跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+197新增论文
+151其中 AI 相关
+16已有论文修订
+91新增 AI 导读

arXiv 最近完成扫描:2026-11-22 06:41 · 全站最近入库:2026-11-21 12:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-21 06:49 — 2026-11-22 06:49(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:2026-11-22 06:41最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 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。

29,158 已收录论文24,181 AI 相关8,992 有修订112 HF 社区发现

浏览论文库

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 21,490 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-22 06:39(北京时间)
新论文 · v1arXiv:2611.20622

Observable Spectral Framework for Noisy Amplitude Amplification: Stopping Stability and Correction

Ge Zhong, Charith Jayasekara, Nisal De Silva, Muhammad Usman, Muhammad Aamir Cheema, Chetan Arora

原文摘要

Stopping rules for noisy amplitude amplification often rely on channel-specific success-probability formulas, limiting their transfer across noise mechanisms. We develop an observable spectral framework for stationary noisy standard amplitude amplification under explicit leading-mode, spectral-separation, and residual-control conditions. Within this class, the first amplification lobe has a damped-oscillatory leading component and a bounded non-leading contribution. We derive how frequency shift, damping, and observable phase offset displace the leading peak, and establish sampled-gap criteria for retaining the ideal integer stopping depth, certifying recalibration, or selecting a robust candidate window. In single-target Grover search, frequency and damping errors have stopping-coordinate sensitivities proportional to the search-space size, whereas phase offsets scale with its square root; residual modes can independently reverse the ordering of neighbouring depths. Analytical constructions, circuit simulations, finite-shot inference, and calibration-derived device-like trajectories examine these predictions and their operational limits. For a 64-element search, an a posteriori sampled-set check selects depth five over six, with success probability increasing from 0.595 to 0.626. The resulting theory links observable spectral dynamics to conditional integer stopping guarantees without requiring a separate microscopic channel solution.

quant-ph · cs.ET

新论文 · v1arXiv:2611.20621

GPromptShield: Elevating Resilience in Graph Prompt Tuning against Adversarial Attacks

Shuhan Song, Ping Li, Ming Dun, Maolei Huang, Huawei Cao, Xiaochun Ye

原文摘要

The paradigm of ``pre-training and prompt-tuning", with its effectiveness and lightweight characteristics, has rapidly spread from the language field to the graph field. Several pioneering studies have designed specialized prompt functions for diverse downstream graph tasks based on various graph pre-training strategies. These prompts concentrate on the compatibility between the pre-training pretext and downstream graph tasks, aiming to bridge the gap between them. However, designing prompts blindly to adapt to downstream tasks based on this concept neglects crucial security issues. By conducting covert attacks on downstream graph data, we find that even when the downstream task data closely matches that of the pre-training tasks, it is still feasible to generate highly misleading prompts using simple deceptive techniques. In this paper, we shift the primary focus of graph prompts from compatibility to vulnerability issues in adversarial attack scenarios. We design a highly extensible shield defense system for the prompts, which enhances their robustness from two perspectives: Direct Handling and Indirect Amplification. When downstream graph data contains unreliable biases, the former directly combats invalid information by incorporating hybrid multi-defense prompts to the input graph's feature space, while the latter adopts a training strategy to bypass the invalid components and amplifies valid part. We provide a theoretical derivation that proves their feasibility, indicating that unbiased prompts exist under certain conditions on unreliable data. Extensive experiments across various scenarios of adversarial attacks indicate that the prompts within our defense system exhibit enhanced resilience and superiority. This paper explores a new perspective in graph prompt learning, offering a novel option for robust prompt tuning in downstream tasks.

cs.LG

版本更新 · v5arXiv:2501.15422

TTC Domains

Sumit Goel, Yuki Tamura

原文摘要

The Top Trading Cycles (TTC) rule is fundamental to object reallocation: it is the only rule that is individually rational, pair efficient, and strategyproof on the unrestricted domain. Despite its desirable properties, at some profiles, TTC precludes longer trading cycles that could lead to more appealing allocations. We explore restricted preference domains as an avenue for finding alternative rules satisfying the three axioms. Our main contribution is to introduce a richness condition on domains, the top-two condition, and show that it characterizes TTC domains: domains to which the above characterization of TTC extends. The condition requires that, within every subset of objects, any two objects that can each be ranked first can also be ranked as the top two, in both possible orders. For every domain that fails the top-two condition, we construct a rule distinct from TTC that is individually rational, Pareto efficient, and strategyproof, and that executes longer trading cycles that TTC precludes.

econ.TH · cs.GT

新论文 · v1arXiv:2611.20616

GEAR$^2$: Pruned Experts Live On through Gated Reuse of Activated Experts and a Missing-Mass Residual

Linliang Chen, Yan Zhong, Wang Kang, Sai Li, Ji Wu, Ji Liu

原文摘要

Mixture-of-Experts (MoE) Large Language Models (LLMs) activate only a few experts per token while generally keeping all the experts in GPU memory for high efficiency. Expert pruning reduces this GPU memory footprint cost, yet discards the contribution of pruned experts. Existing remedies either rely on static substitutes, merge experts at the cost of altering their functions, or require costly recovery training. In this paper, we find that the functionalities of pruned experts can be efficiently reconstructed from the activated remaining experts. We propose GEAR (Gated Expert Activation Reuse), which reuses the scaled outputs of reliable activated experts without any training or additional expert computation. In addition, we propose GEAR$^2$ to further recover what reuse cannot express with an ultra-lightweight residual. On five MoE LLMs from 16B to 106B parameters with 25% and 50% of experts pruned, GEAR improves over its training-free counterparts in nearly every setting (up to 4 pp.) in terms of accuracy, and GEAR$^2$ outperforms baseline approaches (up to 2.75 pp.) in every setting on nine zero-shot benchmarks, with larger advantages (up to 2.72 pp. higher) under more aggressive pruning (up to 50%).

