2026-08论文中文摘要
超越 token 级交叉熵：用于自回归图像生成的 Fréchet 分布后训练
Beyond Token-Level Cross-Entropy: Fréchet Distributional Post-Training for Autoregressive Image Generation
arXiv 2608.00562 · 2026-08-04通过LENS：视觉语言模型表示的局部几何分解
Through the LENS: Local Geometric Decomposition of Vision-Language Model Representations
arXiv 2608.00561 · 2026-08-04用于有效多体哈密顿量的自适应算子生成子空间
Adaptive operator-generated subspaces for effective many-body Hamiltonians
arXiv 2608.00560 · 2026-08-04开放集AIGC检测的测试时课程
Test-Time Curriculum for Open-Set AIGC Detection
arXiv 2608.00559 · 2026-08-04AiFlow：面向流式大语言模型应用的带界背压令牌原生反应式编排框架
AiFlow: Token-Native Reactive Orchestration with Bounded Backpressure for Streaming LLM Applications
arXiv 2608.00558 · 2026-08-04平面图的恰当无冲突7-着色
Proper conflict-free 7-coloring of planar graphs
arXiv 2608.00555 · 2026-08-04DexMani：用于灵巧旋转的人类衍生可操作度引导
DexMani: Human-Derived Manipulability Guidance for Dexterous Rotation
arXiv 2608.00554 · 2026-08-04无自旋Salpeter方程描述的四夸克态质量谱
Mass spectra of Tetraquark States from the Spinless Salpeter Equation
arXiv 2608.00553 · 2026-08-04共享内存领域的一段历程
A Journey in Shared Memory Land
arXiv 2608.00552 · 2026-08-04PHA-Net：用于文本-视频检索的基于原型的分层对齐网络
PHA-Net: Prototype-based Hierarchical Alignment Network for Text-Video Retrieval
arXiv 2608.00551 · 2026-08-04引力红移诱导的光子系统中的多参数量子估计
Multiparameter quantum estimation in a photon system induced by gravitational redshift
arXiv 2608.00550 · 2026-08-04由大语言模型（LLMs）驱动的上下文感知文化遗产指南
A Context-Aware Cultural Heritage Guide Powered by LLMs
arXiv 2608.00549 · 2026-08-04DrawAI：用于生成可编辑光栅图像的智能体基准与工作流
DrawAI: Agentic Benchmark and Workflow for Making Raster Images Editable
arXiv 2608.00548 · 2026-08-04解耦视觉-触觉前瞻：面向世界动作模型的神谕引导式接口发现
Disentangling Visuo-Tactile Foresight: Oracle-Guided Interface Discovery for World Action Models
arXiv 2608.00547 · 2026-08-04超越提示词依从性：语音生成中属性级语音控制的审计
Beyond Prompt Adherence: Auditing Attribute-Level Voice Control in Speech Generation
arXiv 2608.00545 · 2026-08-04零成本虚拟RNA：通过跨模态WSI检索近似免疫治疗特征
Zero-Cost Virtual RNA: Approximating Immunotherapy Signatures via Cross-Modal WSI Retrieval
arXiv 2608.00544 · 2026-08-04鲁棒水印与中毒模型相遇：通过隐秘后门规避扩散语义水印
Robust Watermarks Meet Backdoored Models: Evading Diffusion Semantic Watermarks via Stealthy Backdoor
arXiv 2608.00543 · 2026-08-04智能体图令牌推理
Agentic Graph Token Reasoning
arXiv 2608.00542 · 2026-08-04可用性与法律实践探索：来自德国司法领域数字取证用户的见解
Exploring Usability and Legal Practice: Insights from German Judicial Users of Digital Forensics
arXiv 2608.00541 · 2026-08-04DiffuseAgent-MI：用于忠实视觉推理的基于分布、集成工具的自进化智能体
DiffuseAgent-MI: Distributionally-Grounded,Tool-Integrated Self-Evolving Agents for Faithful Visual Reasoning
arXiv 2608.00540 · 2026-08-04噪声采样下的频率编码
Frequency Coding over Noisy Sampling
arXiv 2608.00539 · 2026-08-04DE-NER：通过大型语言模型的对话引导实现零样本命名实体识别
DE-NER : Zero-shot Named Entity Recognition via Dialogue Elicitation of Large Language Models
arXiv 2608.00538 · 2026-08-04基于潜在流匹配的混合域后验采样用于逆问题
Hybrid-Domain Posterior Sampling for Inverse Problems via Latent Flow Matching
arXiv 2608.00537 · 2026-08-04环面富集的动机布吕阿复形与极大紧群
Torus-enriched Motivic Bruhat Complexes and Maximal Compact Groups
arXiv 2608.00535 · 2026-08-04印度史诗中的性别化结构角色
Gendered Structural Roles in Indian Epics
arXiv 2608.00534 · 2026-08-04低资源东南亚语言中的原生多语言思维链推理
Native Multilingual Chain-of-Thought Reasoning in Low-Resource Southeast Asian Languages
arXiv 2608.00533 · 2026-08-04DTS拟群基础：代数、几何与计数
Basics of DTS quasigroups: algebra, geometry and enumeration
arXiv 2608.00532 · 2026-08-04CURE：面向块并行推测解码的局部不确定性修复
