2026-08论文中文摘要
核查问题：受监管企业在AI产品交付前必须满足什么条件
The Checking Problem: What must be true before AI ships in a regulated firm
arXiv 2607.28666 · 2026-08-03GRU、LSTM与Transformer编码器在自动驾驶系统分类中的敏感性分析
Sensitivity Analysis of GRU, LSTM and Transformer Encoder in Classification of Automated Driving Systems
arXiv 2607.28665 · 2026-08-03矩阵的永久秩的加法保持算子
Additive preservers of permanent rank
arXiv 2607.28664 · 2026-08-03NeuroSynth：一种受生物启发的持续强化学习架构，用于缓解灾难性遗忘
NeuroSynth: A Biologically Inspired Continual Reinforcement Learning Architecture for Mitigating Catastrophic Forgetting
arXiv 2607.28663 · 2026-08-03面向异构文档知识图谱构建的本体引导、去重感知抽取层
An Ontology-Guided, Deduplication-Aware Extraction Layer for Knowledge Graph Construction from Heterogeneous Documents
arXiv 2607.28662 · 2026-08-03大型语言模型（LLMs）的财务推理是否可信？针对长期财务报表的现实测试
Are the Financial Reasoning from LLMs Credible? A Real World Test over Long-Horizon Statements
arXiv 2607.28661 · 2026-08-03无穷维哈密顿流形上的熵几何与归一化均值
Entropy Geometry and Normalized Means on Infinite-Dimensional Hamiltonian Manifolds
arXiv 2607.28660 · 2026-08-03基于混合分词与串行-并行解码的跨域序列推荐增强
Empowering Cross-Domain Sequential Recommendation with Hybrid Tokenization and Serial-Parallel Decoding
arXiv 2607.28659 · 2026-08-03联邦预训练评估：下游微调与内在评估的可靠性研究
Evaluating Federated Pre-Training: On the Reliability of Downstream Fine-Tuning and Intrinsic Evaluation
arXiv 2607.28658 · 2026-08-03TAPR：利用任务感知提示重写器提升大语言模型性能
TAPR: Enhancing LLM Performance with a Task-Aware Prompt Rewriter
arXiv 2607.28657 · 2026-08-03用于快速小区域慢性病估计的地理加权代理模型
Geographically Weighted Surrogate Models for Rapid Small-Area Chronic Disease Estimation
arXiv 2607.28655 · 2026-08-03HenTwin：用于蛋鸡纵向生物状态监测的多模态数字孪生框架
HenTwin: A Multimodal Digital Twin Framework for Longitudinal Biological State Monitoring in Laying Hens
arXiv 2607.28652 · 2026-08-03基于扩展ICAP框架的协作对话认知投入度测量：人类标注、上下文学习与反思型大语言模型智能体的比较
Measuring Cognitive Engagement in Collaborative Discourse with an Extended ICAP Framework: Comparing Human Annotation, In-Context Learning, and Reflective LLM Agents
arXiv 2607.28651 · 2026-08-03COSI-Lab：用于建模多视角多模态社会意图的会议生活实验室
COSI-Lab: Conference Living Lab for Modeling Multi-Perspective Multimodal Social Intention
arXiv 2607.28649 · 2026-08-03ConnectED：面向越南教学课程规划与学生学习的课程对齐AI系统
ConnectED: A Curriculum-Aligned AI System for Vietnamese Instructional Lesson Planning and Student Learning
arXiv 2607.28647 · 2026-08-03不同视角的观察：基于四世界框架的计算创造性中的解释视角建模
Seeing Differently: Modeling Interpretive Perspectives in Computational Creativity using a Four-World Framework
arXiv 2607.28644 · 2026-08-03是否促进：在线讨论中的人类与大语言模型（LLM）促进者倾向
To Facilitate or not to Facilitate: Human and LLM Facilitator Tendencies in Online Discussions
arXiv 2607.28643 · 2026-08-03ThinkReset：面向有限上下文长程推理的可学习中间接口构建
ThinkReset: Learnable Intermediate Interface Construction for Bounded-Context Long-Horizon Reasoning
arXiv 2607.28642 · 2026-08-03形式主义陷阱：社会负荷下作为评判者的大语言模型是否会被共识模仿蒙蔽？
The Formalism Trap: Are LLM-as-a-Judge Evaluators Blinded by Consensus Mimicry under Social Load?