cs.AI

新论文 · v1arXiv:2611.20613

Noise-Aware Testing of NISQ Quantum Programs: Self-Baseline Oracle with a Controlled False-Positive Rate

Mir Mohammad Yousuf, Shabir Ahmad Sofi

原文摘要

Testing quantum programs on Noisy Intermediate-Scale Quantum (NISQ) devices faces a test-oracle problem: hardware noise and finite-shot sampling make the runs of a correct program vary as much as those of a faulty one, so oracles with a fixed acceptance threshold produce false positives that depend on the program and the backend. We present a noise-aware testing framework built on self-baseline calibration: each run of the program under test is compared with repeated executions of a trusted reference program on the same backend, and the acceptance threshold is the $(1-α)$ quantile of the finite-shot null distribution of the Jensen-Shannon divergence, so the oracle's false-positive rate is a level set in advance. A noncentral chi-square model predicts detection power, a homogeneity test and a two-run sentinel detect drift, a mutation engine labels every mutant exactly within a three-class noise-aware taxonomy, and zero-noise extrapolation attributes rejections to faults or noise. In simulation on nine core benchmarks, 32 QASMBench programs, ten real bugs, seven noise regimes and three device snapshots, the oracle's false-positive rate is 0.0011 at $α= 10^{-3}$ over 9,000 correct runs, against 36-74% for fixed thresholds, 8.9% for a QOIN-style filter and 10.8% for a globally calibrated threshold, while attaining the highest F1 on non-equivalent faults (0.941). On an IBM Heron device, using about 14 minutes of QPU time, it rejected none of 82 correct runs and detected 120 of 121 faults and semantic mutants, and a two-run sentinel reduced drift-induced false positives from 27.4% to 1.8%.

quant-ph · cs.SE

新论文 · v1arXiv:2611.20612

Learning Design Taste through Verifiable Synthetic Environments and Agentic Reinforcement Learning

Sujay choubey, Siddhant Saxena, Sarthak Kapil, Suhas Motwani

原文摘要

Code-generating language models produce user interfaces that build and run, but the pages are often weak in hierarchy, spacing, typography and composition. We study whether a small code model can learn design taste through reinforcement learning in verifiable synthetic environments. The method has three components. First, a layered verification framework splits design quality into hard gates, render-grounded metrics, rubric judges, within-group relative preference and a blinded expert panel. A signal enters the reward only if it passes an admission protocol that tests agreement with experts, causal sensitivity, resistance to hack traps and stability. Second, following SCALER [1], synthetic design environments are parameterised generators with measured difficulty ladders: base pages are written once by an agentic coder, and instances are produced by syntax-tree perturbation operators and constraint samplers, scheduled by an adaptive curriculum with a quarantine rule against reward over-optimisation. Third, a 4B code model is trained with agentic reinforcement learning over multi-turn episodes in which it edits and builds a repository and reads a textual report of its own render, while hidden graders compute the reward. The admission protocol admits 7 of 16 render signals and 6 of 10 rubric dimensions. On TD-Bench, agentic training raises the Taste Index from 0.630 to 0.795 in-family and from 0.598 to 0.700 on held-out families, and wins 0.600 of blinded expert comparisons against the base model and 0.569 against a single-turn control. Most of the advantage depends on the inference-time loop, and cross-brief distinctiveness falls slightly below that of the base model, the only success criterion not met.

cs.MA · cs.HC

版本更新 · v2arXiv:2611.15146

Reframing Video Inverse Tone Mapping as Blind Posterior Inference via Latent Luminance Supervision

Hanwool Sul, Chang-Hoon Jeong, Kisung Shin, Bomseok Park, Donghoon Cha, Hansl von Kwon, DoHyung Ku, Byoung-Tak Zhang

原文摘要

High-dynamic-range (HDR) displays are now common, but most video content, such as archival footage, broadcast material, and the output of generative video models, exists only in standard dynamic range (SDR). The need is sharpest for generated video, which has no HDR master at all and is graded alongside live-action and visual-effects material finished in HDR. Converting SDR to HDR means undoing a tone-mapping operation that was never recorded, and clipping destroys the radiance wherever it acts, so HDR recovery is a blind inverse problem. Methods built on generative video priors fill clipped regions with plausible structure, but their reconstructions are systematically biased: in every public method we test, saturated content leans too bright or too dark, and the direction of the lean moves with the tone operator behind the input. We call this failure the luminance error, measure both its magnitude and its sign, and make it the focus of this work. Reframing ITM as posterior inference under an unknown operator locates the failure in two places and yields two corrections. A fixed affine functional of the video-diffusion latent reads log-luminance without decoding; on that coordinate an asymmetric loss selects an expectile of the posterior, applied only where clipping leaves luminance unconstrained. We also synthesize training SDR through an explicit family of tone operators, so the model marginalizes over what the observation cannot show. Trained this way, our luminance error stays flatter across operators and smaller on average than any public method, and the full system achieves the best perceptual quality on all three benchmarks.

