机器学习学术速递[7.15]

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284篇

大模型相关(25篇)

【1】Fusing LLM Capabilities with Routing Data
标题：将LLM功能与路由数据融合
链接：https://arxiv.org/abs/2507.10540

【2】Response Wide Shut? Surprising Observations in Basic Vision Language Model Capabilities
标题：响应大范围关闭？基本视觉语言模型功能的惊人观察
链接：https://arxiv.org/abs/2507.10442

【3】Leveraging RAG-LLMs for Urban Mobility Simulation and Analysis
标题：利用RAG-LLM进行城市移动模拟和分析
链接：https://arxiv.org/abs/2507.10382

【4】Conditional Chemical Language Models are Versatile Tools in Drug Discovery
标题：条件化学语言模型是药物发现中的通用工具
链接：https://arxiv.org/abs/2507.10273

【5】Pimba: A Processing-in-Memory Acceleration for Post-Transformer Large Language Model Serving
标题：Pimba：一种用于Post-Transformer大型语言模型服务的内存处理加速
链接：https://arxiv.org/abs/2507.10178

【6】ElasticMM: Efficient Multimodal LLMs Serving with Elastic Multimodal Parallelism
标题：ElasticMM：高效的多模态LLM，支持弹性多模态并行性
链接：https://arxiv.org/abs/2507.10069

【7】Towards Applying Large Language Models to Complement Single-Cell Foundation Models
标题：应用大型语言模型来补充单细胞基础模型
链接：https://arxiv.org/abs/2507.10039

【8】Memorization Sinks: Isolating Memorization during LLM Training
标题：再同步下沉：在LLM训练期间隔离再同步
链接：https://arxiv.org/abs/2507.09937

【9】Mechanistic Interpretability of LoRA-Adapted Language Models for Nuclear Reactor Safety Applications
标题：核反应堆安全应用的LoRA自适应语言模型的机制解释性
链接：https://arxiv.org/abs/2507.09931

【10】Through the River: Understanding the Benefit of Schedule-Free Methods for Language Model Training
标题：过河：了解语言模型训练的无日程方法的好处
链接：https://arxiv.org/abs/2507.09846

【11】Rethinking Prompt Optimization: Reinforcement, Diversification, and Migration in Blackbox LLMs
标题：重新思考即时优化：Blackbox LLM中的强化、多元化和迁移
链接：https://arxiv.org/abs/2507.09839

【12】Your Pretrained Model Tells the Difficulty Itself: A Self-Adaptive Curriculum Learning Paradigm for Natural Language Understanding
标题：您的预训练模型本身就说明了困难：自然语言理解的自适应课程学习范式
链接：https://arxiv.org/abs/2507.09758

【13】Large Language Models Encode Semantics in Low-Dimensional Linear Subspaces
标题：大型语言模型在低维线性子空间中编码语义
链接：https://arxiv.org/abs/2507.09709

【14】Calibrated and Robust Foundation Models for Vision-Language and Medical Image Tasks Under Distribution Shift
标题：分布转移下视觉语言和医学图像任务的校准且稳健的基础模型
链接：https://arxiv.org/abs/2507.09222

【15】Detecting and Pruning Prominent but Detrimental Neurons in Large Language Models
标题：在大型语言模型中检测和修剪突出但有害的神经元
链接：https://arxiv.org/abs/2507.09185

【16】Tactile-VLA: Unlocking Vision-Language-Action Model's Physical Knowledge for Tactile Generalization
标题：触觉-VLA：解锁视觉-语言-动作模型的物理知识以实现触觉概括
链接：https://arxiv.org/abs/2507.09160

【17】HedraRAG: Coordinating LLM Generation and Database Retrieval in Heterogeneous RAG Serving
标题：HedraRAG：协调异类RAG服务中的LLM生成和数据库检索
链接：https://arxiv.org/abs/2507.09138

【18】Lizard: An Efficient Linearization Framework for Large Language Models
标题：LSYS：大型语言模型的高效线性化框架
链接：https://arxiv.org/abs/2507.09025

【19】On Evaluating Performance of LLM Inference Serving Systems
标题：LLM推理服务系统性能评价研究
链接：https://arxiv.org/abs/2507.09019

【20】PRISM: Reducing Spurious Implicit Biases in Vision-Language Models with LLM-Guided Embedding Projection
标题：PRism：利用LLM引导的嵌入投影减少视觉语言模型中的虚假内隐偏差
链接：https://arxiv.org/abs/2507.08979

【21】How to Train a Leader: Hierarchical Reasoning in Multi-Agent LLMs
标题：如何训练领导者：多代理LL M中的分层推理
链接：https://arxiv.org/abs/2507.08960

【22】ODIA: Oriented Distillation for Inline Acceleration of LLM-based Function Calling
标题：ODIA：基于LLM的函数调用的内联加速的定向蒸馏
链接：https://arxiv.org/abs/2507.08877

【23】Contrastive Language-Image Pre-Training Model based Semantic Communication Performance Optimization
标题：基于对比图-图像预训练模型的语义通信性能优化
链接：https://arxiv.org/abs/2507.08873

【24】wd1: Weighted Policy Optimization for Reasoning in Diffusion Language Models
标题：wd 1：扩散语言模型中推理的加权策略优化
链接：https://arxiv.org/abs/2507.08838

