机器学习学术速递[11.14] na

点击阅读原文访问arxivdaily.com，涵盖CS|物理|数学|经济|统计|金融|生物|电气领域，更有搜索、收藏等功能！

cs.LG 方向，今日共计158篇

大模型相关(15篇)

【1】Instella: Fully Open Language Models with Stellar Performance
标题：Instella：具有Stellar性能的完全开放语言模型
链接：https://arxiv.org/abs/2511.10628

【2】SSR: Socratic Self-Refine for Large Language Model Reasoning
标题：SR：大型语言模型推理的苏格拉底式自我精炼
链接：https://arxiv.org/abs/2511.10621

【3】EDGC: Entropy-driven Dynamic Gradient Compression for Efficient LLM Training
标题：EDGC：用于高效LLM训练的信息驱动动态梯度压缩
链接：https://arxiv.org/abs/2511.10333

【4】OutSafe-Bench: A Benchmark for Multimodal Offensive Content Detection in Large Language Models
标题：OutSafe-Bench：大型语言模型中多模式攻击性内容检测的基准
链接：https://arxiv.org/abs/2511.10287

【5】Lost in Serialization: Invariance and Generalization of LLM Graph Reasoners
标题：迷失在序列化中：LLM图推理器的不变性和推广
链接：https://arxiv.org/abs/2511.10234

【6】Bridging Synthetic and Real Routing Problems via LLM-Guided Instance Generation and Progressive Adaptation
标题：通过LLM引导的实例生成和渐进式适应弥合合成和真实路由问题
链接：https://arxiv.org/abs/2511.10233

【7】DemoTuner: Efficient DBMS Knobs Tuning via LLM-Assisted Demonstration Reinforcement Learning
标题：DemoTuner：通过LLM辅助演示强化学习实现高效的数据库优化
链接：https://arxiv.org/abs/2511.09998

【8】EEGAgent: A Unified Framework for Automated EEG Analysis Using Large Language Models
标题：EEGAgent：使用大型语言模型进行自动脑电分析的统一框架
链接：https://arxiv.org/abs/2511.09947

【9】HierRouter: Coordinated Routing of Specialized Large Language Models via Reinforcement Learning
标题：HierRouter：通过强化学习协调路由专业大型语言模型
链接：https://arxiv.org/abs/2511.09873

【10】Uncertainty-Guided Checkpoint Selection for Reinforcement Finetuning of Large Language Models
标题：大型语言模型强化微调的不确定性引导检查点选择
链接：https://arxiv.org/abs/2511.09864

【11】Unlearning Imperative: Securing Trustworthy and Responsible LLMs through Engineered Forgetting
标题：放弃学习势在必行：通过精心设计的遗忘确保值得信赖和负责任的LLM
链接：https://arxiv.org/abs/2511.09855

【12】ACT as Human: Multimodal Large Language Model Data Annotation with Critical Thinking
标题：像人一样行动：用批判性思维进行多模态大型语言模型数据标注
链接：https://arxiv.org/abs/2511.09833

【13】Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
标题：测试时光谱感知潜在引导，用于视觉语言模型中的Zero-Shot概括
链接：https://arxiv.org/abs/2511.09809

【14】ConstrainedSQL: Training LLMs for Text2SQL via Constrained Reinforcement Learning
标题：约束SQL：通过约束强化学习为Text2SQL训练LLM
链接：https://arxiv.org/abs/2511.09693

【15】Scaling Environments for LLM Agents in the Era of Learning from Interaction: A Survey
标题：交互学习时代LLM代理的扩展环境：调查
链接：https://arxiv.org/abs/2511.09586

Graph相关(图学习|图神经网络|图优化等)(6篇)

【1】Unlocking Dynamic Inter-Client Spatial Dependencies: A Federated Spatio-Temporal Graph Learning Method for Traffic Flow Forecasting
标题：解锁动态客户端间空间依赖：交通流预测的联邦时空图学习方法
链接：https://arxiv.org/abs/2511.10434

【2】GraphSB: Boosting Imbalanced Node Classification on Graphs through Structural Balance
标题：GraphSB：通过结构平衡提高图形上的不平衡节点分类
链接：https://arxiv.org/abs/2511.10022

【3】Towards Multiple Missing Values-resistant Unsupervised Graph Anomaly Detection
标题：抗多缺失值的无监督图异常检测
链接：https://arxiv.org/abs/2511.09917

