2026-09论文中文摘要
T-MoXAI：面向时间多模态数据的层级可解释性框架
T-MoXAI: A Hierarchical Explainability Framework for Temporal Multimodal Data
arXiv 2609.33685 · 2026-09-29改进的基于递归高斯证书的 3-可着色图 SDP 着色
Improved SDP Coloring of 3-Colorable Graphs from Recursive Gaussian Certificates
arXiv 2609.33684 · 2026-09-29MAD-Guard：封闭式多模态取证任务中自回归生成与直接决策接口的受控研究
MAD-Guard: Controlled Study of Autoregressive Generation versus Direct Decision Interfaces for Closed Multimodal Forensic Tasks
arXiv 2609.33683 · 2026-09-29在线社交网络上SDIR流行病模型的一种新方法
A Novel Approach for the SDIR Epidemic Model on Online Social Networks
arXiv 2609.33682 · 2026-09-29面向软体机器人的可观测性信息最优传感器布置
Observability-Informed Optimal Sensor Placement for Soft Robots
arXiv 2609.33681 · 2026-09-29关于并图的定量 Merino--Welsh 不等式
Quantitative Merino--Welsh inequalities for joins
arXiv 2609.33680 · 2026-09-29具有大边界的klt对的最小模型计划的完全终止性
Full termination of MMP for klt pairs with big boundaries
arXiv 2609.33679 · 2026-09-29一般范数与分布下的良性过拟合
Benign Overfitting for General Norms and Distributions
arXiv 2609.33675 · 2026-09-29可达性并不足够：诊断图神经网络中的长程行为
Reachability is not enough: Diagnosing long-range behavior in GNNs
arXiv 2609.33674 · 2026-09-29暗能量与宇宙的负压缩性：基于宇宙状态方程的研究
Dark energy and negative compressibility of the Universe: A study based on cosmic equations of state
arXiv 2609.33673 · 2026-09-29重置并非恢复：通过谄媚滞后评估从虚假对话上下文中恢复的能力
Reset Is Not Recovery: Evaluating Recoverability from False Conversational Context via Sycophancy Hysteresis
arXiv 2609.33672 · 2026-09-29COGNIT-Guard：在显式延迟与假阳性约束下，基于异构CPU-NPU置信级联的校准独立直接决策护栏
COGNIT-Guard: Calibrated Standalone Direct-Decision Guardrails with Heterogeneous CPU-NPU Confidence Cascading under Explicit Latency and False-Positive Constraints
arXiv 2609.33671 · 2026-09-29弥合跨方言差距：查询计划作为文本到SQL中的可移植接口
Closing the Cross-Dialect Gap: Query Plans as a Portable Interface in Text-to-SQL
arXiv 2609.33670 · 2026-09-29RSD-Poker：不完全信息博弈中残差策略的结构自适应与移位鲁棒风险-效用认证
RSD-Poker: Structure-Adaptive and Shift-Robust Risk-Utility Certification for Residual Policies in Imperfect-Information Games
arXiv 2609.33669 · 2026-09-29当噪声遇上长尾：特征-阈值双重校准用于稳健伪标签
When Noise Meets Long-Tail: Feature-Threshold Dual Calibration for Robust Pseudo-Labeling
arXiv 2609.33668 · 2026-09-29训练期间模型行为的可扩展归因与控制
Scalable Attribution and Control of Model Behavior During Training
arXiv 2609.33667 · 2026-09-29一个模型不是群体：多LLM与方面条件化的多样化评论生成
One Model Is Not a Crowd: Multi-LLM and Aspect-Conditioned Diverse Comment Generation
arXiv 2609.33666 · 2026-09-29LieDiscover：用于显式开放形式对称性发现的自适应符号库构建
LieDiscover: Adaptive Symbolic Library Construction for Explicit Open-form Symmetry Discovery
arXiv 2609.33663 · 2026-09-29审计优先VAPO：不完美验证下的风险认证选择性更新
Audit-First VAPO: Risk-Certified Selective Updates under Imperfect Verification
arXiv 2609.33662 · 2026-09-29Kerr-Newman黑洞的扰动：分离与模式稳定性
Perturbations of Kerr-Newman black holes: separation and mode stability
arXiv 2609.33661 · 2026-09-29学习具有证据对齐读出层的多模态嵌入
Learning Multimodal Embeddings with Evidence-Aligned Readout
arXiv 2609.33659 · 2026-09-29AgentBoundary：工具使用型LLM智能体的安全性反事实评估
AgentBoundary: Counterfactual Evaluation of Safety in Tool-Using LLM Agents
arXiv 2609.33658 · 2026-09-29太空中有机磷物种的合成：P$^+$与C$_2$H$_2$的反应
Synthesis of organo-phosphorous species in space: the reaction of P$^+$ with C$_2$H$_2$
arXiv 2609.33657 · 2026-09-29复曲面的RC正性
RC-positivity of complex surfaces
arXiv 2609.33656 · 2026-09-29迈向消除数学推理任务课程学习中的灾难性遗忘
