong, Chenxin Li, Junting Lu, Xinyu Geng, Zhaochen Su, Zhenyu Li 等 11 位作者

原文摘要

Multimodal deep search requires an agent to solve open-world problems by chaining search, tool use, and visual reasoning over evolving textual and visual context. Two bottlenecks limit current systems. First, existing tool-use harnesses treat images returned by search, browsing, or transformation as transient outputs, so intermediate visual evidence cannot be re-consumed by later tools. Second, training data is usually built by fixed curation recipes that cannot track the target agent's evolving capability. To address these challenges, we first introduce a visual-native agent harness centered on an image bank reference protocol, which registers every tool-returned image as an addressable <image:N> reference and makes intermediate visual evidence reusable by later tools. On top of this harness, On-policy Data Evolution (ODE) runs a closed-loop data generator that refines itself across rounds from rollouts of the policy being trained. This per-round refinement makes each round's data target what the current policy still needs to learn. The same framework supports both diverse supervised fine-tuning data and policy-aware reinforcement learning data curation, covering the full training lifecycle of the target agent. Across 8 multimodal deep search benchmarks, ODE improves the Qwen3-VL-8B agent from 24.9% to 39.0% on average. At 30B, ODE raises the average score from 30.6% to 41.5%. Further analyses validate the effectiveness of image bank reuse, especially on complex tasks requiring iterative visual refinement, while rollout-feedback evolution yields more grounded SFT traces and better policy-matched RL tasks than static synthesis.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06751

A Quantum Lower Bound for Triangle Finding

Arjan Cornelissen, Amin Shiraz Gilani, Xingyu Zhou

原文摘要

Triangle finding is a fundamental problem in graph theory that has played a central role in the development of quantum algorithmic techniques. For $N$-vertex graphs given by their adjacency matrices, a long line of work has improved the quantum query upper bound from $O(N^{3/2})$ to $O(N^{5/4})$. However, the best known lower bound has remained the elementary $Ω(N)$ bound obtained from unstructured search. We prove the first superlinear quantum query lower bound for triangle finding, showing that $Ω(N^{13/12})$ queries are necessary. Our result also implies the same lower bound for triangle detection and for detecting every fixed connected non-bipartite graph. The proof builds on the compressed oracle approach of Zhandry and the transfer approaches of Belovs. It compares the computation on graphs with a planted triangle to the computation on random graphs conditioned to be triangle-free, through a map on the input registers that almost commutes with every query. We average over local changes to the input while preserving its distribution, through a procedure known as conditional resampling. Our main technical contribution is a set of estimates that let us control the query error under this averaging, despite the dependence introduced by conditioning a dense graph to be triangle-free.

quant-ph · cs.DS

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.06598

SimForcing: Distilling Simulation Motion Priors into Real-Domain Robot World Models

Xiaodong Wang, Tianle Li, Chuanxin Song, Junliang Xie, Zhanmi Zhong, Suiying Wu, Peixi Peng

原文摘要

Action-conditioned robot world models must respond precisely to robot trajectories while preserving realistic visual dynamics, yet learning both from heterogeneous robot videos remains challenging. Simulation offers structured motion supervision, but appearance differences hinder direct transfer, and inaccurate simulation predictions can misguide real-video generation. We present SimForcing, a simulation-guided framework that uses simulation both as a source of transferable motion knowledge and as a controllable reference for prediction. First, we transfer motion knowledge from a simulation teacher through latent-motion distillation, aligning temporal changes in latent space to internalize motion priors while mitigating the influence of appearance differences. Second, we introduce multi-block simulation conditioning with condition dropout to exploit predicted simulation trajectories without relying excessively on their accuracy. Our simulation-conditioning classifier-free guidance scheme unifies these two ideas by balancing predictions based on internalized motion knowledge with those additionally guided by simulation latents. The jointly trained student generates both simulation conditions and real-domain videos, requiring no additional world model at inference. On Bridge, SimForcing achieves the best PSNR, SSIM, LPIPS, and FVD among the compared methods without external embodied pretraining. Evaluation on InternData-A1 further supports its applicability across robot datasets. Moreover, using our trained world model to initialize a vision-language-action model improves LIBERO success, suggesting its utility for downstream policy learning. https://github.com/Wang-Xiaodong1899/SimForcing