【4】Probability-Biased Attention over Directed Bipartite Graphs for Long-Tail ICD Coding
标题：长尾ICD编码中有向二部图上的概率偏差注意
链接：https://arxiv.org/abs/2511.09559

【5】Beyond empirical models: Discovering new constitutive laws in solids with graph-based equation discovery
标题：超越经验模型：通过基于图形的方程发现发现固体中新的本构定律
链接：https://arxiv.org/abs/2511.09906

【6】TomoGraphView: 3D Medical Image Classification with Omnidirectional Slice Representations and Graph Neural Networks
标题：TomoGraphView：使用全方向切片表示和图神经网络的3D医学图像分类
链接：https://arxiv.org/abs/2511.09605

Transformer(3篇)

【1】Impact of Layer Norm on Memorization and Generalization in Transformers
标题：《Transformer》中分层规范对简化和概括的影响
链接：https://arxiv.org/abs/2511.10566

【2】Physics informed Transformer-VAE for biophysical parameter estimation: PROSAIL model inversion in Sentinel-2 imagery
标题：物理知识Transformer-VAE用于生物物理参数估计：Sentinel-2图像中的PROSAIL模型倒置
链接：https://arxiv.org/abs/2511.10387

【3】Heuristic Transformer: Belief Augmented In-Context Reinforcement Learning
标题：启发式Transformer：信念增强的上下文强化学习
链接：https://arxiv.org/abs/2511.10251

GAN|对抗|攻击|生成相关(9篇)

【1】Tight Robustness Certification through the Convex Hull of $\ell_0$ Attacks
链接：https://arxiv.org/abs/2511.10576

【2】Benchmarking Diversity in Image Generation via Attribute-Conditional Human Evaluation
标题：通过属性条件人类评估对图像生成中的多样性进行基准测试
链接：https://arxiv.org/abs/2511.10547

【3】EPO: Diverse and Realistic Protein Ensemble Generation via Energy Preference Optimization
标题：促红细胞生成：通过能量偏好优化产生多样化且现实的蛋白质群
链接 ：https://arxiv.org/abs/2511.10165

【4】eXIAA: eXplainable Injections for Adversarial Attack
标题：eXIAA：对抗性攻击的可扩展注射剂
链接：https://arxiv.org/abs/2511.10088

【5】Incremental Generation is Necessity and Sufficient for Universality in Flow-Based Modelling
标题：增量生成是基于流的建模普遍性的必要性和充分性
链接：https://arxiv.org/abs/2511.09902

【6】Simulator and Experience Enhanced Diffusion Model for Comprehensive ECG Generation
标题：模拟器和体验增强型扩散模型，用于全面的心电图生成
链接：https://arxiv.org/abs/2511.09895

【7】Hail to the Thief: Exploring Attacks and Defenses in Decentralised GRPO
标题：向小偷致敬：探索分散式GRPO中的攻击和防御
链接：https://arxiv.org/abs/2511.09780

【8】SEBA: Sample-Efficient Black-Box Attacks on Visual Reinforcement Learning
标题：SEBA：对视觉强化学习的样本高效黑匣子攻击
链接：https://arxiv.org/abs/2511.09681

【9】Edge Machine Learning for Cluster Counting in Next-Generation Drift Chambers
标题：下一代漂移室中用于簇计数的边缘机器学习
链接：https://arxiv.org/abs/2511.10540

半/弱/无/有监督|不确定性|主动学习(2篇)

【1】Improving Perturbation-based Explanations by Understanding the Role of Uncertainty Calibration
标题：通过了解不确定度校准的作用来改进基于微扰的解释
链接：https://arxiv.org/abs/2511.10439

【2】Torch-Uncertainty: A Deep Learning Framework for Uncertainty Quantification
标题：火炬不确定性：用于不确定性量化的深度学习框架
链接：https://arxiv.org/abs/2511.10282

迁移|Zero/Few/One-Shot|自适应(3篇)

【1】Panda: Test-Time Adaptation with Negative Data Augmentation
标题：Panda：通过负数据增强进行测试时适应
链接：https://arxiv.org/abs/2511.10481

【2】AdaptViG: Adaptive Vision GNN with Exponential Decay Gating
标题：AdaptViG：具有指数衰减门控的自适应视觉GNN
链接：https://arxiv.org/abs/2511.09942

【3】Brian Intensify: An Adaptive Machine Learning Framework for Auditory EEG Stimulation and Cognitive Enhancement in FXS
标题：Brian Intensify：FXS中用于听觉脑电刺激和认知增强的自适应机器学习框架
链接：https://arxiv.org/abs/2511.09765

