2026-07论文中文摘要
我们能信任项目反应理论进行人工智能评估吗？
Can We Trust Item Response Theory for AI Evaluation?
arXiv 2607.15190 · 2026-07-20T^2MLR：具有时间中层循环的Transformer
T^2MLR: Transformer with Temporal Middle-Layer Recurrence
arXiv 2607.15178 · 2026-07-20使用有限VRAM进行长上下文微调
Long-Context Fine-Tuning with Limited VRAM
arXiv 2607.15105 · 2026-07-20数字万神殿：使用大语言模型智能体模拟和审计联盟形成
Digital Pantheon: Simulating and Auditing Coalition Formation with LLM Agents
arXiv 2607.15095 · 2026-07-20BrainPilot：通过智能研究实现大脑发现自动化
BrainPilot: Automating Brain Discovery with Agentic Research
arXiv 2607.15079 · 2026-07-20在路径积分蒙特卡罗中学习费米子符号结构
Learning the Fermion sign structure in path-integral Monte Carlo
arXiv 2607.15060 · 2026-07-20具有多个内幕人士的离散时间凯尔模型的存在性和收敛性
Existence and convergence of discrete-time Kyle models with multiple insiders
arXiv 2607.15057 · 2026-07-20非光滑区域中斯托克斯系统的\(L^p\)诺伊曼问题
The $L^p$ Neumann problem for the Stokes system in nonsmooth domains
arXiv 2607.15042 · 2026-07-20视频 = 世界 + 事件流
Video = World + Event Stream
arXiv 2607.15038 · 2026-07-20cGAP：用于高维分类数据可视化的带HOMALS引导热图的广义关联图
cGAP: Generalized Association Plots with HOMALS-Guided Heatmaps for Visualization of High-Dimensional Categorical Data
arXiv 2607.15018 · 2026-07-20通过几何极值图形模型进行洪水风险估计
Flood risk estimation via geometric extremal graphical models
arXiv 2607.15000 · 2026-07-20因果IPD-QIM网络流水印的排队稳定性准则
A Queueing-Stability Criterion for Causal IPD-QIM Network Flow Watermarking
arXiv 2607.14954 · 2026-07-20代数拟阵的识别是不可判定的
Recognition of algebraic matroids is undecidable
arXiv 2607.14907 · 2026-07-20VST ATLAS巡天-IV：通过ACT CMB引力透镜测量星系、大质量红移星系和类星体的偏差及晕占据分布
The VST ATLAS Survey IV: Galaxy, LRG and QSO bias and HODs via ACT CMB Lensing
arXiv 2607.14904 · 2026-07-20基于大语言模型的房地产搜索重排
LLM-Based Re-Ranking for Real Estate Search
arXiv 2607.14835 · 2026-07-20ATLAS18生产ITk条形传感器的质量控制（QC）总结
Summary of quality control (QC) of ATLAS18 production ITk strip sensors
arXiv 2607.14691 · 2026-07-20关于太阳活动区13664/8的日冕物质抛射生产率
The coronal mass ejection productivity of solar active region 13664/8
arXiv 2607.14636 · 2026-07-20记忆驱动的自我表露与关系转折点：人机交互的纵向多模态研究
Memory-Driven Self-Disclosure and Relational Turning Points: A Longitudinal Multimodal Study of Human-AI Interaction
arXiv 2607.14593 · 2026-07-20近阈单中子共振的普适性质
Universal Properties of Near-Threshold Single-Neutron Resonances
arXiv 2607.14464 · 2026-07-20补偿设计
Compensation Design
arXiv 2607.14438 · 2026-07-20一种具有超二次加速的弱非线性等离子体物理的端到端量子算法
An end-to-end quantum algorithm for weakly nonlinear plasma physics with superquadratic speedup
arXiv 2607.14308 · 2026-07-20关于双曲域的伯格曼数
On the Bergman number of hyperbolic domains
arXiv 2607.14265 · 2026-07-20在量子计算机上对纠缠分布网络进行仿真
Emulation of Entanglement Distribution Networks on a Quantum Computer
arXiv 2607.14260 · 2026-07-20偶数均匀超图的摩尔界
The Hypergraph Moore Bound
arXiv 2607.14068 · 2026-07-20GigaWorld-Policy-0.5：由自动研究赋能的更快更强的世界行动模型
GigaWorld-Policy-0.5: A Faster and Stronger WAM Empowered by AutoResearch
arXiv 2607.13960 · 2026-07-20生成式编译：人工智能生成代码时的即时编译器反馈
Generative Compilation: On-the-Fly Compiler Feedback as AI Generates Code
arXiv 2607.13921 · 2026-07-20MxGPS：用于电网基础模型的多路图变换器
MxGPS: Multiplex Graph Transformers for a Power Grid Foundation Model
arXiv 2607.13763 · 2026-07-20约束驱动的模型优化：现代机器学习系统中选择压缩和加速技术的行业框架
Constraint-Driven Model Optimization: An Industry Framework for Selecting Compression and Acceleration Techniques in Modern Machine Learning Systems
arXiv 2607.13735 · 2026-07-20通过反馈应对不确定威胁
Meeting Uncertain Threats with Feedback
arXiv 2607.13648 · 2026-07-20深度套期保值是强化学习吗？
Is Deep Hedging Reinforcement Learning?