【25】VIP: Visual Information Protection through Adversarial Attacks on Vision-Language Models
标题：VIP：通过对视觉语言模型的对抗性攻击保护视觉信息
链接：https://arxiv.org/abs/2507.08982

Graph相关(图学习|图神经网络|图优化等)(16篇)

【1】Graph World Model
标题：图表世界模型
链接：https://arxiv.org/abs/2507.10539

【2】A Graph Sufficiency Perspective for Neural Networks
标题：神经网络的图充分性观点
链接：https://arxiv.org/abs/2507.10215

【3】T-GRAB: A Synthetic Diagnostic Benchmark for Learning on Temporal Graphs
标题：T-GRAB：一个用于时态图学习的综合诊断基准
链接：https://arxiv.org/abs/2507.10183

【4】Large-Scale Graph Building in Dynamic Environments: Low Latency and High Quality
标题：动态环境中的大规模图形构建：低延迟和高质量
链接：https://arxiv.org/abs/2507.10139

【5】Wavelet-Enhanced Neural ODE and Graph Attention for Interpretable Energy Forecasting
标题：基于小波增强的神经网络常微分方程和图形注意力的可解释能源预测
链接：https://arxiv.org/abs/2507.10132

【6】Forecasting Coccidioidomycosis (Valley Fever) in Arizona: A Graph Neural Network Approach
标题：预测亚利桑那州的球孢子菌病（山谷热）：图形神经网络方法
链接：https://arxiv.org/abs/2507.10014

【7】Hierarchical Job Classification with Similarity Graph Integration
标题：具有相似度图集成的分层职位分类
链接：https://arxiv.org/abs/2507.09949

【8】Soft Graph Clustering for single-cell RNA Sequencing Data
标题：单细胞RNA测序数据的软图聚集
链接：https://arxiv.org/abs/2507.09890

【9】Federated Learning with Graph-Based Aggregation for Traffic Forecasting
标题：基于图的聚合的联邦学习用于交通预测
链接：https://arxiv.org/abs/2507.09805

【10】Toward accurate RUL and SOH estimation using reinforced graph-based PINNs enhanced with dynamic weights
标题：使用动态权重增强的基于图的增强PINN实现准确的RUL和SOH估计
链接：https://arxiv.org/abs/2507.09766

【11】Geometric Generative Modeling with Noise-Conditioned Graph Networks
标题：具有噪音条件图网络的几何生成建模
链接：https://arxiv.org/abs/2507.09391

【12】Towards Interpretable Drug-Drug Interaction Prediction: A Graph-Based Approach with Molecular and Network-Level Explanations
标题：迈向可解释的药物相互作用预测：基于图形的方法，具有分子和网络级解释
链接：https://arxiv.org/abs/2507.09173

【13】Heterogeneous Graph Prompt Learning via Adaptive Weight Pruning
标题：通过自适应权重修剪的异类图提示学习
链接：https://arxiv.org/abs/2507.09132

【14】Graph Neural Network Enhanced Sequential Recommendation Method for Cross-Platform Ad Campaign
标题：跨平台广告活动的图神经网络增强序列推荐方法
链接：https://arxiv.org/abs/2507.08959

【15】History Matching under Uncertainty of Geological Scenarios with Implicit Geological Realism Control with Generative Deep Learning and Graph Convolutions
标题：地质场景不确定性下的历史匹配，隐式地质现实主义控制，生成式深度学习和图形卷积
链接：https://arxiv.org/abs/2507.10201

【16】Signed Graph Learning: Algorithms and Theory
标题：符号图学习：算法与理论
链接：https://arxiv.org/abs/2507.09717

Transformer(12篇)

【1】Extracting Important Tokens in E-Commerce Queries with a Tag Interaction-Aware Transformer Model
标题：使用标签交互感知Transformer模型提取电子商务收件箱中的重要令牌
链接：https://arxiv.org/abs/2507.10385

【2】TAT: Temporal-Aligned Transformer for Multi-Horizon Peak Demand Forecasting
标题：STAT：用于多层面峰值需求预测的时间对齐Transformer
链接：https://arxiv.org/abs/2507.10349

【3】Should We Ever Prefer Decision Transformer for Offline Reinforcement Learning?
标题：我们应该更喜欢使用决策Transformer来进行离线强化学习吗？
链接：https://arxiv.org/abs/2507.10174

【4】Extracting Cause-Effect Pairs from a Sentence with a Dependency-Aware Transformer Model
标题：使用依赖性感知Transformer模型从句子中提取因果对
链接：https://arxiv.org/abs/2507.09925

【5】MLoRQ: Bridging Low-Rank and Quantization for Transformer Compression
标题：MLoRQ：桥接低秩和量化的Transformer压缩
链接：https://arxiv.org/abs/2507.09616

【6】Transformers Don't In-Context Learn Least Squares Regression
标题：Transformer不会在上下文中学习最小二乘回归
链接：https://arxiv.org/abs/2507.09440

【7】Adversarial Activation Patching: A Framework for Detecting and Mitigating Emergent Deception in Safety-Aligned Transformers
标题：对抗激活修补：检测和减轻安全调整Transformer中紧急欺骗的框架
链接：https://arxiv.org/abs/2507.09406