强化学习(5篇)

【1】Towards Emotionally Intelligent and Responsible Reinforcement Learning
标题：迈向情感智能和负责任的强化学习
链接：https://arxiv.org/abs/2511.10573

【2】Causal Model-Based Reinforcement Learning for Sample-Efficient IoT Channel Access
标题：基于因果模型的强化学习用于样本高效的物联网通道访问
链接：https://arxiv.org/abs/2511.10291

【3】Improved Offline Reinforcement Learning via Quantum Metric Encoding
标题：通过量子度量编码改进离线强化学习
链接：https://arxiv.org/abs/2511.10187

【4】Optimistic Reinforcement Learning with Quantile Objectives
标题：具有分位数目标的乐观强化学习
链接：https://arxiv.org/abs/2511.09652

【5】Operator Models for Continuous-Time Offline Reinforcement Learning
标题：连续时间离线强化学习的操作员模型
链接：https://arxiv.org/abs/2511.10383

医学相关(3篇)

【1】EgoEMS: A High-Fidelity Multimodal Egocentric Dataset for Cognitive Assistance in Emergency Medical Services
标题：EgoEMS：用于紧急医疗服务认知辅助的高保真多模式自我中心数据集
链接：https://arxiv.org/abs/2511.09894

【2】A Fourier-Based Global Denoising Model for Smart Artifacts Removing of Microscopy Images
标题：基于傅里叶的显微图像智能伪影去除全局去噪模型
链接：https://arxiv.org/abs/2511.09734

【3】Prostate-VarBench: A Benchmark with Interpretable TabNet Framework for Prostate Cancer Variant Classification
标题：前列腺-VarBench：具有可解释TabNet框架的前列腺癌变体分类基准
链接：https://arxiv.org/abs/2511.09576

蒸馏|知识提取(1篇)

【1】PRISM: Diversifying Dataset Distillation by Decoupling Architectural Priors
标题：PRISM：通过解耦架构先验实现多样化的数据集蒸馏
链接：https://arxiv.org/abs/2511.09905

聚类(3篇)

【1】Enhancing Kernel Power K-means: Scalable and Robust Clustering with Random Fourier Features and Possibilistic Method
标题：增强核功率K-means：使用随机傅里叶特征和可能性方法的可扩展和鲁棒性集群
链接：https://arxiv.org/abs/2511.10392

【2】A General Anchor-Based Framework for Scalable Fair Clustering
标题：基于锚点的通用可扩展公平集群框架
链接：https://arxiv.org/abs/2511.09889

【3】Koopman Invariants as Drivers of Emergent Time-Series Clustering in Joint-Embedding Predictive Architectures
标题：Koopman不变量作为联合嵌入预测架构中紧急时间序列集群的驱动器
链接：https://arxiv.org/abs/2511.09783

超分辨率|去噪|去模糊|去雾(2篇)

【1】OpenSR-SRGAN: A Flexible Super-Resolution Framework for Multispectral Earth Observation Data
标题：OpenSR-SRGAN：多光谱地球观测数据的灵活超分辨率框架
链接：https://arxiv.org/abs/2511.10461

【2】DenoGrad: Deep Gradient Denoising Framework for Enhancing the Performance of Interpretable AI Models
标题：DenoGrad：用于增强可解释人工智能模型性能的深度梯度去噪框架
链接：https://arxiv.org/abs/2511.10161

联邦学习|隐私保护|加密(3篇)

【1】FedCure: Mitigating Participation Bias in Semi-Asynchronous Federated Learning with Non-IID Data
标题：FedCure：缓解使用非IID数据的半同步联邦学习中的参与偏差
链接：https://arxiv.org/abs/2511.10227

【2】SMoFi: Step-wise Momentum Fusion for Split Federated Learning on Heterogeneous Data
标题：SMoFi：用于异类数据上的分离联邦学习的分步动量融合
链接：https://arxiv.org/abs/2511.09828

【3】Data Heterogeneity and Forgotten Labels in Split Federated Learning
标题：分裂联邦学习中的数据异构和遗忘标签
链接：https://arxiv.org/abs/2511.09736

推理|分析|理解|解释(7篇)

【1】Generalizing Analogical Inference from Boolean to Continuous Domains
标题：类比推理从布尔域推广到连续域
链接：https://arxiv.org/abs/2511.10416