cs.CV · cs.AI

新论文 · v1arXiv:2611.20609

Newton-PINet: A fast physics-informed neural network with Newton linearization for meta-learning nonlinear PDEs

Yuchen Fan, Chang Wei, Pao-Hsiung Chiu, Chin Chun Ooi, Heyang Wang, Jian Cheng Wong

原文摘要

Scientific machine learning has opened new avenues for solving parameterized partial differential equations (PDEs), enabling models to learn a family of PDEs and generalize to unseen instances. In this context, data-driven operator learning methods typically require large training datasets, while physics-informed neural networks (PINNs) suffer from difficult optimization and limited generalization, especially for nonlinear PDEs. We propose Newton-PINet, a physics-informed network enhanced by Newton linearization, offering an effective meta-learning framework for nonlinear PDEs. Newton-PINet (i) employs a physics-informed multilayer network with skip connections, where the output-layer weights are solved by least squares; (ii) adopts a two-stage learning strategy that first leverages gradient-based training to learn robust representations from the available training tasks, and then performs gradient-free fine-tuning on the output layer for fast task-specific generalization; and (iii) incorporates a Newton linearization method to speed up the least-squares iteration for nonlinear PDE problems. On a challenging nonlinear reaction-diffusion benchmark, Newton-PINet achieves up to three orders of magnitude lower relative error than recent neural solvers, while using 16x fewer training tasks and over an order of magnitude less training time (about 5 minutes versus several hours). This work advances the meta-learning of PINNs toward data-efficient, fast, and generalizable physics solvers. The code and datasets are publicly available at: https://github.com/fycAIwj/Newton-PINet.

cs.LG · cs.AI · cs.CE · physics.comp-ph

版本更新 · v2arXiv:2602.20584

Long-Term Multi-Session 3D Reconstruction Under Substantial Appearance Change

Beverley Gorry, Tobias Fischer, Michael Milford, Alejandro Fontan

原文摘要

Long-term environmental monitoring requires the ability to reconstruct and align 3D models across repeated site visits separated by months or years to enable the detection of structural change and the assessment of ecosystem health. However, existing Structure-from-Motion (SfM) pipelines are generally designed for imagery captured within a limited time window and consequently fail in long-term monitoring scenarios such as repeated coral reef surveys with substantial visual and structural change. In this paper, we show that the primary limitation of current approaches lies in their inability to maintain visual correspondences under substantial temporal appearance change. We address this limitation by enforcing cross-session correspondences within a joint SfM reconstruction. Our approach combines handcrafted and learned visual features to establish correspondences across large temporal gaps, enabling the reconstruction of a single coherent 3D model from imagery captured years apart, where standard SfM fails to connect sessions. To ensure scalability and further improve alignment robustness on large datasets, we restrict expensive learned feature matching to a small set of likely cross-session image pairs identified via visual place recognition, which reduces computational cost while improving the reliability of cross-session correspondence establishment. We evaluate our method on long-term underwater datasets exhibiting significant real-world change, and demonstrate consistent joint reconstruction across sessions in cases where existing methods fail to connect the sessions or align them only coarsely.

cs.CV · cs.RO

新论文 · v1arXiv:2611.20607

DNA: A Double-Helix Hierarchical Deep Interaction Network Architecture for Industrial Digital Twin

Haoran Yang, Yanan Xu, Yinan Zhang, Wenjie Zhang, Peiyu Liu, Jiahui Xiang, Wenhai Wang

原文摘要

High-fidelity modeling is fundamental to industrial digital twin, yet a trade-off persists between the physical transparency of mechanistic models and the superior predictive power of data-driven approaches. Existing hybrid methods often fail to bridge this gap, suffering from shallow integration and rigid one-shot modeling strategies that struggle with complex industrial systems. To address these issues, we propose a \textbf{Double-Helix Hierarchical Deep Interaction Network Architecture (DNA)}. Inspired by the double-helix structure of biological \textit{DNA}, our architecture intertwines mechanistic and data-driven modeling strands through two novel interaction strategies, forming a symbiotic architecture in which the two strands mutually enhance predictive capability. Furthermore, the hierarchical structure of DNA allows it to model complex systems in a multi-stage manner through two tailored deep learning frameworks. We evaluate the proposed approach on two distinct industrial systems, where DNA achieves reductions in MAE of 3.704\% and 3.175\%, respectively, compared to the strongest baseline. These results demonstrate the effectiveness of DNA and suggest its potential to provide new insights for the development of industrial digital twins.

cs.LG

新论文 · v1arXiv:2611.20606

Necessary for What? Auditing Demographic Disclosure to LLM Synthetic Respondents

Aspen Wang, Yen Yu Shan, Lun-Wei Ku

原文摘要

Large language models (LLMs) increasingly answer surveys on behalf of real panelists, and each panelist's recorded demographic attributes (e.g., age, education, ethnicity) are often placed in the prompt on the premise that richer profiles make more faithful synthetic respondents. While data minimization limits personal data to what a purpose requires, a synthetic survey serves several purposes at once, from predicting individual answers to estimating how opinions are distributed and comparing groups. We ask which of these purposes require the respondent's record, and whether improved estimates support sound conclusions about groups. We vary whose profile a model receives, where it comes from, and how many fields it contains, in two U.S. surveys with eleven open-weight and three hosted models, and compare each model with a statistical model fit to the same information. We make four contributions. First, we show that the record adds little to individual predictions and that, for the open-weight models, most of its gain in population estimates remains when it is replaced by census data or a synthetic persona. Second, we find that group comparisons improve only with the respondent's own record, and that even then the models' group differences rest more on the demographic label in the prompt than on the individual respondent, which the usual check of reproducing known group differences does not reveal. Third, we find that the models make groups look more alike than they are, keeping differences in circumstances while flattening or reversing differences in attitudes, most sharply for religious affiliation and ethnic origin, both special categories of personal data. Fourth, we outline an audit that a survey organization can run on questions it has already fielded to decide which estimates may use the record.