【8】TPP-SD: Accelerating Transformer Point Process Sampling with Speculative Decoding
标题：TPP-SD：用推测解码加速Transformer点过程采样
链接：https://arxiv.org/abs/2507.09252

【9】POIFormer: A Transformer-Based Framework for Accurate and Scalable Point-of-Interest Attribution
标题：POIFormer：一个基于转换器的框架，用于准确且可扩展的兴趣点归因
链接：https://arxiv.org/abs/2507.09137

【10】Domain-Adaptive Diagnosis of Lewy Body Disease with Transferability Aware Transformer
标题：利用可转移性感知Transformer对路易体病进行区域适应性诊断
链接：https://arxiv.org/abs/2507.08839

【11】Representation learning with a transformer by contrastive learning for money laundering detection
标题：基于对比学习的Transformer表示学习算法在洗钱检测中的应用
链接：https://arxiv.org/abs/2507.08835

【12】DepViT-CAD: Deployable Vision Transformer-Based Cancer Diagnosis in Histopathology
标题：DepViT-CAD：可部署的基于视觉变换器的组织学癌症诊断
链接：https://arxiv.org/abs/2507.10250

GAN|对抗|攻击|生成相关(11篇)

【1】Bridging Robustness and Generalization Against Word Substitution Attacks in NLP via the Growth Bound Matrix Approach
标题：通过增长界矩阵方法在NLP中弥合鲁棒性和通用性对抗词替换攻击
链接：https://arxiv.org/abs/2507.10330

【2】Learning Private Representations through Entropy-based Adversarial Training
标题：通过基于熵的对抗训练学习私人表示
链接：https://arxiv.org/abs/2507.10194

【3】Towards High Supervised Learning Utility Training Data Generation: Data Pruning and Column Reordering
标题：迈向高监督学习实用程序训练数据生成：数据修剪和列重新排序
链接：https://arxiv.org/abs/2507.10088

【4】Efficient Molecular Conformer Generation with SO(3)-Averaged Flow Matching and Reflow
标题：利用SO（3）平均流匹配和回流高效生成分子Conformer
链接：https://arxiv.org/abs/2507.09785

【5】Do we need equivariant models for molecule generation?
标题：我们需要等变模型来生成分子吗？
链接：https://arxiv.org/abs/2507.09753

【6】CAN-Trace Attack: Exploit CAN Messages to Uncover Driving Trajectories
标题：CAN跟踪攻击：利用CAN消息揭露驾驶轨迹
链接：https://arxiv.org/abs/2507.09624

【7】DRAGD: A Federated Unlearning Data Reconstruction Attack Based on Gradient Differences
标题：BRAWD：基于梯度差异的联邦取消学习数据重建攻击
链接：https://arxiv.org/abs/2507.09602

【8】La-Proteina: Atomistic Protein Generation via Partially Latent Flow Matching
标题：La-Proteina：通过部分潜伏流匹配产生原子蛋白质
链接：https://arxiv.org/abs/2507.09466

【9】Shortening the Trajectories: Identity-Aware Gaussian Approximation for Efficient 3D Molecular Generation
标题：缩短轨迹：识别高斯逼近以实现高效3D分子生成
链接：https://arxiv.org/abs/2507.09043

【10】Next-Generation Travel Demand Modeling with a Generative Framework for Household Activity Coordination
标题：具有家庭活动协调生成框架的下一代旅行需求建模
链接：https://arxiv.org/abs/2507.08871

【11】Recurrent Expansion: A Pathway Toward the Next Generation of Deep Learning
标题：循环扩展：迈向下一代深度学习的途径
链接：https://arxiv.org/abs/2507.08828

半/弱/无/有监督|不确定性|主动学习(7篇)

【1】Self-supervised Learning on Camera Trap Footage Yields a Strong Universal Face Embedder
标题：相机陷阱镜头上的自监督学习产生了强大的通用人脸嵌入器
链接：https://arxiv.org/abs/2507.10552

【2】CLA: Latent Alignment for Online Continual Self-Supervised Learning
标题：CLA：在线持续自我监督学习的潜在一致
链接：https://arxiv.org/abs/2507.10434

【3】Enhancing ALS Progression Tracking with Semi-Supervised ALSFRS-R Scores Estimated from Ambient Home Health Monitoring
标题：通过从周围家庭健康监测中估计的半监督ALSFRS-R分数来增强ALS进展跟踪
链接：https://arxiv.org/abs/2507.09460

【4】Impute With Confidence: A Framework for Uncertainty Aware Multivariate Time Series Imputation
标题：充满信心的指责：意识不确定性的多元时间序列插补框架
链接：https://arxiv.org/abs/2507.09353

【5】Multimodal Cardiovascular Risk Profiling Using Self-Supervised Learning of Polysomnography
标题：使用多导睡眠图的自我监督学习进行多模式心血管风险分析
链接：https://arxiv.org/abs/2507.09009

【6】Simulation as Supervision: Mechanistic Pretraining for Scientific Discovery
标题：模拟作为监督：科学发现的机械预训练
链接：https://arxiv.org/abs/2507.08977

【7】Discovering Governing Equations in the Presence of Uncertainty
标题：在不确定性的情况下发现治理方程
链接：https://arxiv.org/abs/2507.09740