【2】BuddyMoE: Exploiting Expert Redundancy to Accelerate Memory-Constrained Mixture-of-Experts Inference
标题：BuddyMoE：利用专家冗余来加速记忆限制的混合专家推理
链接：https://arxiv.org/abs/2511.10054

【3】ConSurv: Multimodal Continual Learning for Survival Analysis
标题：ConSurv：用于生存分析的多模式持续学习
链接：https://arxiv.org/abs/2511.09853

【4】Privacy-Preserving Explainable AIoT Application via SHAP Entropy Regularization
标题：通过SHAP熵正规化保护隐私的可解释AIoT应用
链接：https://arxiv.org/abs/2511.09775

【5】Theory and computation for structured variational inference
标题：结构化变分推理的理论与计算
链接：https://arxiv.org/abs/2511.09897

【6】Lithological Controls on the Permeability of Geologic Faults: Surrogate Modeling and Sensitivity Analysis
标题：岩石对地质断层渗透率的控制：替代建模和敏感性分析
链接：https://arxiv.org/abs/2511.09674

【7】Analysis of the TAIGA-HiSCORE Data Using the Latent Space of Autoencoders
标题：利用自编码器的潜在空间分析TAIGA-HiSCORE数据
链接：https://arxiv.org/abs/2511.09655

检测相关(3篇)

【1】Revisiting Evaluation of Deep Neural Networks for Pedestrian Detection
标题：重新审视用于行人检测的深度神经网络评估
链接：https://arxiv.org/abs/2511.10308

【2】Fault Detection in Solar Thermal Systems using Probabilistic Reconstructions
标题：使用概率重建的太阳能热力系统故障检测
链接：https://arxiv.org/abs/2511.10296

【3】Out-of-Context Misinformation Detection via Variational Domain-Invariant Learning with Test-Time Training
标题：通过测试时训练的变分域不变学习进行上下文外错误信息检测
链接：https://arxiv.org/abs/2511.10213

分类|识别(5篇)

【1】How does My Model Fail? Automatic Identification and Interpretation of Physical Plausibility Failure Modes with Matryoshka Transcoders
标题：我的模型如何失败？使用Matryoshka代码转换器自动识别和解释物理合理性故障模式
链接：https://arxiv.org/abs/2511.10094

【2】A Novel Data-Dependent Learning Paradigm for Large Hypothesis Classes
标题：大型假设类的新型数据依赖学习范式
链接：https://arxiv.org/abs/2511.09996

【3】Constrained Best Arm Identification with Tests for Feasibility
标题：带可行性测试的约束最佳手臂识别
链接：https://arxiv.org/abs/2511.09808

【4】CaReTS: A Multi-Task Framework Unifying Classification and Regression for Time Series Forecasting
标题：CaReTS：统一时间序列预测分类和回归的多任务框架
链接：https://arxiv.org/abs/2511.09789

【5】NeuroLingua: A Language-Inspired Hierarchical Framework for Multimodal Sleep Stage Classification Using EEG and EOG
标题：NeuronLingua：一个受启发的分层框架，使用脑电和EOG进行多模式睡眠阶段分类
链接：https://arxiv.org/abs/2511.09773

表征(2篇)

【1】Efficient Hyperdimensional Computing with Modular Composite Representations
标题：具有模块化复合表示的高效超维计算
链接：https://arxiv.org/abs/2511.09708

【2】DynamicRTL: RTL Representation Learning for Dynamic Circuit Behavior
标题：DynamicRTL：动态电路行为的RTL表示学习
链接 ：https://arxiv.org/abs/2511.09593

优化|敛散性(7篇)

【1】Pretrained Joint Predictions for Scalable Batch Bayesian Optimization of Molecular Designs
标题：分子设计可扩展批量Bayesian优化的预训练联合预测
链接：https://arxiv.org/abs/2511.10590

【2】Opinion: Towards Unified Expressive Policy Optimization for Robust Robot Learning
标题：观点：迈向稳健的机器人学习的统一表达政策优化
链接：https://arxiv.org/abs/2511.10087

【3】Tree-Based Stochastic Optimization for Solving Large-Scale Urban Network Security Games
标题：基于树的随机优化解决大规模城市网络安全博弈
链接：https://arxiv.org/abs/2511.10072

【4】On the Convergence of Overparameterized Problems: Inherent Properties of the Compositional Structure of Neural Networks
标题：关于过度参数化问题的收敛：神经网络组成结构的固有性质
链接：https://arxiv.org/abs/2511.09810