cs.CY · cs.CL

新论文 · v1arXiv:2611.20605

AdMem: Expectation-Grounded Procedural Memory for Long-Horizon Task-Solving Agents

Runzhe Wang, Huilin Lu, Shengjie Liu, Li Dong, Zhihuai Zhu

原文摘要

An agent that carries memory across tasks is only as good as the entries it keeps. We begin from an observation that is easy to reproduce and uncomfortable for the field: naively accumulating cross-task experience can make an agent \emph{worse} than one with no memory at all. On AgentBoard, a workflow-induction memory solves fewer tasks than a memoryless ReAct agent in seven of eight domains, and a per-step procedural store with no surrounding structure loses half of ReAct's completions on Jericho. We argue that the missing ingredients are per-action credit assignment and an explicit life cycle for stored entries, and present AdMem, which supplies both. Before each action the actor commits to an expected outcome; a concurrently running critic scores the realized outcome against that expectation and writes a procedural entry carrying a binary reward and a reflection, so credit lands on the action that earned it and no environment reward is required. Each entry then carries a scalar utility, updated online under a noisy-OR reward model, that ranks retrieval jointly with context similarity and drives eviction and merging, giving the store a forgetting and consolidation policy instead of unbounded growth. Across 707 tasks in eight AgentBoard domains AdMem solves 441 against 388 for ReAct and 312 for the workflow-induction baseline --- a pooled gain significant under a two-sided Fisher exact test, concentrated in the domains that supply transferable cross-task structure. All memory stays human-readable text, which is what makes the failures above diagnosable --- and, as we discuss, also makes the store an auditable but attackable surface.

cs.AI

新论文 · v1arXiv:2611.20603

BiGSyn: A Bi-Level Heterogeneous Graph Contrastive Learning Framework for Interpretable Drug Synergy Prediction

Senquan Wang, Hui Fang, Haoran Nan, Xingwei Wang, Qiang He

原文摘要

Drug combination therapy has become a mainstream of modern medicine, with the potential to enhance efficacy, reduce toxicity, and overcome drug resistance. However, identifying synergistic drug combinations remains challenging, as it requires modeling complex interactions across multiple biological scales---from atom-level molecular structures to entity-level relations in biological networks. Existing approaches typically model these levels in isolation, overlooking complementary signals and long-range dependencies. To address these limitations, we propose BiGSyn, a Bi-level heterogeneous Graph contrastive learning framework for drug Synergy prediction. Specifically, BiGSyn leverages heterogeneous graph transformers to extract high-order dependencies from molecular graphs and biological networks, and introduces a bi-level contrastive learning objective to bridge the cross-level semantic gap. Extensive experiments on two benchmark datasets show that BiGSyn consistently outperforms state-of-the-art baselines. Moreover, its attention mechanism highlights key molecular substructures and biological interactions driving synergy, providing interpretable insights.

cs.AI

版本更新 · v2arXiv:2610.33433

Examining form-based linguistic productivity in people and large language models

Michael Wilson, Robert Frank

原文摘要

Linguistic productivity, as evidenced by the use of words in novel contexts, is one of the most striking facts about human language. However, how abstract the knowledge that enables such productivity is remains unclear. We compare people's and computational language models' linguistic productivity in contexts where successful generalization depends on form-based information. We focus on extending the use of a novel verb learned in the active voice to passive contexts. We find people extend active to passive uses in a way consistent with abstract generalizations concerning linguistic form, even when no plausible meaning can be assigned to the novel verb. In contrast, large language models succeeded only when a meaning like an existing verb's could be assigned to the novel verb. We conclude that computational models' syntactic knowledge remains tied to the distribution and meaning of specific lexical items in a way that human syntactic knowledge is not.

cs.CL

新论文 · v1arXiv:2611.20595

Invariant-Aware Harness Enforcement to Protect LLM Agents from Silent API Failures

Sajib Acharjee Dip, Brian Ouyang, Lala Motlhabi

原文摘要

LLM agents answer questions by calling APIs and trusting the results. An API can fail silently. It ignores part of a request, such as a misnamed filter, yet returns HTTP 200 and a plausible but wrong result. Current fixes, such as checklists, reviewer models and LLM judges, ask the model to catch it. But models often see the problem and answer anyway. We present SilentGuard, which moves the decision from the model to the harness running tool calls. It tests each result against invariants that any correct API must satisfy and withholds values that break one. We find silent failures in 11 of 15 public research APIs, with counts inflated up to 691 times. Current fixes leave over 70% of answers wrong, and agents report 77.4% of counts a judge flags. Across 7 open models, SilentGuard cuts wrong answers on its target faults from 74.0% to 1.7%, with no loss on fault-free tasks. On three frontier models they fall from 70.2% to 0.5%. A warning from the same check leaves 70.1% wrong, so enforcement is the key. On new fault types, the harness alone cuts wrong answers from 61.1% to 24.5%. It avoids wrong answers by abstaining rather than recovering.

