2026-10论文中文摘要
关于Silvia Calderazzo、Manuel Wiesenfarth、Vivienn Weru和Annette Kopp-Schneider所著《具有外部对照数据信息借用设计的双臂临床试验中原则性I型错误率膨胀》的讨论
Discussion on "Principled type I error rate inflation in two-arm clinical trial designs with external control information borrowing" by Silvia Calderazzo, Manuel Wiesenfarth, Vivienn Weru, and Annette Kopp-Schneider
arXiv 2610.00711 · 2026-10-02ReLiveGym：在数周重放现实中评估长寿智能体
ReLiveGym: Evaluating Long-Lived Agents over Weeks of Replayed Reality
arXiv 2610.00710 · 2026-10-02极化簇中超曲面的自同构
Automorphisms of hypersurfaces in polarized varieties
arXiv 2610.00709 · 2026-10-02超越单峰基底：多模态数据的拉回几何
Beyond Unimodal Bases: Pullback Geometry for Multimodal Data
arXiv 2610.00708 · 2026-10-02初始化提升大语言模型驱动的发现
Initialization Improves LLM-Driven Discovery
arXiv 2610.00707 · 2026-10-02AnchorPrompt：用于稳健音频-语言模型的自蒸馏软提示
AnchorPrompt: Self-Distilled Soft Prompts for Robust Audio-Language Models
arXiv 2610.00706 · 2026-10-02元多智能体强化学习用于交互策略的快速适应及其在自动驾驶中的应用
Meta-Multi-Agent Reinforcement Learning for Fast Adaptation of Interactive Policies with Applications to Autonomous Driving
arXiv 2610.00705 · 2026-10-02SkillSpec: 基于共识门控与表示特化的智能体技能演化
SkillSpec: Consensus-Gated Agent Skill Evolution via Representation Specialization
arXiv 2610.00704 · 2026-10-02与Dunkl-Schrödinger算子相关的Hardy空间的Riesz变换刻画
Riesz transform characterization of Hardy spaces associated with Dunkl--Schrödinger operators
arXiv 2610.00703 · 2026-10-02Ditto：广义可重配置的线性化读取
Ditto: Generalized Reconfigurable Linearizable Reads
arXiv 2610.00702 · 2026-10-02从图像到任务：野外多模态大语言模型交互的特征刻画
From Images to Tasks: Characterizing Multimodal LLM Interactions in the Wild
arXiv 2610.00701 · 2026-10-02R-GroundBench：Markush分子编辑中R基团定位的诊断基准
R-GroundBench: A Diagnostic Benchmark for R-Group Groundingin Markush Molecular Editing
arXiv 2610.00700 · 2026-10-02奇异夸克星与中子星的共存：原中子星中的亚稳态与成核
Coexistence of strange quark stars and neutron stars: metastability and nucleation in proto-neutron stars
arXiv 2610.00699 · 2026-10-02三角输运的多保真度公式
Multifidelity Formulations for Triangular Transport
arXiv 2610.00698 · 2026-10-02认证交易传播的带宽费用机制
Bandwidth Fee Mechanisms for Certified Transaction Dissemination
arXiv 2610.00697 · 2026-10-02半格点多边形的 Pick 型定理
A Pick-type theorem for halfway-lattice polygons
arXiv 2610.00696 · 2026-10-02渐进分辨率联邦学习安全聚合
Progressive-Resolution Secure Aggregation for Federated Learning
arXiv 2610.00695 · 2026-10-02压缩语言模型中散度如何转化为决策翻转
How Divergence Becomes Decision Flips in Compressed Language Models
arXiv 2610.00694 · 2026-10-02FedMAD：面向遥感图像分类的联邦学习中的调制感知方向聚合
FedMAD: Modulation-Aware Directional Aggregation for Federated Learning in Remote Sensing Image Classification
arXiv 2610.00693 · 2026-10-02外籍培养教师与美国高影响力科学的协作组织
Foreign-trained faculty and the collaborative organization of high-impact U.S. science
arXiv 2610.00692 · 2026-10-02突触布局反映果蝇下行神经元的共享输入
Synaptic placement reflects shared input in Drosophila descending neurons
arXiv 2610.00690 · 2026-10-02迈向稳健的数值声明验证
Towards Robust Numerical Claim Verification
arXiv 2610.00689 · 2026-10-02双选择线性探测的威力
The Power of Two-Choice Linear Probing
arXiv 2610.00688 · 2026-10-02Leto：在幸存硬件上实现LLM训练的快速原地恢复
Leto: Fast In-Place Recovery for LLM Training on Surviving Hardware
arXiv 2610.00687 · 2026-10-02SemanTok：用于高效自回归视频生成的可预测语义标记
SemanTok: Predictable Semantic Tokens for Efficient Autoregressive Video Generation
arXiv 2610.00686 · 2026-10-02LoRA微调大语言模型的后门净化：基于零空间投影
Backdoor Purification for LoRA-Tuned LLMs via Null-Space Projection
arXiv 2610.00685 · 2026-10-02从滚动到销售：探究互动类型与设备价格对TikTok广告的影响