【5】Parametric Expensive Multi-Objective Optimization via Generative Solution Modeling
标题：通过生成解建模的参数化昂贵多目标优化
链接：https://arxiv.org/abs/2511.09598

【6】HeatGen: A Guided Diffusion Framework for Multiphysics Heat Sink Design Optimization
标题：HeatGen：用于多物理场散热器设计优化的引导扩散框架
链接：https://arxiv.org/abs/2511.09578

【7】Global Convergence of Four-Layer Matrix Factorization under Random Initialization
标题：随机搜索下四层矩阵分解的全局收敛性
链接：https://arxiv.org/abs/2511.09925

预测|估计(11篇)

【1】Know Your Limits: Entropy Estimation Modeling for Compression and Generalization
标题：了解你的极限：用于压缩和概括的熵估计建模
链接：https://arxiv.org/abs/2511.10618

【2】Oya: Deep Learning for Accurate Global Precipitation Estimation
标题：Oya：深度学习可准确估计全球降水量
链接：https://arxiv.org/abs/2511.10562

【3】Weak Relation Enforcement for Kinematic-Informed Long-Term Stock Prediction with Artificial Neural Networks
标题：基于神经网络的动态信息股票长期预测的弱关系增强
链接：https://arxiv.org/abs/2511.10494

【4】PITE: Multi-Prototype Alignment for Individual Treatment Effect Estimation
标题：PITE：个体治疗效果估计的多原型对齐
链接：https://arxiv.org/abs/2511.10320

【5】Beyond MSE: Ordinal Cross-Entropy for Probabilistic Time Series Forecasting
标题：超越均方误差：概率时间序列预测的有序交叉熵
链接：https://arxiv.org/abs/2511.10200

【6】RI-Loss: A Learnable Residual-Informed Loss for Time Series Forecasting
标题：RI损失：时间序列预测的可学习剩余知情损失
链接：https://arxiv.org/abs/2511.10130

【7】AI-Integrated Decision Support System for Real-Time Market Growth Forecasting and Multi-Source Content Diffusion Analytics
标题：实时市场增长预测和多源内容扩散分析的人工智能集成决策支持系统
链接：https://arxiv.org/abs/2511.09962

【8】MDMLP-EIA: Multi-domain Dynamic MLPs with Energy Invariant Attention for Time Series Forecasting
标题：MDMLP-EIA：具有时间序列预测能量不变关注的多域动态MLP
链接：https://arxiv.org/abs/2511.09924

【9】History Rhymes: Macro-Contextual Retrieval for Robust Financial Forecasting
标题：历史押韵：稳健金融预测的宏观上下文检索
链接：https://arxiv.org/abs/2511.09754

【10】FlowCast: Advancing Precipitation Nowcasting with Conditional Flow Matching
标题：FlowCast：通过有条件流量匹配推进降水临近预报
链接：https://arxiv.org/abs/2511.09731

【11】Let the Experts Speak: Improving Survival Prediction & Calibration via Mixture-of-Experts Heads
标题：让专家说话：通过专家混合头改进生存预测和校准
链接：https://arxiv.org/abs/2511.09567

其他神经网络|深度学习|模型|建模(31篇)

【1】Robot Crash Course: Learning Soft and Stylized Falling
标题：机器人速成课程：柔软学习和风格化坠落
链接：https://arxiv.org/abs/2511.10635

【2】Semi-Unified Sparse Dictionary Learning with Learnable Top-K LISTA and FISTA Encoders
标题：使用可学习的Top-K LISTA和FISTA编码器进行半统一稀疏词典学习
链接：https://arxiv.org/abs/2511.10575

【3】Belief Net: A Filter-Based Framework for Learning Hidden Markov Models from Observations
标题：Belief Net：一个基于过滤器的隐马尔可夫模型学习框架
链接：https://arxiv.org/abs/2511.10571

【4】Maximizing Efficiency of Dataset Compression for Machine Learning Potentials With Information Theory
标题：利用信息论最大化数据集压缩效率以提高机器学习潜力
链接：https://arxiv.org/abs/2511.10561

【5】Intrinsic Dimensionality as a Model-Free Measure of Class Imbalance
标题：内在主观性作为阶级失衡的无模型衡量标准
链接：https://arxiv.org/abs/2511.10475

【6】Neuronal Fluctuations: Learning Rates vs Participating Neurons
标题：神经元波动：学习率与参与神经元
链接：https://arxiv.org/abs/2511.10435