arXiv 2607.28641 · 2026-08-03TokenSwap：多模态大语言模型中模态差距的基准测试与缩减
TokenSwap: Benchmarking and Reducing the Modality Gap in Multimodal LLMs
arXiv 2607.28640 · 2026-08-03从经验中学习有状态的预测知识
Learning Stateful Predictive Knowledge From Experience
arXiv 2607.28638 · 2026-08-03ZeroR@CHiPSAL 2026：用于尼泊尔表情包分类的对比学习两阶段视觉-语言适配
ZeroR@CHiPSAL 2026: Two-Stage Vision-Language Adaptation with Contrastive Learning for Nepali Meme Classification
arXiv 2607.28637 · 2026-08-03模型链：面向偏见鲁棒性大语言模型评判器的跨模型审计
Chain-of-Models: Cross-Model Auditing for Bias-Robust LLM Judges
arXiv 2607.28636 · 2026-08-03基于高斯混合模型（GMM）与大语言模型（LLM）的定向数据增强用于不平衡数据聚类
Imbalanced Data Clustering via Targeted Data Augmentation Using GMM and LLM
arXiv 2607.28635 · 2026-08-03大语言模型真的能理解题目难度等级吗？对使用大语言模型进行自动题目生成的启示
Can LLMs Really Understand Item Difficulty Levels? Implications for Automated Item Generation Using LLMs
arXiv 2607.28634 · 2026-08-03用于发现重要数学猜想的大语言模型框架：AI对下一个黎曼猜想的探索
LLM Framework for Discovering Major Mathematical Conjectures: AI's Quest for the Next Riemann Hypothesis
arXiv 2607.28632 · 2026-08-03AI能否评估AI科学家？基于自动化多模型评审的自主研究生成系统基准研究
Can AI Evaluate AI Scientists? A Benchmarking Study of Autonomous Research Generation Systems Using Automated Multi-Model Review
arXiv 2607.28631 · 2026-08-03为生成式AI（GenAI）搭建批判性参与支架：在提示工程任务中转变少数民族预科生的协作话语
Scaffolding Critical Engagement with GenAI: Transforming Ethnic Minority Preparatory Students' Collaborative Discourse in Prompt Engineering Tasks
arXiv 2607.28630 · 2026-08-03智能体AI中的OpenClaw与Ollama：迈向完全自主且可扩展的AI智能体系统
OpenClaw and Ollama in Agentic AI: Toward Fully Autonomous and Scalable AI Agent Systems
arXiv 2607.28629 · 2026-08-03泡利编码与不可克隆加密
Pauli Encodings & Unclonable Encryption
arXiv 2607.28602 · 2026-08-03基于稳定块输入输出反馈结构的两参数稳定控制器实现方法
On the Implementation of Two-Parameter Stabilizing Controllers Using Stable-Block Input-Output Feedback Structures
arXiv 2607.28469 · 2026-08-03SVR：基于联合判决-置信度强化学习的自适应测试时计算自验证优化
SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute
arXiv 2607.28457 · 2026-08-03超越单一评判者：用于生成式UI评估的社会角色面板模拟
Beyond a Single Judge: The Evidence-Grounded, Social-Weighted Persona Panel for Generative UI Evaluation
arXiv 2607.28439 · 2026-08-03揭秘Solana机器人：从GitHub蓝图到链上指纹
Demystifying Solana Bots: From GitHub Blueprints to On-Chain Fingerprints
arXiv 2607.28424 · 2026-08-03QAdapt：用于量子纠错的噪声自适应神经预解码框架
QAdapt: A Noise-Adaptive Neural Pre-Decoding Framework for Quantum Error Correction