Towards Eliminating Catastrophic Forgetting in the Curriculum Learning of Math Reasoning Tasks
arXiv 2609.33655 · 2026-09-29基于非线性参与因子的电力系统模型降阶方法应对近共振条件
Nonlinear Participation Factor-based Power System Model Reduction Addressing Near-Resonance Conditions
arXiv 2609.33654 · 2026-09-29无演示的成功概率奖励学习用于通用机器人策略
Demonstration-Free Success-Probability Reward Learning for Generalist Robot Policies
arXiv 2609.33653 · 2026-09-29Tsubame：基于扩散的投机解码的树重放
Tsubame: Tree Replay for Diffusion-Based Speculative Decoding
arXiv 2609.33652 · 2026-09-29平坦排列多项式的对数凹性
Log-concavity of flat arrangement polynomials
arXiv 2609.33651 · 2026-09-29FuseAlign：野外环境下的强制对齐
FuseAlign: Forced Alignment in the Wild
arXiv 2609.33650 · 2026-09-29从分布到随机过程：测度值映射的神经逼近
From Distributions to Stochastic Processes: Neural Approximation of Measure-Valued Maps
arXiv 2609.33649 · 2026-09-29面向波束斜视下宽带集成感知与通信的联合收发机与组索引调制设计
Joint Transceiver and Group Index Modulation Design for Wideband Integrated Sensing and Communications Under Beam Squint
arXiv 2609.33648 · 2026-09-29InfraVLA：利用基础设施摄像头扩展视觉-语言-动作导航
InfraVLA: Extending Vision-Language-Action Navigation with Infrastructure Cameras
arXiv 2609.33647 · 2026-09-29探针即行动：通过主动视觉探测提升浏览器使用智能体
Probe to Act: Elevating Browser-Use Agent via Active Visual Probing
arXiv 2609.33646 · 2026-09-29AI对猜想9.1的解答
AI's solution of Conjecture 9.1
arXiv 2609.33644 · 2026-09-29学习从上下文中学习：基于扰动公开文档的合成训练
Learning to Learn from Context: Synthetic Training from Perturbed Public Documents
arXiv 2609.33642 · 2026-09-29可扩展且数据驱动的维护、修理和大修过程决策支持
Scalable and Data-Driven Decision Support in the Maintenance, Repair, and Overhaul Process
arXiv 2609.33641 · 2026-09-29SafeMol：分子多模态模型的双模态安全对齐
SafeMol: Dual-Modality Safety Alignment for Molecular Multimodal Models
arXiv 2609.33640 · 2026-09-29基于LLM智能体的轨迹遗忘
Trajectory Unlearning on LLM-based Agents
arXiv 2609.33639 · 2026-09-29通信丢失下仓库机器人协调的分层多智能体强化学习
Hierarchical Multi-agent Reinforcement Learning for Warehouse Robot Coordination under Communication Loss
arXiv 2609.33637 · 2026-09-29$m$ 叉递归树中的距离
Distances in $m$-ary recursive trees
arXiv 2609.33636 · 2026-09-29安全重构：通过掩码扩散进行生成式建模构建强大的安全护栏
Safety Reconstructed: Generative Modeling via Masked Diffusion Builds Strong Safety Guardrails
arXiv 2609.33634 · 2026-09-29渐近秩处无线性填充的CAT(0)空间
CAT(0) spaces without linear filling at the asymptotic rank
arXiv 2609.33633 · 2026-09-29K3 曲面上的刚性及非刚性各向同性 Nef 类
Rigid and Nonrigid Isotropic Nef Classes on K3 Surfaces
arXiv 2609.33632 · 2026-09-29交变磁体中时间反演偶和奇自旋霍尔电导率张量的计算对称性层级
Computational symmetry hierarchy of time-reversal-even and -odd spin Hall conductivity tensors in altermagnets
arXiv 2609.33631 · 2026-09-29lapanda：一种用于非凸约束优化层的无矩阵可微求解器
lapanda: A Matrix-Free Differentiable Solver for Nonconvex Constrained Optimization Layers
arXiv 2609.33630 · 2026-09-29StoryEngine：一种基于状态锚定的智能体视频叙事框架
StoryEngine: A State-Grounded Agentic Framework for Video Storytelling
arXiv 2609.33627 · 2026-09-29$S^2\times S^2$ 上具有两个不动点的光滑哈密顿微分同胚
A Smooth Hamiltonian Diffeomorphism with Two Fixed Points on $S^2\times S^2$
arXiv 2609.33626 · 2026-09-29连续投影逆定理及唯一可解码码对的加权构造
A Continuous Projection Converse and a Weighted Construction for Uniquely Decodable Code Pairs
arXiv 2609.33625 · 2026-09-29从用户视角刻画人机交互中的记忆错位
Characterizing Memory Misalignment in Human-LLM Interaction From User Perspectives
arXiv 2609.33623 · 2026-09-29