CURE: Local Uncertainty Repair for Block-Parallel Speculative Decoding
arXiv 2608.00531 · 2026-08-04释放文本的力量：面向复杂退化场景的文本引导流匹配图像融合
Unleashing the Power of Text: Text-Guided Flow Matching for Image Fusion under Complex Degradations
arXiv 2608.00530 · 2026-08-04用于量子多体态高维积分与采样的薛定谔生成器
Schrödinger Generator for High-Dimensional Integration and Sampling on Quantum Many-Body States
arXiv 2608.00529 · 2026-08-04SSTG-Nav：用于可重复物体导航的度量空间语义拓扑图
SSTG-Nav: Metric-Grounded Spatial-Semantic Topological Graphs for Reusable Object Navigation
arXiv 2608.00527 · 2026-08-04量子自旋霍尔绝缘体铋烯及其前驱相的自旋分辨与角分辨光电子能谱研究
Spin- and Angle-resolved Photoelectron Spectroscopy Study of the Quantum Spin Hall Insulator Bismuthene and its Precursor Phase
arXiv 2608.00526 · 2026-08-04基于边缘辅助混合量子优化的协同车队路由与调度
Cooperative Platoon Routing and Dispatching via Edge-Assisted Hybrid Quantum Optimization
arXiv 2608.00524 · 2026-08-04采用子问题集与随机离散速度策略的内存高效确定性多尺度气体流模拟方法
A memory-efficient deterministic method for multiscale gas flows using an ensemble-of-subproblems strategy with stochastic discrete velocities
arXiv 2608.00521 · 2026-08-04利用高分辨率光谱学研究系外行星大气的组成、温度与动力学
Using High-Resolution Spectroscopy to Study the Composition, Temperature, and Dynamics of Exoplanet Atmospheres
arXiv 2608.00520 · 2026-08-04离壳Kerr几何的凸性判据与径向分布响应：作为解析基准的模糊暗物质分布
Convexity criterion and radial-profile response for off-shell Kerr geometries: a fuzzy-dark-matter profile as an analytic benchmark
arXiv 2608.00519 · 2026-08-04GuideGround：用于3D视觉定位的VLM引导语义理解与视点感知推理
GuideGround: VLM-guided Semantic Understanding and Viewpoint-aware Reasoning for 3D Visual Grounding
arXiv 2608.00518 · 2026-08-04通过支付规则模糊性实现精确预算平衡：激励保持、转移能力与参与度
Exact Budget Balance via Payment-Rule Ambiguity: Incentive Preservation, Transfer Capacity, and Participation
arXiv 2608.00517 · 2026-08-04LLM辅导器中教学式泄露的可审计发布控制
Auditable Release Control for Pedagogical Leakage in LLM Tutors
arXiv 2608.00515 · 2026-08-04采用子问题集成策略结合随机离散速度缓解稀薄流动模拟中的射线效应
Mitigating ray effects in rarefied flow simulations using an ensemble-of-subproblems strategy with stochastic discrete velocities
arXiv 2608.00514 · 2026-08-04BayesSeg：用于用电量时间序列状态分割的贝叶斯优化框架
BayesSeg: A Bayesian Optimization Framework for State Segmentation of Electricity Consumption Time Series
arXiv 2608.00513 · 2026-08-04无限层镍酸盐中$d_{x^2-y^2}$波超导性的从头算预测
{\it Ab initio} prediction of $d_{x^2-y^2}$-wave superconductivity in infinite-layer nickelates
arXiv 2608.00512 · 2026-08-04用于比较人口普查数据收集方法的总统计误差框架
A Total Statistical Error Framework for Comparing Census Data Collection Methods
arXiv 2608.00511 · 2026-08-04基于动态可靠性引导的骨盆骨分割模型测试时适应
Test-time Adaptation of Pelvic Bone Segmentation Models via Dynamic Reliability-Guided
arXiv 2608.00510 · 2026-08-04两个与双重和相关的q超同余式
Two $q$-Supercongruences Related to Double Sums
arXiv 2608.00509 · 2026-08-04RadYOLO：面向CT和MRI的计算高效型3D目标检测与分割方法
RadYOLO: Computationally Efficient 3D Object Detection and Segmentation in CT and MRI
arXiv 2608.00508 · 2026-08-04学习目标调控感知窄化：自监督语音编码器的跨语言、分层、十种子研究
The Learning Objective Governs Perceptual Narrowing: A Cross-Lingual, Layer-Wise, Ten-Seed Study of Self-Supervised Speech Encoders
arXiv 2608.00507 · 2026-08-04向量空间在t-覆盖数约束下的极值交叉t-相交族
Extremal cross $t$-intersecting families under $t$-covering number constraints for vector spaces
arXiv 2608.00505 · 2026-08-04递归高斯过程与贝叶斯大脑
Recursive Gaussian Processes and the Bayesian Brain
arXiv 2608.00503 · 2026-08-04SpatialAfford：教紧凑视觉语言模型（VLMs）关注何处与定位何处以实现 affordance
SpatialAfford: Teaching Compact VLMs Where to Look and Where to Ground for Affordance
arXiv 2608.00502 · 2026-08-04