迁移|Zero/Few/One-Shot|自适应(21篇)

【1】Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation
标题：迭代混合：学习动态迭代深度以实现自适应令牌级计算
链接：https://arxiv.org/abs/2507.10524

【2】RAPNet: A Receptive-Field Adaptive Convolutional Neural Network for Pansharpening
标题：RAPNet：一种用于Panshering的感受场自适应卷积神经网络
链接：https://arxiv.org/abs/2507.10461

【3】Efficient Federated Learning with Heterogeneous Data and Adaptive Dropout
标题：具有异类数据和自适应辍学的高效联邦学习
链接：https://arxiv.org/abs/2507.10430

【4】Kernel-Adaptive PI-ELMs for Forward and Inverse Problems in PDEs with Sharp Gradients
标题：具有尖锐影响的偏置方程中正向和逆问题的核自适应PI-ELM
链接：https://arxiv.org/abs/2507.10241

【5】Domain Borders Are There to Be Crossed With Federated Few-Shot Adaptation
标题：联邦Few-Shot改编将跨越领域边界
链接：https://arxiv.org/abs/2507.10160

【6】A Transfer Learning-Based Method for Water Body Segmentation in Remote Sensing Imagery: A Case Study of the Zhada Tulin Area
标题：基于迁移学习的遥感影像水体分割方法--以札达图林地区为例
链接：https://arxiv.org/abs/2507.10084

【7】Bridging Neural Networks and Dynamic Time Warping for Adaptive Time Series Classification
标题：自适应时间序列分类的神经网络和动态时间扭曲的桥梁
链接：https://arxiv.org/abs/2507.09826

【8】NegRefine: Refining Negative Label-Based Zero-Shot OOD Detection
标题：NegRefine：完善基于负标签的Zero-ShotOOD检测
链接：https://arxiv.org/abs/2507.09795

【9】Energy Dissipation Rate Guided Adaptive Sampling for Physics-Informed Neural Networks: Resolving Surface-Bulk Dynamics in Allen-Cahn Systems
标题：物理信息神经网络的能量消耗率引导自适应采样：解析艾伦-卡恩系统中的表面-体动力学
链接：https://arxiv.org/abs/2507.09757

【10】VDInstruct: Zero-Shot Key Information Extraction via Content-Aware Vision Tokenization
标题：VDDirecct：通过内容感知视觉令牌化的Zero-Shot关键信息提取
链接：https://arxiv.org/abs/2507.09531

【11】Domain Adaptation and Multi-view Attention for Learnable Landmark Tracking with Sparse Data
标题：稀疏数据可学习地标跟踪的领域适应和多视角关注
链接：https://arxiv.org/abs/2507.09420

【12】Real-Time Adaptive Motion Planning via Point Cloud-Guided, Energy-Based Diffusion and Potential Fields
标题：通过点云引导、基于能量的扩散和势场的实时自适应运动规划
链接：https://arxiv.org/abs/2507.09383

【13】ViT-ProtoNet for Few-Shot Image Classification: A Multi-Benchmark Evaluation
标题：用于Few-Shot图像分类的ViT-ProtoNet：多基准评估
链接：https://arxiv.org/abs/2507.09299

【14】Controllable Patching for Compute-Adaptive Surrogate Modeling of Partial Differential Equations
标题：偏微方程计算自适应代理建模的可控修补
链接：https://arxiv.org/abs/2507.09264

【15】Behavioral Exploration: Learning to Explore via In-Context Adaptation
标题：行为探索：学会通过上下文适应进行探索
链接：https://arxiv.org/abs/2507.09041

【16】DAFOS: Dynamic Adaptive Fanout Optimization Sampler
标题：DAFOS：动态自适应风扇优化采样器
链接：https://arxiv.org/abs/2507.08845

【17】Gradients as an Action: Towards Communication-Efficient Federated Recommender Systems via Adaptive Action Sharing
标题：作为行动的参与者：通过自适应行动共享实现通信高效的联邦推荐系统
链接：https://arxiv.org/abs/2507.08842

【18】Zero-Shot Neural Architecture Search with Weighted Response Correlation
标题：具有加权响应相关性的Zero-Shot神经架构搜索
链接：https://arxiv.org/abs/2507.08841

【19】An Enhanced Classification Method Based on Adaptive Multi-Scale Fusion for Long-tailed Multispectral Point Clouds
标题：基于自适应多尺度融合的长尾多光谱点云增强分类方法
链接：https://arxiv.org/abs/2412.11407

【20】Nesterov Finds GRAAL: Optimal and Adaptive Gradient Method for Convex Optimization
标题：Nesterov发现GRAAL：凸优化的最佳自适应梯度方法
链接：https://arxiv.org/abs/2507.09823

【21】A Generalization Theory for Zero-Shot Prediction
标题：一种推广的零炮预测理论
链接：https://arxiv.org/abs/2507.09128

强化学习(7篇)

【1】Reasoning or Memorization? Unreliable Results of Reinforcement Learning Due to Data Contamination
标题：推理还是精简？由于数据污染，强化学习的结果不可靠
链接：https://arxiv.org/abs/2507.10532