From Scroll to Sale: Exploring the Impact of Interaction Type and Device Price on TikTok Advertisements
arXiv 2610.00684 · 2026-10-02本体引导、推理器验证的科学人工智能大语言模型推理评估基准
Ontology-Grounded, Reasoner-Verified Benchmarks for Evaluating LLM Reasoning in Scientific AI
arXiv 2610.00682 · 2026-10-02半填充光学Su-Schrieffer-Heeger-Hubbard模型中的单轴应变
The half-filled optical Su-Schrieffer-Heeger-Hubbard model with uniaxial strain
arXiv 2610.00681 · 2026-10-02hZACH-ViT 中的曲率攻击：规范对称性、边界饱和与对抗失败
Curvature Under Attack in hZACH-ViT: Gauge Symmetry, Boundary Saturation, and Adversarial Failure
arXiv 2610.00680 · 2026-10-02贝叶斯微调使语言模型在信念允许的范围内达到贝叶斯行为
Bayesian Fine-tuning Yields Language Models that are as Bayesian as their Beliefs Allow
arXiv 2610.00679 · 2026-10-02利用视觉-语言模型进行增强现实中的感知质量评估与自主内容调整
Harnessing Vision-Language Models for Perceptual Quality Assessment and Autonomous Content Adjustment in Augmented Reality
arXiv 2610.00677 · 2026-10-02使用目标条件双模拟学习可迁移技能
Learning Transferable Skills using Goal-Conditioned Bisimulation
arXiv 2610.00676 · 2026-10-02LabBook：利用实验历史实现高效的LLM驱动发现
LabBook: Harnessing Experimental History for Efficient LLM-Driven Discovery
arXiv 2610.00675 · 2026-10-02ROBIN-PIP：基于物理信息先验的鲁棒贝叶斯场级推断
ROBIN-PIP: Robust Bayesian Field-Level Inference with Physics-Informed Priors
arXiv 2610.00674 · 2026-10-02闭环：循环语言模型的实用训练方案
Closing the Loop: Practical Training Recipes for Looped Language Models
arXiv 2610.00673 · 2026-10-02ORBIT-FMIB：通过ESM-2追踪阶序解析的上位性信息
ORBIT-FMIB: Tracking Order-Resolved Epistatic Information Through ESM-2
arXiv 2610.00672 · 2026-10-02MegaFlux：通过流水线专家复制实现抗偏斜的MoE超级内核
MegaFlux: Skew-Resilient MoE Megakernels via Pipelined Expert Replication
arXiv 2610.00671 · 2026-10-02全息互信息的单婚量子修正
Monogamous quantum corrections to holographic mutual information
arXiv 2610.00670 · 2026-10-02量子电路剪枝：从NISQ架构到容错操作
Quantum Circuit Pruning: From NISQ Architectures to Fault-Tolerant Operations
arXiv 2610.00669 · 2026-10-02一种用于规范引导决策的简单信念道义逻辑
A Simple Doxastic Deontic Logic for Norm-Guided Decision Making
arXiv 2610.00668 · 2026-10-02驱动-耗散量子系统的算符语言费曼规则：从平均场到非高斯光子关联
Operator-language Feynman rules for driven-dissipative quantum systems: from mean field to non-Gaussian photon correlations
arXiv 2610.00667 · 2026-10-02VisionQ：用于计算机视觉定性分析的VLM-as-a-Judge分类体系、数据集与基准
VisionQ: VLM-as-a-Judge Taxonomy, Dataset and Benchmark for Qualitative Analysis in Computer Vision
arXiv 2610.00666 · 2026-10-02量化与高效适配蛋白质语言模型的分析
Analysis of Quantized and Efficiently Adapted Protein Language Models
arXiv 2610.00665 · 2026-10-02PhysicsMate：基于课程的中学物理问答孟加拉语基准与小模型适配
PhysicsMate: A Curriculum-Grounded Bengali Benchmark for Secondary Physics QA with Small-Model Adaptation
arXiv 2610.00664 · 2026-10-02通过专家隔离与关闭实现大语言模型后门遏制
Backdoor Containment via Expert Quarantine and Shutdown in LLMs
arXiv 2610.00663 · 2026-10-02Silence-the-Mimic：加速针对语音克隆的不可感知扰动生成
Silence-the-Mimic: Accelerating Imperceptible Perturbation Generation Against Voice Cloning
arXiv 2610.00662 · 2026-10-02探索更多，推理更优：面向扩散语言模型的逐步风险敏感GRPO
Exploring More, Reasoning Better: Stepwise Risk-Sensitive GRPO for Diffusion Language Models
arXiv 2610.00661 · 2026-10-02互连离散时间系统的固定时间输入状态稳定性，第二部分：有限时域Lyapunov函数
Fixed-Time ISS for Interconnected Discrete-Time Systems, Part II: Finite-Horizon Lyapunov Functions
arXiv 2610.00660 · 2026-10-02开放星团 NGC 2437 (M 46) 中未分辨的三合星系统经 KMOS VVVX-GalCen 光谱巡天确认
Unresolved triples in the open cluster NGC 2437 (M 46) confirmed by the KMOS VVVX-GalCen Spectroscopic Survey
arXiv 2610.00659 · 2026-10-02