cs.RO · cs.AI · cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06747

Fixed-Radius Neighbourhoods in Three-Dimensional Unit-Cube Graphs

Guangjian Zhang

原文摘要

We prove Conjecture 4 of Chan, Chang, Gao, Kisfaludi-Bak, Le, and Zheng (SoCG 2026): for every fixed integer $r\ge0$, the radius-$r$ neighbourhoods of a unit-cube graph in $\mathbb{R}^3$ have bounded VC-dimension. Our bound is $2^{2^{O(r)}}$, independent of the number of vertices. The main ingredient is a VC-dimension bound for reachability through arbitrary fixed-length sequences of three-dimensional generalized dominance relations. Its proof uses groups ordered in two coordinates, with neighbouring groups acting as barriers that preserve this invariant when one layer is removed. Combining these bounds with the existing interval-representation framework, we obtain a Las Vegas algorithm that, for every fixed integer $Δ\ge1$, decides whether a graph given by $n$ cube centres has diameter at most $Δ$ in expected $O_Δ(n^{2-\varepsilon_Δ})$ time, for an explicit constant $\varepsilon_Δ>0$. Finally, for every integer $r\ge7$, we construct unit-cube graphs with rational centres whose radius-$r$ neighbourhoods have VC-dimension $Ω(r^3)$.

cs.CG · math.CO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06745

Generalized Sorting via Four-Cycles

Soheil Behnezhad, Shane Ferrante

原文摘要

We study the generalized sorting problem where only a subset of the pairs can be compared. Equivalently, we are given an $n$-vertex $m$-edge graph $G=(V, E)$ where the vertex set $V$ is to be sorted, and the edge set $E$ indicates possible comparisons. The goal is to sort $V$ with as few (adaptive) queries to $E$ as possible. When $G$ is the complete graph, the problem is equivalent to standard sorting for which $Θ(n \log n)$ queries are sufficient and necessary. All classical sorting algorithms crucially rely on the ability to compare any pair and do not imply an $o(n^2)$-query algorithm in the general case. This was improved 15 years ago by Huang, Kannan, and Khanna [FOCS'11] who achieved a bound of $\widetilde{O}(n\sqrt{n})$. Kuszmaul and Narayanan [FOCS'21] later gave a bound of $\widetilde{O}(\sqrt{m n})$, which is more efficient for sparse graphs, but matches the $\widetilde{O}(n\sqrt{n})$ barrier for dense graphs. In this paper, we present an algorithm for generalized sorting that makes at most $\widetilde{O}(n^{4/3})$ queries, therefore breaking the longstanding $\widetilde{O}(n\sqrt{n})$ barrier for all graphs. The key to our result is to query edges that belong to many (specific types of) four-cycles. We show that querying such edges allows the algorithm to verify the direction of many other edges (specifically, their opposing edges in the four-cycles) and thus achieve a much better amortized query cost.

cs.DS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.06743

Using AI Predictions to Augment Rather Than Replace Surveys

Lai Wei, Yifei Zhu

原文摘要

LLMs can predict human survey responses with nontrivial accuracy. What this means for the future of social surveys remains hotly contested. Optimists anticipate a coming era of synthetic respondents; skeptics see a fragile and overhyped substitute. This article advocates a position: AI predictions cannot replace surveys, but they can augment them. Using prediction-powered inference (PPI), a statistical framework rooted in survey methodology and known in sociology as "mixed subjects design", the study tests whether AI predictions can improve the statistical efficiency of social surveys when combined with census-scale auxiliary information. Applying the framework to 83 social outcomes in the General Social Survey (GSS) and National Health Interview Survey (NHIS), with American Community Survey (ACS) being the auxiliary file, we find that supervised predictors yield effective-sample-size gains of about 7-12 percent when estimating the mean, while LLM gains are smaller but positive.

cs.CY

查看论文
arXiv ↗PDF ↗