arXiv 2607.13353 · 2026-07-20评估能力并不意味着优化效用：闭环表格识别中的大语言模型作为评判信号
LLM-as-a-Judge Scores Are Unreliable Optimization Signals in Closed-Loop Table Recognition
arXiv 2607.13347 · 2026-07-20自然语言断言的忠实自动形式化
Faithful Autoformalization of Natural Language Assertions
arXiv 2607.13303 · 2026-07-20逻辑S4.1的拓扑平方
Topological square of logic S4.1
arXiv 2607.13240 · 2026-07-20关于格里博夫视界谱几何的一些评论
Some Remarks on the Spectral Geometry of the Gribov Horizon
arXiv 2607.13228 · 2026-07-20模型表达、抑制和抗拒的内容：使用角色向量审计开放权重语言模型
What Models Express, Suppress, and Resist: Auditing Open-Weight LLMs with Persona Vectors
arXiv 2607.13162 · 2026-07-20世界合一演示：用于学习开放世界移动操作的合成数据引擎
Worlds in One Demo: A Synthetic Data Engine for Learning Open-World Mobile Manipulation
arXiv 2607.13154 · 2026-07-20对称防波堤上极端波浪破碎的能量学与随机学
Energetics and Stochastics of Extreme Waves Breaking over a Symmetrical Breakwater
arXiv 2607.12941 · 2026-07-20刘维尔型狄利克雷问题解的严格凸性
Strict Convexity for Solution of Liouville-Type Dirichlet Problems
arXiv 2607.12849 · 2026-07-20关于多个系统发育树中的一致子树
On Agreement Subtrees in Multiple Phylogenetic Trees
arXiv 2607.12778 · 2026-07-20基于最大可满足性的反馈在数独中引导视觉语言模型
MaxSAT-Based Feedback for Guiding Vision-Language Models in Sudoku
arXiv 2607.12711 · 2026-07-20用于视觉语言导航的实例增强语义地图
Instance-Enriched Semantic Maps for Visual Language Navigation
arXiv 2607.12630 · 2026-07-20通过半周长的 k - 凸多联骨牌
$k$-Convex Polyominoes by Semi-perimeter
arXiv 2607.12448 · 2026-07-20Code-MUE：通过基于执行的语义交互图测量代码语言模型的不确定性
Code-MUE: Measuring Code LLMs' Uncertainty through Execution-based Semantic Interaction Graphs
arXiv 2607.12273 · 2026-07-20从重建到解释：X射线断层扫描数据的零设置多相分割
From Reconstruction to Interpretation: Zero-Setup Multi-Phase Segmentation of X-ray Tomography Data
arXiv 2607.12175 · 2026-07-20完全可达道路着色
Completely Reachable Road Coloring
arXiv 2607.12078 · 2026-07-20用于稀缺神经数据的尺度感知注意力：基于睡眠脑电信号的RG流变压器
The RG-Flow Transformer: Encoding Scale-Free Dynamics in Scarce EEG
arXiv 2607.11950 · 2026-07-20扩展即时语言模型
Scaling Point-in-Time Language Models
arXiv 2607.11889 · 2026-07-20神经执行器：用于机器人动力学和外力感知的神经驱动建模
NeuralActuator: Neural Actuation Modeling for Robot Dynamics and External Force Perception
arXiv 2607.11734 · 2026-07-20多模态焦点的潮起潮落：为基于视觉的语言模型推理调度视觉中继窗口
The Ebb and Flow of Multimodal Focus: Scheduling Visual Relay Windows for Grounded VLM Reasoning
arXiv 2607.11436 · 2026-07-20空间异质景观上的元群落持久性
Metacommunity persistence on spatially heterogeneous landscapes
arXiv 2607.11291 · 2026-07-20