【7】SHRUG-FM: Reliability-Aware Foundation Models for Earth Observation
标题：SHRUG-FM：可靠性感知的地球观测基础模型
链接：https://arxiv.org/abs/2511.10370

【8】Product distribution learning with imperfect advice
标题：建议不完善的产品分销学习
链接：https://arxiv.org/abs/2511.10366

【9】Gradient Flow Equations for Deep Linear Neural Networks: A Survey from a Network Perspective
标题：深度线性神经网络的梯度流方程：网络角度的综述
链接：https://arxiv.org/abs/2511.10362

【10】T2IBias: Uncovering Societal Bias Encoded in the Latent Space of Text-to-Image Generative Models
标题：T2 IBias：揭露文本到图像生成模型潜在空间中编码的社会偏见
链接：https://arxiv.org/abs/2511.10089

【11】Physics-informed Machine Learning for Static Friction Modeling in Robotic Manipulators Based on Kolmogorov-Arnold Networks
标题：基于Kolmogorov-Arnold网络的机器人机械手静摩擦建模物理信息机器学习
链接：https://arxiv.org/abs/2511.10079

【12】Temporal Latent Variable Structural Causal Model for Causal Discovery under External Interferences
标题：外部干扰下因果发现的时间潜在变量结构原因模型
链接：https://arxiv.org/abs/2511.10031

【13】SVD-NO: Learning PDE Solution Operators with SVD Integral Kernels
标题：SVD-NO：使用DDD积分核学习PCE解运算符
链接：https://arxiv.org/abs/2511.10025

【14】Towards Robust Multimodal Learning in the Open World
标题：在开放世界中迈向稳健的多模式学习
链接：https://arxiv.org/abs/2511.09989

【15】MultiTab: A Scalable Foundation for Multitask Learning on Tabular Data
标题：MultiTab：表格数据多任务学习的可扩展基础
链接：https://arxiv.org/abs/2511.09970

【16】Explore and Establish Synergistic Effects Between Weight Pruning and Coreset Selection in Neural Network Training
标题：探索并建立神经网络训练中权重修剪与核心集选择之间的协同效应
链接：https://arxiv.org/abs/2511.09901

【17】Expandable and Differentiable Dual Memories with Orthogonal Regularization for Exemplar-free Continual Learning
标题：可扩展和可区分的双存储器，具有垂直正规化，用于无示例的连续学习
链接：https://arxiv.org/abs/2511.09871

【18】Learning Intersections of Halfspaces under Factorizable Distribution
标题：可分解分布下半空间的学习交
链接：https://arxiv.org/abs/2511.09832

【19】Beyond Monotonicity: Revisiting Factorization Principles in Multi-Agent Q-Learning
标题：超越单调性：重新审视多智能体Q学习中的因子分解原则
链接：https://arxiv.org/abs/2511.09792

【20】ProbLog4Fairness: A Neurosymbolic Approach to Modeling and Mitigating Bias
标题：Problog 4公平性：建模和缓解偏见的神经符号方法
链接：https://arxiv.org/abs/2511.09768

【21】Gradient-Guided Exploration of Generative Model's Latent Space for Controlled Iris Image Augmentations
标题：受试者引导的生成模型潜在空间探索受控虹膜图像增强
链接：https://arxiv.org/abs/2511.09749

【22】TawPipe: Topology-Aware Weight Pipeline Parallelism for Accelerating Long-Context Large Models Training
标题：TawPipe：用于加速长上下文大型模型训练的具有布局意识的权重管道并行主义
链接：https://arxiv.org/abs/2511.09741

【23】Boosted GFlowNets: Improving Exploration via Sequential Learning
标题：增强的GFlowNets：通过顺序学习改善探索
链接：https://arxiv.org/abs/2511.09677

【24】PriVi: Towards A General-Purpose Video Model For Primate Behavior In The Wild
标题：PriVi：开发野外灵长类动物行为的通用视频模型
链接：https://arxiv.org/abs/2511.09675

【25】GEM+: Scalable State-of-the-Art Private Synthetic Data with Generator Networks
标题：GEM+：具有生成器网络的可扩展最先进的私有合成数据
链接：https://arxiv.org/abs/2511.09672

【26】Generalization Can Emerge in Tabular Foundation Models From a Single Table
标题：泛化可以从单个表中出现在表格基础模型中
链接：https://arxiv.org/abs/2511.09665

【27】Filtering Jump Markov Systems with Partially Known Dynamics: A Model-Based Deep Learning Approach
标题：过滤具有部分已知动态的跳跃马尔科夫系统：基于模型的深度学习方法
链接：https://arxiv.org/abs/2511.09569