cs.CL

版本更新 · v2arXiv:2601.21177

Flow Perturbation++: Multi-Step Unbiased Jacobian Estimation for High-Dimensional Boltzmann Sampling

Xin Peng, Ang Gao

原文摘要

The scalability of continuous normalizing flows (CNFs) for unbiased Boltzmann sampling remains limited in high-dimensional systems due to the cost of Jacobian-determinant evaluation, which requires $D$ backpropagation passes through the flow layers. Existing stochastic Jacobian estimators such as the Hutchinson trace estimator reduce computation but introduce bias, while the recently proposed Flow Perturbation method is unbiased yet suffers from high variance. We present \textbf{Flow Perturbation++}, a variance-reduced extension of Flow Perturbation that discretizes the probability-flow ODE and performs unbiased stepwise Jacobian estimation at each integration step. This multi-step construction retains the unbiasedness of Flow Perturbation while achieves substantially lower estimator variance. Integrated into a Sequential Monte Carlo framework, Flow Perturbation++ achieves significantly improved equilibrium sampling on a 1000D Gaussian Mixture Model and the all-atom Chignolin protein compared with Hutchinson-based and single-step Flow Perturbation baselines.

cs.LG

版本更新 · v2arXiv:2611.12961

RBLA+: Efficient Rank-Specific LoRA Adapter Extraction for Federated Fine-Tuning with Heterogeneous Clients

Shuaijun Chen, Omid Tavallaie, Xin Chen, Wentao Gao, Nguyen H. Tran, Albert Y. Zomaya

原文摘要

Federated fine-tuning with low-rank adaptation (LoRA) reduces communication and computational overhead by training and exchanging compact trainable LoRA factors. Clients may use different LoRA ranks to meet adapter communication budgets and accommodate different local adaptation needs. The server must process these differences during aggregation and construct adapters at each client's rank for local training. We distinguish pre-aggregation alignment, which handles factor dimensions and scaling conventions before aggregation and rank-specific LoRA adapter extraction, which constructs client-sized pairs from the aggregated global model. Two common approaches to rank-specific extraction are direct rank truncation and Singular Value Decomposition (SVD) of the dense update. Direct prefix truncation keeps the leading columns of one aggregated factor and the corresponding rows of the other. It is cheap but can increase reconstruction error. The dense SVD approach first forms the product of aggregated LoRA factors. It then decomposes this matrix and retains the leading singular components to construct adapters at the target rank. This gives the optimal rank-constrained approximation in Frobenius norm but incurs significant computational overhead. To address this trade-off, we propose RBLA+ as an extension of RBLA. It uses established compact QR/core-SVD decomposition to minimize Frobenius error for the extracted LoRA update without forming the dense update. It also accounts for client and server scaling conventions during upload and broadcast. Across vision and language tasks, RBLA+ improves accuracy over RBLA in multiple heterogeneous settings. In CPU benchmarks with synthetic adapters, RBLA+ has lower extraction time than the dense partial-SVD baseline. RBLA+ preserves each client's prescribed adapter size and per-round communication payload.

cs.AI · cs.DC

新论文 · v1arXiv:2611.21688

DeltaHarness: Self-Evolving Harnesses for Resource-Adaptive Tool Execution

Lecheng Yan, Jianze Lin, Xicheng Han, Xiantao Xu, Wenxi Li, Chenyang Lyu

原文摘要

Tool-using agents rely on execution harnesses to organize and schedule tool operations. The same task and resource description can conceal runtime conditions that reverse the latency ordering of valid execution plans. Resolving this uncertainty requires firsthand execution in the target setting. We introduce DeltaHarness, an execution-guided policy-search method that adapts agent harnesses from cold-start measurements. Its QHI loop uses the observed response to update runtime dispatch, revise physical plans, and select experiments for later proposals. An independent confirmation panel controls policy replacement. With fixed dispatch, a revised materialization policy reduces complete-delivery latency by 30.0% on one video-editing task across three fresh pairs. On held-out tasks, the evolved policies reduce latency by 8.2% relative to the best fixed plan in one seed and by 4.6% relative to the initial proposal in another.

cs.LG · cs.PF

新论文 · v1arXiv:2611.20593

JEPTERA: A Platform Agnostic JEPA-Assisted Model Predictive Planner for Urban UAV Navigation

Gibran Fathoni Belva, Mansur Arief, Ahmad Alfan Alfian Irfan

原文摘要

Urban navigation for Unmanned Aerial Vehicles (UAVs) requires planning that accounts for environmental geometry, vehicle dynamics, and moving obstacles. This paper presents JEPTERA, a navigation architecture that integrates a Joint-Embedding Predictive Architecture (JEPA), Model Predictive Control (MPC), PID control, and a safety layer. JEPA predicts the UAV's motion response to candidate commands for MPC action selection, while the safety layer checks trajectories against occupied space and predicted obstacle motion. Simulator adapters and platform-specific YAML configurations separate the planning core from simulator interfaces. An ablation study evaluated six configurations across three gym scenarios, totaling 180 runs. The complete architecture achieved the highest aggregate success rate, completing 28 of 30 runs. Additional tests completed one route across three city maps without collisions. Separate integration tests demonstrated operation in Gazebo/ROS 2, MuJoCo, and CARLA without modifying the core planning logic. These results support the feasibility of the integrated architecture for simulated urban UAV navigation and its portability across simulator interfaces.