【2】Adaptability in Multi-Agent Reinforcement Learning: A Framework and Unified Review
标题：多智能体强化学习的适应性：框架和统一评论
链接：https://arxiv.org/abs/2507.10142

【3】Intersection of Reinforcement Learning and Bayesian Optimization for Intelligent Control of Industrial Processes: A Safe MPC-based DPG using Multi-Objective BO
标题：强化学习和Bayesian优化的交叉用于工业过程智能控制：使用多目标BO的安全基于MPC的DPG
链接：https://arxiv.org/abs/2507.09864

【4】Consistency Trajectory Planning: High-Quality and Efficient Trajectory Optimization for Offline Model-Based Reinforcement Learning
标题：一致性轨迹规划：基于模型的离线强化学习的高质量、高效的轨迹优化
链接：https://arxiv.org/abs/2507.09534

【5】Continual Reinforcement Learning by Planning with Online World Models
标题：通过使用在线世界模型进行规划进行持续强化学习
链接：https://arxiv.org/abs/2507.09177

【6】Deep Reinforcement Learning with Gradient Eligibility Traces
标题：具有梯度资格痕迹的深度强化学习
链接：https://arxiv.org/abs/2507.09087

【7】Assuring the Safety of Reinforcement Learning Components: AMLAS-RL
标题：确保强化学习组件的安全性：AMLAS-RL
链接：https://arxiv.org/abs/2507.08848

元学习(1篇)

【1】Meta-autoencoders: An approach to discovery and representation of relationships between dynamically evolving classes
标题：元自动编码器：一种动态演化类间关系的发现和表示方法
链接：https://arxiv.org/abs/2507.09362

分层学习(1篇)

【1】Hierarchical Abstraction Enables Human-Like 3D Object Recognition in Deep Learning Models
标题：分层抽象在深度学习模型中实现类人3D对象识别
链接：https://arxiv.org/abs/2507.09830

医学相关(7篇)

【1】Lightweight Model for Poultry Disease Detection from Fecal Images Using Multi-Color Space Feature Optimization and Machine Learning
标题：使用多色空间特征优化和机器学习从粪便图像中检测家禽疾病的轻量级模型
链接：https://arxiv.org/abs/2507.10056

【2】TolerantECG: A Foundation Model for Imperfect Electrocardiogram
标题：耐受性心电图：不完美心电图的基础模型
链接：https://arxiv.org/abs/2507.09887

【3】An Automated Classifier of Harmful Brain Activities for Clinical Usage Based on a Vision-Inspired Pre-trained Framework
标题：基于视觉启发的预训练框架的临床有害大脑活动自动分类器
链接：https://arxiv.org/abs/2507.08874

【4】A Brain Tumor Segmentation Method Based on CLIP and 3D U-Net with Cross-Modal Semantic Guidance and Multi-Level Feature Fusion
标题：基于CLIP和3D U-Net的跨模式语义引导和多层特征融合脑肿瘤分割方法
链接：https://arxiv.org/abs/2507.09966

【5】Advanced U-Net Architectures with CNN Backbones for Automated Lung Cancer Detection and Segmentation in Chest CT Images
标题：具有CNN主干的高级U-Net架构，用于胸部CT图像中的自动肺癌检测和分割
链接：https://arxiv.org/abs/2507.09898

【6】PanoDiff-SR: Synthesizing Dental Panoramic Radiographs using Diffusion and Super-resolution
标题：PanoDiff-SR：使用扩散和超分辨率合成牙科全景射线照片
链接：https://arxiv.org/abs/2507.09227

【7】Multi-omic Prognosis of Alzheimer's Disease with Asymmetric Cross-Modal Cross-Attention Network
标题：不对称跨模式交叉注意力网络对阿尔茨海默病的多组预后
链接：https://arxiv.org/abs/2507.08855

蒸馏|知识提取(1篇)

【1】Feature Distillation is the Better Choice for Model-Heterogeneous Federated Learning
标题：特征提取是模型异类联邦学习的更好选择
链接：https://arxiv.org/abs/2507.10348

推荐(4篇)

【1】Benchmarking and Evaluation of AI Models in Biology: Outcomes and Recommendations from the CZI Virtual Cells Workshop
标题：生物学中人工智能模型的基准和评估：CZZ虚拟细胞研讨会的成果和建议
链接：https://arxiv.org/abs/2507.10502

【2】Radial Neighborhood Smoothing Recommender System
标题：辐射邻近平滑推荐系统
链接：https://arxiv.org/abs/2507.09952

【3】Identifying Offline Metrics that Predict Online Impact: A Pragmatic Strategy for Real-World Recommender Systems
标题：识别预测在线影响的离线收件箱：现实世界推荐系统的实用策略
链接：https://arxiv.org/abs/2507.09566

【4】S2SRec2: Set-to-Set Recommendation for Basket Completion with Recipe
标题：S2 SRec2：用配方完成篮子的套对套推荐
链接：https://arxiv.org/abs/2507.09101

聚类(1篇)

【1】Average Sensitivity of Hierarchical $k$-Median Clustering
标题：分层$k$的平均敏感性-中位数聚集
链接：https://arxiv.org/abs/2507.10296

超分辨率|去噪|去模糊|去雾(1篇)

