2026-10论文中文摘要
大距离梯度未必可靠：面向长时程自回归预测的可靠性加权信用分配
Large Distant Gradients Need Not Be Reliable: reliability-weighted credit assignment for long-horizon autoregressive forecasting
arXiv 2609.12890 · 2026-10-08AKTS：面向语言模型智能体的亚微秒级内核策略切换
AKTS: Sub-Microsecond Kernel Policy Switching for Language-Model Agents
arXiv 2609.12276 · 2026-10-08量化对称性破缺
Quantifying Symmetry Breaking with Metric Adjusted Quantum Geometric Tensors
arXiv 2609.11926 · 2026-10-08数据稀缺与模型稀疏：混合专家模型对重复数据的过拟合更严重
Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data
arXiv 2609.11917 · 2026-10-08面向细粒度情感与时长控制的自然语言后训练零样本文本到语音合成
Post-Training Zero-Shot TTS for Fine-Grained Emotion and Duration Control via Natural Language
arXiv 2609.11523 · 2026-10-08广义分数匹配用于凸域上的参数估计
Generalised Score Matching on Convex Domains
arXiv 2609.11521 · 2026-10-08曲线的 $p$-adic 幺模定理
A $p$-adic monodromy theorem for curves
arXiv 2609.11106 · 2026-10-08马尔可夫范畴中的连接（Joinings）
Joinings in Markov categories
arXiv 2609.10912 · 2026-10-08Mordukhovich导数的一些性质及其在Banach空间局部变分不等式中的应用
Some Properties of Mordukhovich Derivatives with Applications to Locally Variational Inequalities in Banach Spaces
arXiv 2609.10721 · 2026-10-08非热太阳停滞：氙中非弹性标量信号
Solar Capture and Suppressed Annihilation of Inelastic Scalar Dark Matter
arXiv 2609.10636 · 2026-10-08行动、等待或实验：弃权情形下零售定价优化的因果治理框架
ACT, WAIT, or EXPERIMENT: A Causal Governance Framework for Retail Price Optimization Under Abstentions
arXiv 2609.10615 · 2026-10-08混合 $Q$-曲率刚性的尖锐阈值
A sharp threshold for mixed $Q$-curvature rigidity
arXiv 2609.10527 · 2026-10-08语义瓶颈：利用语义表示进行非侵入式语音解码
The Semantic Bottleneck: Leveraging Semantic Representations for Non-Invasive Speech Decoding
arXiv 2609.10296 · 2026-10-08Alice条件化量子Bell证书的无界次数开销
Contact geometry and sharp degree costs of quantum Bell certificates
arXiv 2609.10162 · 2026-10-08统计证据何时足够强？利用假设检验为数据收集赋值
When is statistical evidence strong enough? Using hypothesis tests to value data collection
arXiv 2609.09544 · 2026-10-08基于机器学习推断的德拜质量的底夸克偶素抑制
Bottomonium suppression with a machine-learning-informed Debye mass
arXiv 2609.09047 · 2026-10-08零和网络化可分离马尔可夫博弈中策略动力学的末次迭代收敛
Last-Iterate Convergence of Policy Dynamics in Zero-Sum Networked Separable Markov Games
arXiv 2609.08823 · 2026-10-08多智能体LLM辩论中分歧与回答质量的分层分析
What Does Multi-Agent Debate Actually Change?