cs.RO

新论文 · v1arXiv:2611.20587

Stability Principle Inherent in Wheel Gait of Planar X-shaped Walker Generated Using Constant Torque Drive and Mechanical Stoppers

Fumihiko Asano, Cong Yan

原文摘要

Since the late 19th century when the first walking toys were developed, it has been known that mechanical stoppers at the hip joint are crucial for generating stable passive dynamic walking. Recent research on passive-dynamic and limit-cycle walkers has also confirmed that mechanical stoppers at the hip and knee joints are effective for generating stable walking motion, but theoretical research on how this mechanical constraint enhances the overall gait stability remains insufficient. This paper introduces a planar X-shaped walker equipped with mechanical stoppers at the hip joint, and investigates the effect of the mechanical constraint on the stability of the wheel gait generated by constant torque drive on a downslope. By simply falling forward while using the stoppers to constrain itself to the target impact posture, the robot can generate a highly stable wheel gait. We divide the motion of one step into four phases, derive approximate analytical solutions for the state error transition function matrix in each phase using a linearized model, and analyze the increase or decrease in the state error norm using metrics such as its maximum singular value. Numerical simulations demonstrate that while two phases are unstable in terms of the increase in the state error norm, the remaining two phases are stable, resulting in overall asymptotic stability.

cs.RO

新论文 · v1arXiv:2611.20584

A Self-Evolving Agent Runs Like a Gas: Describing and Predicting Its Evolution with Fluid Dynamics

Haipeng Wang

原文摘要

A self-evolving LLM agent improves in service by rewriting its own plugin library under a validation gate. Agent engineering is converging on exactly such agents: single systems in continuous operation, carrying populations of runtime instances (plugins, skills, tools) past the point where anyone can read the library and still predict what the system will do. Yet the run itself, the aggregate motion of that population across hundreds of edit rounds, is unstudied: no theory says whether it relaxes, oscillates, or freezes. A recent kinetic theory supplies the missing level: under four architectural properties, the plugin instances of one agent behave as identical hard spheres whose density obeys a master equation, and the theory is candid that its gate injects and drains credit, momentum is not conserved, and waves are excluded by construction. We argue the boundary is architectural; keeping the four safety theorems, we stack conservation axioms on top. A closed-credit ledger makes credit a collision invariant and unlocks acoustic waves at a measurable, engineerable speed; a zero-sum projection gate conserves centroid momentum unconditionally, for any proposal stream. Both layers are proven and executed: the axis total is conserved to the digit over 66 interventions under a live gate, and the pre-registered decisive probe returns the theory's directional verdict: linear wavefronts where the ledger holds, dissipation where it does not, replicated across two population sizes. For agents that must never stop, the gas description is not a metaphor: it is the measurement interface through which their future becomes readable.

cs.GR

新论文 · v1arXiv:2611.20583

State-Machine Governance of Adaptive Enforcement: Evidence from a Paired-Scenario Evaluation

Ifeanyi Williams Agbedo, Yanzhen Qu

原文摘要

AI-assisted security systems can recommend rapid policy changes, but governing their execution requires explicit authorization, evidence, and recovery rules. This study evaluated a machine-verifiable, state-machine-governed transition procedure implemented in ZTGuard using 36 matched scripted scenarios and deterministic vendor-connector simulations. Each pair held the starting state, telemetry, AI recommendation, and scheduled vendor condition constant. The sole dependent variable, Control, was the equal-weight mean of traceability, accountability, audit-chain completeness, and reliable rollback capability. Mean Control was 0.720 (SD = 0.218) for the scripted operator-driven baseline and 0.972 (SD = 0.131) for the governed procedure. A two-tailed paired Wilcoxon signed-rank test yielded V = 359, p < .001, with a matched-pairs rank-biserial correlation of .899. Twenty-six pairs favored treatment, nine were tied, and one favored baseline. The findings support controlled artifact utility, not production readiness. Interpretation is limited by purposive sampling, scripted operators, simulated connectors, overlapping components, and arm-specific scoring rules.

cs.CR

版本更新 · v2arXiv:2607.07984

Agentic Neural Architecture Search

Seokhoon Jeong, Mijung Kim, Taehwan Kim

原文摘要

Neural architecture search (NAS) methods have grown increasingly efficient, but they still operate within manually designed search spaces that require domain expertise and must be rebuilt for each new task. Large language models (LLMs) can write architectures in an open-ended code space, yet it is unclear how the work should be divided between LLM-driven design and NAS-driven search. We propose a way to combine the two. An LLM first designs a seed architecture and then decomposes it into a slotted architecture: a scaffold with named, interchangeable module slots that defines a bounded, task-specific search space, which a conventional NAS algorithm then explores without any manual engineering. We implement this in AgentNAS, a three-phase pipeline whose phases can be evaluated separately. On 17 tasks from NAS-Bench-360 and Unseen NAS, covering classification, dense regression, segmentation, and multi-label tagging, AgentNAS outperforms published baselines, including task-specific expert designs, on 11 tasks. Extensive ablation studies show that LLMs find strong architectures quickly but soon plateau, and that switching to NAS at this point yields further gains, indicating that the two approaches are complementary. Experiments with five LLMs show that seed architecture design is effective across all of them, whereas constructing a usable search space requires a minimum level of model capability. Our code is available at https://github.com/alroimfebruary/AgentNAS.