【1】Investigating the Robustness of Extreme Precipitation Super-Resolution Across Climates
标题：不同气候条件下极端降水超分辨率的稳健性研究
链接：https://arxiv.org/abs/2507.09166

自动驾驶|车辆|车道检测等(3篇)

【1】Multi-residual Mixture of Experts Learning for Cooperative Control in Multi-vehicle Systems
标题：基于多残差混合专家学习的多车系统协同控制
链接：https://arxiv.org/abs/2507.09836

【2】DAA*: Deep Angular A Star for Image-based Path Planning
标题：DAA*：Deep Angular A Star，基于图像的路径规划
链接：https://arxiv.org/abs/2507.09305

【3】On the Fragility of Multimodal Perception to Temporal Misalignment in Autonomous Driving
标题：自动驾驶中多模式感知对时间失调的脆弱性
链接：https://arxiv.org/abs/2507.09095

联邦学习|隐私保护|加密(2篇)

【1】MTF-Grasp: A Multi-tier Federated Learning Approach for Robotic Grasping
标题：MTF-Grasp：一种用于机器人抓取的多层联邦学习方法
链接：https://arxiv.org/abs/2507.10158

【2】Lightweight Federated Learning over Wireless Edge Networks
标题：无线边缘网络上的联合轻量级学习
链接：https://arxiv.org/abs/2507.09546

推理|分析|理解|解释(20篇)

【1】Improving Remote Sensing Classification using Topological Data Analysis and Convolutional Neural Networks
标题：利用布局数据分析和卷积神经网络改进遥感分类
链接：https://arxiv.org/abs/2507.10381

【2】MoCap-Impute: A Comprehensive Benchmark and Comparative Analysis of Imputation Methods for IMU-based Motion Capture Data
标题：MoCap-Impute：基于IMU的运动捕获数据的综合基准和插补方法的比较分析
链接：https://arxiv.org/abs/2507.10334

【3】Visual Analytics for Explainable and Trustworthy Artificial Intelligence
标题：用于可解释且值得信赖的人工智能的视觉分析
链接：https://arxiv.org/abs/2507.10240

【4】Understanding the Rank of Tensor Networks via an Intuitive Example-Driven Approach
标题：通过直观的示例驱动方法了解张量网络的排名
链接：https://arxiv.org/abs/2507.10170

【5】Analysis of AI Techniques for Orchestrating Edge-Cloud Application Migration
标题：描述边缘云应用程序迁移的人工智能技术分析
链接：https://arxiv.org/abs/2507.10119

【6】Explainable AI in Genomics: Transcription Factor Binding Site Prediction with Mixture of Experts
标题：基因组学中的可解释人工智能：混合专家预测转录因子结合位点
链接：https://arxiv.org/abs/2507.09754

【7】Holistix: A Dataset for Holistic Wellness Dimensions Analysis in Mental Health Narratives
标题：Holistix：心理健康叙事中整体健康维度分析的数据集
链接：https://arxiv.org/abs/2507.09565

【8】Assessing reliability of explanations in unbalanced datasets: a use-case on the occurrence of frost events
标题：评估不平衡数据集中解释的可靠性：霜冻事件发生的用例
链接：https://arxiv.org/abs/2507.09545

【9】An Analysis of Action-Value Temporal-Difference Methods That Learn State Values
标题：学习状态值的时间值时间差方法分析
链接：https://arxiv.org/abs/2507.09523

【10】MI CAM: Mutual Information Weighted Activation Mapping for Causal Visual Explanations of Convolutional Neural Networks
标题：MI CAM：卷积神经网络因果视觉推理的互信息加权激活映射
链接：https://arxiv.org/abs/2507.09092

【11】Infinite Video Understanding
标题：无限的视频理解
链接：https://arxiv.org/abs/2507.09068

【12】BrainLesion Suite: A Flexible and User-Friendly Framework for Modular Brain Lesion Image Analysis
标题：BrainLesion套件：一个灵活且用户友好的模块化脑损伤图像分析框架
链接：https://arxiv.org/abs/2507.09036

【13】Accuracy and Consumption analysis from a compressed model by CompactifAI from Multiverse Computing
标题：Multiverse Computing的CompacifAI对压缩模型进行准确性和消耗分析
链接：https://arxiv.org/abs/2507.08836

【14】Efficient Triple Modular Redundancy for Reliability Enhancement of DNNs Using Explainable AI
标题：使用可解释AI的DNN可靠性增强的高效三模块冗余
链接：https://arxiv.org/abs/2507.08829

【15】Think Clearly: Improving Reasoning via Redundant Token Pruning
标题：清晰思考：通过冗余令牌修剪改进推理
链接：https://arxiv.org/abs/2507.08806

【16】Lightweight Cloud Masking Models for On-Board Inference in Hyperspectral Imaging
标题：用于高光谱成像机载推理的轻量级云遮蔽模型
链接：https://arxiv.org/abs/2507.08052

【17】Regret Analysis of Posterior Sampling-Based Expected Improvement for Bayesian Optimization
标题：基于后验抽样的Bayesian优化预期改进的遗憾分析
链接：https://arxiv.org/abs/2507.09828

【18】Sensitivity Analysis of Transport and Radiation in NeuralPlasmaODE for ITER Burning Plasmas
标题：ITER燃烧等离子体神经等离子体传输和辐射敏感性分析
链接：https://arxiv.org/abs/2507.09432