【28】Completion of partial structures using Patterson maps with the CrysFormer machine learning model
标题：使用Patterson地图和CrysFormer机器学习模型完成部分结构
链接：https://arxiv.org/abs/2511.10440

【29】Symmetry aware Reynolds Averaged Navier Stokes turbulence models with equivariant neural networks
标题：具有对称性的Reynolds平均Navier Stokes湍流模型和等变神经网络
链接：https://arxiv.org/abs/2511.09769

【30】Masked Mineral Modeling: Continent-Scale Mineral Prospecting via Geospatial Infilling
标题：掩蔽矿物建模：通过地理空间填充进行大陆规模矿产勘探
链接：https://arxiv.org/abs/2511.09722

【31】Siegel Neural Networks
标题：西格尔神经网络
链接：https://arxiv.org/abs/2511.09577

其他(37篇)

【1】Querying Labeled Time Series Data with Scenario Programs
标题：使用场景程序查询已标记的时间序列数据
链接：https://arxiv.org/abs/2511.10627

【2】Algorithm Design and Stronger Guarantees for the Improving Multi-Armed Bandits Problem
标题：改进多臂强盗问题的算法设计和更强的保证
链接：https://arxiv.org/abs/2511.10619

【3】Multitask GLocal OBIA-Mamba for Sentinel-2 Landcover Mapping
标题：多任务GLocal OBIA-Mamba用于Sentinel-2土地覆盖制图
链接：https://arxiv.org/abs/2511.10604

【4】Bi-Level Contextual Bandits for Individualized Resource Allocation under Delayed Feedback
标题：延迟反馈下个性化资源分配的双层上下文盗贼
链接：https://arxiv.org/abs/2511.10572

【5】Two Americas of Well-Being: Divergent Rural-Urban Patterns of Life Satisfaction and Happiness from 2.6 B Social Media Posts
标题：两个幸福的美洲：来自2.6 B社交媒体帖子的不同城乡生活满意度和幸福感模式
链接：https://arxiv.org/abs/2511.10542

【6】Holonorm
标题：霍洛尼茨
链接：https://arxiv.org/abs/2511.10504

【7】Don't Waste It: Guiding Generative Recommenders with Structured Human Priors via Multi-head Decoding
标题：不要浪费它：通过多头解码用结构化的人类先验引导生成性推荐者
链接：https://arxiv.org/abs/2511.10492

【8】AgentEvolver: Towards Efficient Self-Evolving Agent System
标题：AgentEvolver：迈向高效的自我进化代理系统
链接：https://arxiv.org/abs/2511.10395

【9】Robust Decentralized Multi-armed Bandits: From Corruption-Resilience to Byzantine-Resilience
标题：强大的分散多武装强盗：从腐败韧性到拜占庭韧性
链接：https://arxiv.org/abs/2511.10344

【10】Unitho: A Unified Multi-Task Framework for Computational Lithography
标题：Unitho：计算平板印刷的统一多任务框架
链接：https://arxiv.org/abs/2511.10255

【11】Fractional neural attention for efficient multiscale sequence processing
标题：分数神经注意力用于高效的多尺度序列处理
链接：https://arxiv.org/abs/2511.10208

【12】Towards Leveraging Sequential Structure in Animal Vocalizations
标题：利用动物发声中的序列结构
链接：https://arxiv.org/abs/2511.10190

【13】Generalizing to Unseen Disaster Events: A Causal View
标题：概括为看不见的灾难事件：因果观
链接：https://arxiv.org/abs/2511.10120

【14】FAQNAS: FLOPs-aware Hybrid Quantum Neural Architecture Search using Genetic Algorithm
标题：FAQNAS：使用遗传算法的FLOPs感知混合量子神经架构搜索
链接：https://arxiv.org/abs/2511.10062

【15】From Static Structures to Ensembles: Studying and Harnessing Protein Structure Tokenization
标题：从静态结构到整体：研究和利用蛋白质结构代币化
链接：https://arxiv.org/abs/2511.10056

【16】Multi-agent In-context Coordination via Decentralized Memory Retrieval
标题：通过去中心化记忆检索的多智能体上下文协调
链接：https://arxiv.org/abs/2511.10030

【17】Interaction as Interference: A Quantum-Inspired Aggregation Approach
标题：相互作用即干扰：量子启发的聚集方法
链接：https://arxiv.org/abs/2511.10018