arXiv 2607.28422 · 2026-08-033-Sasakian轨形商空间上的SU(3)-结构
SU(3)-structures on quotients of 3-Sasakian orbifolds
arXiv 2607.28337 · 2026-08-03引力场宇宙学探测中的动力学引力编织（KGB）特征
Signatures of kinetic gravity braiding in cosmological probes of the gravitational field
arXiv 2607.28207 · 2026-08-03光前方法研究四维无质量高自旋相互作用
Light-Front approach to $4d$ massless Higher-Spin interactions
arXiv 2607.28183 · 2026-08-03$a_0(980)^+ f_0(500,980)$与$a_1(1260)^+η$产生机制在衰变$D_s^+ \to π^+ π^0 π^0 η$中的作用
Roles of $a_0(980)^+ f_0(500,980)$ and $a_1(1260)^+η$ production mechanisms in decay $D_s^+ \to π^+ π^0 π^0 η$
arXiv 2607.28169 · 2026-08-03利用感知能力：针对多模态大语言模型智能体的隐蔽并发音频提示注入攻击
Piggybacking on Perception: Stealthy Concurrent Audio Prompt Injections against Multimodal LLM Agents
arXiv 2607.28165 · 2026-08-03极化激元凝聚体中的三维 Kardar–Parisi–Zhang 标度
Three-Dimensional Kardar--Parisi--Zhang Scaling in Polariton Condensates
arXiv 2607.28106 · 2026-08-03回归类问题中相关特征子集与子空间的联合同步学习
On a joint simultaneous learning of relevant feature subsets and subspaces in regression-like problems
arXiv 2607.28080 · 2026-08-03带跳跃信号的多资产效用最大化
Multi-Asset Utility Maximization with Jump Signals
arXiv 2607.28051 · 2026-08-03关于Gorenstein对称猜想的一个证明
On a proof of the Gorenstein Symmetry Conjecture
arXiv 2607.28011 · 2026-08-03氧化物纳米管支架上的可调谐保形石墨烯生长：面向超润湿分级2D-3D结构
Tunable Conformal Graphene Growth on Oxide Nanotube scaffolds: Towards Superwettable Hierarchical 2D-3D Architectures
arXiv 2607.27980 · 2026-08-0322年超级神冈（Super-Kamiokande）太阳$^8\rm{B}$中微子数据的周期图对比分析：经典方法、基于相位的方法及信息论方法
Comparative Periodogram Analysis of 22 Years of Super-Kamiokande Solar $^{8}\mathrm{B}$ Neutrino Data: Classical, Phase-Based, and Information Theoretic Methods
arXiv 2607.27979 · 2026-08-03TriShield：通过正交梯度投影与优化器状态纠缠实现联邦语言模型微调中隐私后门的零效用损失防御
TriShield: Zero-Utility-Loss Defense Against Privacy Backdoors in Federated Language Model Fine-Tuning via Orthogonal Gradient Projection and Optimizer State Entanglement
arXiv 2607.27940 · 2026-08-03不要信任AI生态系统：分析受感染开源组件中的隐私泄露
Don't Trust the AI Ecosystem: Analyzing Privacy Leakage in Compromised Open-Source Components
arXiv 2607.27886 · 2026-08-03超越借用的历史：用于交互式角色扮演评估的人物对齐用户模拟
Beyond Borrowed Histories: Person-Aligned User Simulation for Interactive Role-Playing Evaluation
arXiv 2607.27816 · 2026-08-03RIPPLE：生成多通道相位，而非恢复相位
RIPPLE: Generating Multi-Channel Phase, Not Recovering It
arXiv 2607.27775 · 2026-08-03