【19】Predictive Causal Inference via Spatio-Temporal Modeling and Penalized Empirical Likelihood
标题：通过时空建模和惩罚经验可能性的预测性因果推理
链接：https://arxiv.org/abs/2507.08896

【20】Mind the Gap: Navigating Inference with Optimal Transport Maps
标题：注意差距：利用最佳交通地图进行推理
链接：https://arxiv.org/abs/2507.08867

检测相关(7篇)

【1】BenchReAD: A systematic benchmark for retinal anomaly detection
标题：BenchReAD：视网膜异常检测的系统基准
链接：https://arxiv.org/abs/2507.10492

【2】From BERT to Qwen: Hate Detection across architectures
标题：从BERT到Qwen：跨架构的仇恨检测
链接：https://arxiv.org/abs/2507.10468

【3】Evaluating Fake Music Detection Performance Under Audio Augmentations
标题：评估音频增强下的假音乐检测性能
链接：https://arxiv.org/abs/2507.10447

【4】DNS Tunneling: Threat Landscape and Improved Detection Solutions
标题：DNS隧道：威胁格局和改进的检测解决方案
链接：https://arxiv.org/abs/2507.10267

【5】Credit Card Fraud Detection Using RoFormer Model With Relative Distance Rotating Encoding
标题：使用RoFormer模型和相对距离旋转编码的信用卡欺诈检测
链接：https://arxiv.org/abs/2507.09385

【6】Advanced Health Misinformation Detection Through Hybrid CNN-LSTM Models Informed by the Elaboration Likelihood Model (ELM)
标题：通过由EQUALY模型（ELM）告知的混合CNN-LSTM模型进行高级健康错误信息检测
链接：https://arxiv.org/abs/2507.09149

【7】Fair-FLIP: Fair Deepfake Detection with Fairness-Oriented Final Layer Input Prioritising
标题：Fair-FLIP：具有面向公平的最终层输入优先级的公平Deepfake检测
链接：https://arxiv.org/abs/2507.08912

分类|识别(8篇)

【1】Recognizing Dementia from Neuropsychological Tests with State Space Models
标题：利用状态空间模型从神经心理学测试中识别痴呆症
链接：https://arxiv.org/abs/2507.10311

【2】Play Style Identification Using Low-Level Representations of Play Traces in MicroRTS
标题：使用MicroRTS中播放痕迹的低级表示进行播放风格识别
链接：https://arxiv.org/abs/2507.10172

【3】Long-Tailed Data Classification by Increasing and Decreasing Neurons During Training
标题：通过训练期间增加和减少神经元进行长尾数据分类
链接：https://arxiv.org/abs/2507.09940

【4】Post-Training Quantization of Generative and Discriminative LSTM Text Classifiers: A Study of Calibration, Class Balance, and Robustness
标题：生成性和区分性LSTM文本分类器的训练后量化：校准、类别平衡和稳健性研究
链接：https://arxiv.org/abs/2507.09687

【5】Theory-Informed Improvements to Classifier-Free Guidance for Discrete Diffusion Models
标题：离散扩散模型无分类器指导的理论改进
链接：https://arxiv.org/abs/2507.08965

【6】Simulating Biases for Interpretable Fairness in Offline and Online Classifiers
标题：离线和在线分类器中模拟可解释公平性的偏差
链接：https://arxiv.org/abs/2507.10154

【7】Conformation-Aware Structure Prediction of Antigen-Recognizing Immune Proteins
标题：抗原识别免疫蛋白的认知结构预测
链接：https://arxiv.org/abs/2507.09054

【8】Fixed-Confidence Multiple Change Point Identification under Bandit Feedback
标题：Bandit反馈下的固定置信度多变点识别
链接：https://arxiv.org/abs/2507.08994

表征(5篇)

【1】Text-Driven Causal Representation Learning for Source-Free Domain Generalization
标题：用于无源领域概括的文本驱动因果表示学习
链接：https://arxiv.org/abs/2507.09961

【2】Discrete Differential Principle for Continuous Smooth Function Representation
标题：连续光滑函数表示的离散微分原理
链接：https://arxiv.org/abs/2507.09480

【3】Fair CCA for Fair Representation Learning: An ADNI Study
标题：公平代表学习的公平国家评估：ADNI研究
链接：https://arxiv.org/abs/2507.09382

【4】Learning Diffusion Models with Flexible Representation Guidance
标题：具有灵活表示指导的学习扩散模型
链接：https://arxiv.org/abs/2507.08980

【5】WellPINN: Accurate Well Representation for Transient Fluid Pressure Diffusion in Subsurface Reservoirs with Physics-Informed Neural Networks
标题：WellPINN：利用物理信息神经网络准确描述地下储层中的瞬时流体压力扩散
链接：https://arxiv.org/abs/2507.09330

优化|敛散性(19篇)

【1】On the Performance of Differentially Private Optimization with Heavy-Tail Class Imbalance
标题：重尾类不平衡差分私有优化算法的性能研究
链接：https://arxiv.org/abs/2507.10536

【2】Some remarks on gradient dominance and LQR policy optimization
标题：关于梯度主导和LQR政策优化的一些看法
链接：https://arxiv.org/abs/2507.10452