【18】The Role of Advanced Computer Architectures in Accelerating Artificial Intelligence Workloads
标题：先进计算机架构在加速人工智能工作负载方面的作用
链接：https://arxiv.org/abs/2511.10010

【19】Rediscovering the Lunar Equation of the Centre with AI Feynman via Embedded Physical Biases
标题：通过嵌入式物理偏差与AI Feynman重新发现中心的月球方程
链接：https://arxiv.org/abs/2511.09979

【20】Autonomous Concept Drift Threshold Determination
标题：自主概念漂移阈值确定
链接：https://arxiv.org/abs/2511.09953

【21】Harnessing Bounded-Support Evolution Strategies for Policy Refinement
标题：利用有界支持进化策略进行政策细化
链接：https://arxiv.org/abs/2511.09923

【22】Steering Pretrained Drafters during Speculative Decoding
标题：在推测解码期间指导经过预先训练的起草者
链接：https://arxiv.org/abs/2511.09844

【23】A Robust Task-Level Control Architecture for Learned Dynamical Systems
标题：学习动态系统的鲁棒任务级控制架构
链接：https://arxiv.org/abs/2511.09790

【24】Is nasty noise actually harder than malicious noise?
标题：令人讨厌的噪音实际上比恶意的噪音更难吗？
链接：https://arxiv.org/abs/2511.09763

【25】Assessing the Applicability of Natural Language Processing to Traditional Social Science Methodology: A Case Study in Identifying Strategic Signaling Patterns in Presidential Directives
标题：评估自然语言处理对传统社会科学方法的适用性：识别总统指令中战略信号模式的案例研究
链接：https://arxiv.org/abs/2511.09738

【26】Out-of-Distribution Generalization with a SPARC: Racing 100 Unseen Vehicles with a Single Policy
标题：带收件箱的分销外概括：用单一政策赛车100辆隐形车辆
链接：https://arxiv.org/abs/2511.09737

【27】Generalizing PDE Emulation with Equation-Aware Neural Operators
标题：使用方程感知神经运算符进行广义的偏出方程模拟
链接：https://arxiv.org/abs/2511.09729

【28】Baby Sophia: A Developmental Approach to Self-Exploration through Self-Touch and Hand Regard
标题：婴儿索菲亚：通过自我触摸和手注视进行自我探索的发展方法
链接：https://arxiv.org/abs/2511.09727

【29】Classifying Phonotrauma Severity from Vocal Fold Images with Soft Ordinal Regression
标题：用软有序回归对喉音图像中的音素创伤严重程度进行分类
链接：https://arxiv.org/abs/2511.09702

【30】Making Every Head Count: Sparse Attention Without the Speed-Performance Trade-off
标题：让每一个人都计数：稀疏的注意力，而不需要权衡速度和性能
链接：https://arxiv.org/abs/2511.09596

【31】Group Averaging for Physics Applications: Accuracy Improvements at Zero Training Cost
标题：物理应用的小组平均化：以零训练成本提高准确性
链接：https://arxiv.org/abs/2511.09573

【32】SynthTools: A Framework for Scaling Synthetic Tools for Agent Development
标题：SynthTools：扩展代理开发合成工具的框架
链接：https://arxiv.org/abs/2511.09572

【33】Global Solutions to Non-Convex Functional Constrained Problems with Hidden Convexity
标题：具有隐性凸性的非凸函数约束问题的整体解
链接：https://arxiv.org/abs/2511.10626

【34】Continuum Dropout for Neural Differential Equations
标题：神经元微分方程的连续丢弃
链接：https://arxiv.org/abs/2511.10446

【35】Generalized infinite dimensional Alpha-Procrustes based geometries
标题：广义无限维Alpha-Procrustes几何
链接：https://arxiv.org/abs/2511.09801

【36】Modelos Empiricos de Pos-Dupla Selecao por LASSO: Discussoes para Estudos do Transporte Aereo
标题：模式Empiricos de Pos-Dupla Selecao por LANSO：讨论航空运输研究
链接：https://arxiv.org/abs/2511.09767

【37】The Data Fusion Labeler (dFL): Challenges and Solutions to Data Harmonization, Labeling, and Provenance in Fusion Energy
标题：数据融合标签器（dFL）：融合能源中数据协调、标签和起源的挑战和解决方案
链接：https://arxiv.org/abs/2511.09725

机器翻译由腾讯交互翻译提供，仅供参考

点击“阅读原文”获取带摘要的学术速递