cs.AI

新论文 · v1arXiv:2611.20579

Language Models Meet Vibration Testing for LEO Constellation Satellites with Calibrated Response Prediction from Few-show Sweeps

Jun Wu, Deming Li, Zhenming Zheng, Honghao Yue, Fei Yang, Hao Sun

原文摘要

Large language models (LLMs) increasingly drive engineering analysis end to end, turning textual records into the inputs of numerical programs, but a plausible and wrong value can pass silently into a decision. We propose deterministic validation gates for LLM-driven engineering analysis, a general paradigm in which a typed configuration produced by LLMs is consumed only after it passes deterministic checks on units, physical consistency, verbatim quotes, agreement between independent models, consistency with measurement exports and the applicability of trained models. The gates target hallucinations from two directions, since quote checking catches fabricated content while agreement between models catches errors of omission that quote checking misses. We instantiate the paradigm on sine vibration acceptance testing for low Earth orbit mega-constellation satellites, where DeepSeek V4.1 Flash and Qwen-plus read Chinese test records and extract a configuration of 76 fields. Fact-level F1 ranged from 0.98 to 1.0 on five input variants, with all repairs frozen on development inputs and evaluated on held-out inputs. With all gates, 20 of 20 configurations produced notching results bit for bit identical to the gold run, and all 17 real and injected errors in consumed fields were stopped, of which 16 would have run without gates. Retraining the bound model on a dataset built from an LLM configuration reproduced its results within GPU nondeterminism. We also report a negative result of practical value. Direct LLM curve forecasting only matches linear extrapolation, whereas a small physics-structured network reduces the over-notched bandwidth by 86% to 89% at calibrated coverage c = 0.90, which delineates where direct LLM prediction should give way to a trained model.

cs.CL · cs.AI · cs.LG

新论文 · v1arXiv:2611.20576

Partial Borrows in Rust

Sasha Pak, Richard Willie, Umang Mathur, Fabian Muehlboeck, Alex Potanin

原文摘要

Rust ensures memory safety by requiring exclusive access to mutably borrowed data. However, a borrow can prevent access to more data than necessary. For example, a method with an &mut self parameter borrows the entire struct, even when the method uses only some of its fields. Calling the method can therefore conflict with an existing borrow of a field the method never uses, causing Rust to reject the program. Resolving such conflicts can require programmers to compromise useful abstractions, make non-local changes to their programs, or worse, resort to unsafe code. We introduce view types, a backward-compatible refinement of reference types that specifies which fields a reference may access. This allows methods to partially borrow a struct, accessing only the fields they need without blocking access to the rest. We establish the first fully mechanized foundation for Aeneas in Rocq. Along the way, we uncover and fix gaps in the original semantics and proofs. Building on this foundation, we formalize partial borrows and prove the extended borrow checking sound. We implement view types in rustc and extend the Aeneas interpreter to support partial borrows. Our evaluation shows how view types systematically resolve borrowing conflicts across method boundaries at no runtime cost.

cs.PL

版本更新 · v2arXiv:2605.09446

Astro Generative Network: A Variational Framework for Controlled Node Insertion in Incomplete Complex Networks

Mehrdad Jalali, Binh Vu, Swati Chandna, Chen Ding

原文摘要

Empirical networked systems are often only partially observed: sampling frames, crawling policies, privacy constraints, and temporal gaps can leave actors and edges unobserved. This complicates robustness and sensitivity analysis because many graph-learning pipelines implicitly treat the observed node set as exhaustive. Link prediction and graph completion repair structure among known vertices, whereas full-graph generators synthesize new graphs rather than extending an observed one as a fixed backbone. We study the complementary task of controlled node insertion: generating plausible new actors and attaching them to an existing graph while preserving interpretable global topology. We introduce the Astro Generative Network (AGN), a variational graph autoencoder that samples latent vectors to decode node features and then integrates new vertices through similarity-based attachment to the observed backbone. We distinguish the recommended configuration, AGN, from AGN-original, a diagnostic baseline that permits generated-generated edges. Across three synthetic regimes, AGN-original forms dense generated-generated subgraphs that artificially inflate clustering and density. Disabling those edges removes this artifact while preserving degree and path-length behavior. In our experiments, AGN keeps clustering and modularity changes modest relative to pre-insertion values, while novelty diagnostics show non-trivial separation from existing nodes without claiming domain-grounded identities. Our contribution is methodological: a reproducible insertion protocol and evaluation lens for incomplete network science and engineering

cs.SI

版本更新 · v4arXiv:2611.12657

When Does Sliced Inference Pay Off? Resolution, Tiling and Dense Fusion for Small Object Detection

Stefan Becker, Jens Bayer, David Münch, Wolfgang Hübner

原文摘要

Sliced inference such as SAHI is the standard remedy for small objects in high-resolution imagery, but its gains are usually measured against low-resolution baselines. We revisit sliced inference with scale as the explicit variable and ask when tiling actually pays off. We compare nine dense detectors, eight YOLO models of four generations and FCOS, on VisDrone and DOTA, under pre-specified hypotheses, calibration-only tuning and a single evaluation on held-out test data. Two findings emerge. First, much of the benefit attributed to slicing is a resolution effect: on VisDrone, a single pass trained at a higher resolution outperforms tiling with box fusion (merging decoded boxes) for every detector that could be trained at that resolution, and at equal compute a higher resolution helps more than a larger model. Second, tiles carry complementary information that box fusion wastes. We introduce SCAHI (Slicing with Coverage-Aware Hyper Inference), which accumulates the dense outputs of all passes with weights that give every image position the same total coverage, and reads objects off a peak grid fine enough to keep crowded neighbours apart. On identical passes, SCAHI outperforms box fusion for every YOLO detector on both datasets and matches it for FCOS. On DOTA, whose images far exceed the resolution a single pass can be trained at, tiles with SCAHI beat the best single pass for every detector, by 4.3 to 5.6 mAP50 at 2048 px, and replacing only SAHI's merge step by SCAHI adds 2.3 to 4.0 mAP50. An ablation attributes the gain to the border weighting and the resolution of the peak grid. Source code is available at https://github.com/beckerio/scahi.