arXiv 2609.08016 · 2026-10-08黎曼zeta函数在短区间内的简单临界零点和互异零点
Simple critical zeros and distinct zeros of the Riemann zeta function in short intervals
arXiv 2609.07918 · 2026-10-08分解引导的扩散语言模型用于惯性约束聚变预测
Decomposition-Guided Diffusion Language Models for Inertial Confinement Fusion Prediction
arXiv 2609.07756 · 2026-10-08微波无 $^{13}$C 金刚石颗粒超极化：基于魔角旋转与NV色心
Microwave-Free $^{13}$C Hyperpolarization of Diamond Particles Enabled by Magic Angle Spinning and NV Centers
arXiv 2609.07653 · 2026-10-08抛物双曲空间中的几何有限性
Geometric finiteness in paracomplex hyperbolic spaces
arXiv 2609.07628 · 2026-10-08COINS：任意阶段有效且面向效用的序贯共形预测
COINS: Any-Stage-Valid and Utility-Oriented Sequential Conformal Prediction
arXiv 2609.07112 · 2026-10-08LZ暗物质信号的多信使与古探测器探测
A Common $\sim$55 GeV Inelastic Dark Matter Origin of the LZ, Fermi-LAT, and AMS-02 Excesses and Paleo-Detector Prospects
arXiv 2609.06640 · 2026-10-08Wigner Time：面向量子科学与技术的实验时间线创建的数据导向方法
Wigner Time: a data-oriented approach to experimental timeline creation for quantum science and technology
arXiv 2609.06230 · 2026-10-08成功何处失效：面向鲁棒视觉-语言-动作模型的失败边界学习
Where Success Breaks: Failure-Boundary Learning for Robust Vision-Language-Action Models
arXiv 2609.06114 · 2026-10-08动态开放耦合SYK马约拉纳结中的红外记忆与搅乱
The Junction That Remembers: Infrared Response of Two Coupled SYK Majorana Systems
arXiv 2609.05336 · 2026-10-08混合Sobolev空间中的浅层神经网络逼近
Shallow neural network approximation in mixed Sobolev spaces
arXiv 2609.05263 · 2026-10-08匹配的起点，分化的对象：人机协作如何形成其解释内容
Matched Starts, Divergent Objects: Research-Object Formation in Human-AI Scholarly Work
arXiv 2609.04542 · 2026-10-08并发随机博弈的鲁棒PAC学习
Robust PAC Learning of Concurrent Stochastic Games
arXiv 2609.04189 · 2026-10-08递归查询语言的属性测试
Property Testing for Recursive Query Languages
arXiv 2609.03908 · 2026-10-08HalluPeer：用于检测科学同行评审中幻觉的分类驱动基准
HalluPeer: A Taxonomy-driven Benchmark for Detecting Hallucinations in Scientific Peer Reviews
arXiv 2609.03580 · 2026-10-08颗粒棒在更密集的障碍物场中下落更快
Granular Rods Fall Faster in Denser Obstacle Fields
arXiv 2609.03473 · 2026-10-08低轨卫星网络中用于双层空中在线联邦学习的自适应波束跳变与功率控制
Adaptive Beam Hopping and Power Control for Dual-Layer Over-the-Air Online Federated Learning in LEO Satellite Networks
arXiv 2609.03202 · 2026-10-08原子热力学景观的生成嵌套采样
Generative Nested Sampling of Atomistic Thermodynamic Landscapes
arXiv 2609.03193 · 2026-10-08交幂图与幂图的差图的嵌入问题
On embeddings of the difference graph of the intersection power graph and the power graph
arXiv 2609.02945 · 2026-10-08POSEIDON III：围绕热恒星WASP-195的热海王星的对齐轨道
POSEIDON III: The Aligned Orbit of the Hot Neptune Around the Hot Star WASP-195
arXiv 2609.02888 · 2026-10-08语言不可读性对大语言模型（LLM）安全性的影响
The Implications of Linguistic Illegibility for LLM Security
arXiv 2609.02852 · 2026-10-08关于素计数函数的黎曼-冯·曼戈尔特显式公式相关的发散问题
On divergence related to Riemann--von Mangoldt's explicit formula of the prime-counting function
arXiv 2609.02713 · 2026-10-08雾海之上的希格斯微子
Higgsino Above the Sea of Fog
arXiv 2609.01504 · 2026-10-08d≥3时斜高度函数与非交叉曲面的刚性
On the rigidity of sloped height functions in $d\ge 3$ and non-crossing surfaces
arXiv 2609.01499 · 2026-10-08趋化作用诱导的反应-扩散模型中的线性不稳定性与斑图形成
Chemotaxis-induced linear instabilities and pattern formation in a reaction-diffusion model
arXiv 2609.01159 · 2026-10-08趋化作用对模式形成的非单调控制
Re-entrance or monotonic: How chemotaxis controls pattern formation
arXiv 2609.01127 · 2026-10-08用于建模迭代问题解决的数据集
A Dataset for Modeling Iterative Problem-Solving
arXiv 2609.00940 · 2026-10-08答案并非论证
The Answer Is Not the Argument
arXiv 2609.00264 · 2026-10-08采用考虑路段流量传播引导的强化学习方法的动态OD矩阵在线估计
Online Estimation of Dynamic Origin-Destination Matrices Using Reinforcement Learning with Link-Flow Propagation Guidance
arXiv 2608.30317 · 2026-10-08低Q²下³He自旋结构函数及其矩的测量
Measurement of the $^3$He Spin Structure Functions and Their Moments at Low Q$^2$
arXiv 2608.30101 · 2026-10-08EDGE：基于图结构DSL配置的对话模拟实现确定性图评估的引擎
EDGE: Engine for Deterministic Graph Evaluation through Conversation Simulation from Graph Structured DSL Configuration
arXiv 2608.29971 · 2026-10-08S矩阵猜想
An independent proof of the even-dimensional S-matrix inequality
arXiv 2608.29750 · 2026-10-08面向自主云 MLOps 的前置式全栈工程
Forward-Deployed Full-Stack Engineering for Autonomous Cloud MLOps
arXiv 2608.29615 · 2026-10-08