cs.CV

新论文 · v1arXiv:2611.20570

OrthoMark: Place Recognition with Point-Anchored Orthographic Height Mark

Haichuan Li

原文摘要

LiDAR place recognition provides the loop closures of Simultaneous Localization and Mapping(SLAM). Perspective from Z-axis, a gravity-aligned scan is an orthographic map of the scene, and a revisit from another position and heading moves this map rigidly in the plane, a motion that a polar Fourier spectrum ignores and phase correlation recovers. Along the vertical, the heights of the returns describe the structures, but which heights are observed depends on the range, because the sensor's vertical field of view cuts nearby structures low; height thus misleads a global key, yet is decisive once two scans are aligned and compared cell by cell where they overlap. We present \method, a training-free method built on one principle: two LiDAR scans can only be compared through features computed from the same geometric surfaces in both scans. \method retrieves candidates with a translation- and rotation-invariant polar spectrum of the bird's-eye view (BEV), 1.9,KB per keyframe, and verifies each candidate by estimating its planar pose through phase correlation and scoring it with \emph{point-anchored height columns}: every BEV cell keeps a 16-bit mask of the heights of its own returns; within a session, the score is weighted by the expected overlap of the two scans' sensing ranges, computed from their distance. We show that this vertical link pays off in verification but hurts in retrieval, where what a height slice contains depends on the range to the structure. For single robot, \method has the highest AP of the training-free methods without 3D registration on the KITTI development sequences; across robots, on all multi-robot sequences, it matches or exceeds RING++ at 1/350 of its descriptor size and closes more inter-robot loops than the learned BEVPlace++ on all tests.

cs.RO · cs.CV

新论文 · v1arXiv:2611.20569

ToolGraph as a Q-Function for Test-Time Scaling of Tool-Use Agents

Ke Shen, Ankita Bhaumik, Li Dong, Shengjie Liu, Shuzhen Li, Naveed Ahmed Janvekar

原文摘要

Tool-use agents must chain interdependent calls, yet which outputs satisfy which later inputs in a deployed catalog is not something a general-purpose language model reliably infers from tool descriptions alone. We compile that structure once into a reusable external representation. ToolGraph is a directed, weighted dependency graph over a tool catalog: parameter matching and interventional probes, which execute a candidate chain with and without its proposed predecessor, propose edges whose evidence is combined into an edge confidence. Prior work uses such graphs to retrieve tools, search for tool paths, or score the next call. We ask instead whether the same structure can supply a \emph{multi-step} value by composing graph paths into a goal-conditioned action value \(Q(s,a,g)\). We study three estimators (dynamic programming over entity sets, a GraphSAGE estimator distilled from it, and an LLM scorer conditioned on graph evidence) plus a verbalizer that returns action--goal feedback to the agent, which keeps final action selection. Across two tool-use benchmarks and three backbones, we compare value ranking against personalized-PageRank retrieval over the identical graph, and against direct and reflective agents. Composition pays off only when the estimator also reads the concrete state: the LLM estimator improves on PageRank in every benchmark--backbone setting, by \(8.3\) points on average, while the two purely structural estimators fall below it, because a value defined over tool types cannot separate a correct tool called with a wrong argument. Graph construction is a one-time cost; we report it separately rather than assume it amortized away (at our scale it is comparable to request-time cost) and sweep four knobs controlling how much graph evidence to expose against success and tokens.

cs.AI

新论文 · v1arXiv:2611.20568

sqrtVINS-Based 360 Visual-Inertial SLAM Systems: Hilti-Trimble SLAM Challenge 2026 Entry

Jungwoo Lee, Hogyun Kim, Jiwon Choi, Jinyong Jeong, Younggun Cho

原文摘要

This report describes our submission to the 2026 Hilti-Trimble SLAM Challenge. The system is a 360-camera visual-inertial pipeline that uses only the provided dual-fisheye image streams and IMU measurements. First, raw images are preprocessed with low-light enhancement and static/dynamic masking, and then fed into a $\sqrt{\text{VINS}}$ frontend as the local odometry, with deep learning-based SuperPoint features, ELSED line segments, and LightGlue matching. The resulting trajectory is refined using a GTSAM-based backend that performs tightly-coupled factor graph optimization with SALAD global descriptor and visual bundle adjustment (BA). The backend performs local-to-global structureless BA, tightly regularized intrinsic and extrinsic refinement, and global scale adjustment supported by the known inter-camera baseline and cross-view constraints. The design targets long indoor sequences with dark regions, repetitive corridors, dynamic objects, and limited geometric parallax, while keeping the pipeline reproducible in the provided ROS2 workspace. Fig. 1 demonstrates that our proposed method aligns well with the floorplan across several indoor sequences.

cs.RO