【3】Non-exchangeable Conformal Prediction with Optimal Transport: Tackling Distribution Shifts with Unlabeled Data
标题：具有最优传输的不可交换保形预测：用未标记数据解决分布变化
链接：https://arxiv.org/abs/2507.10425

【4】Convergence of Agnostic Federated Averaging
标题：不可知联邦平均的收敛
链接：https://arxiv.org/abs/2507.10325

【5】A Variance-Reduced Cubic-Regularized Newton for Policy Optimization
标题：用于政策优化的方差缩减三次正规牛顿
链接：https://arxiv.org/abs/2507.10120

【6】Compression Method for Deep Diagonal State Space Model Based on $H^2$ Optimal Reduction
标题：基于$H ' 2 $最优约简的深对角状态空间模型压缩方法
链接：https://arxiv.org/abs/2507.10078

【7】PRISM: Fine-Grained Paper-to-Paper Retrieval with Multi-Aspect-Aware Query Optimization
标题：PRism：具有多方面感知查询优化的细粒度纸到纸检索
链接：https://arxiv.org/abs/2507.10057

【8】Memory-Efficient Personalization of Text-to-Image Diffusion Models via Selective Optimization Strategies
标题：通过选择性优化策略实现文本到图像扩散模型的内存高效个性化
链接：https://arxiv.org/abs/2507.10029

【9】Neural Two-Stage Stochastic Optimization for Solving Unit Commitment Problem
标题：求解机组组合问题的神经两阶段随机优化
链接：https://arxiv.org/abs/2507.09503

【10】On Information Geometry and Iterative Optimization in Model Compression: Operator Factorization
标题：模型压缩中的信息几何和迭代优化：运算符因式分解
链接：https://arxiv.org/abs/2507.09428

【11】Scaling Laws for Optimal Data Mixtures
标题：最佳数据混合的缩放定律
链接：https://arxiv.org/abs/2507.09404

【12】Revisiting Convergence: Shuffling Complexity Beyond Lipschitz Smoothness
标题：重温融合：超越利普希茨光滑的复杂性
链接：https://arxiv.org/abs/2507.08913

【13】Counterfactual optimization for fault prevention in complex wind energy systems
标题：复杂风能系统故障预防的反事实优化
链接：https://arxiv.org/abs/2507.08849

【14】Principled Foundations for Preference Optimization
标题：偏好优化的原则基础
链接：https://arxiv.org/abs/2507.07855

【15】Information Must Flow: Recursive Bootstrapping for Information Bottleneck in Optimal Transport
标题：信息必须流动：最佳运输中信息瓶颈的渐进引导
链接：https://arxiv.org/abs/2507.10443

【16】Aligning Generative Speech Enhancement with Human Preferences via Direct Preference Optimization
标题：通过直接偏好优化将生成语音增强与人类偏好保持一致
链接：https://arxiv.org/abs/2507.09929

【17】Optimal High-probability Convergence of Nonlinear SGD under Heavy-tailed Noise via Symmetrization
标题：通过对称化实现重尾噪音下非线性BCD的最优高概率收敛
链接：https://arxiv.org/abs/2507.09093

【18】A Method for Learning to Solve Parametric Bilevel Optimization with Coupling Constraints
标题：具有耦合约束的参数二层优化学习方法
链接：https://arxiv.org/abs/2507.09050

【19】Stochastic Approximation with Block Coordinate Optimal Stepsizes
标题：具有块坐标最优步距的随机逼近
链接：https://arxiv.org/abs/2507.08963

预测|估计(16篇)

【1】SentiDrop: A Multi Modal Machine Learning model for Predicting Dropout in Distance Learning
标题：SentiDrop：用于预测远程学习辍学的多模式机器学习模型
链接：https://arxiv.org/abs/2507.10421

【2】Spatial Lifting for Dense Prediction
标题：密集预测的空间提升
链接：https://arxiv.org/abs/2507.10222

【3】NeuTSFlow: Modeling Continuous Functions Behind Time Series Forecasting
标题：NeuTSFlow：建模时间序列预测背后的连续函数
链接：https://arxiv.org/abs/2507.09888

【4】Frequency-aware Surrogate Modeling With SMT Kernels For Advanced Data Forecasting
标题：使用贴片核心进行频率感知代理建模，用于高级数据预测
链接：https://arxiv.org/abs/2507.09694

【5】Symptom-Driven Personalized Proton Pump Inhibitors Therapy Using Bayesian Neural Networks and Model Predictive Control
标题：使用Bayesian神经网络和模型预测控制的症状驱动个性化质子泵抑制剂治疗
链接：https://arxiv.org/abs/2507.09685

【6】Conformal Prediction for Privacy-Preserving Machine Learning
标题：隐私保护机器学习的保形预测
链接：https://arxiv.org/abs/2507.09678

【7】Lightweight Deep Learning-Based Channel Estimation for RIS-Aided Extremely Large-Scale MIMO Systems on Resource-Limited Edge Devices
标题：资源有限边缘设备上的RIS辅助超大规模CDMA系统的轻量级基于深度学习的信道估计
链接：https://arxiv.org/abs/2507.09627

【8】Fourier Basis Mapping: A Time-Frequency Learning Framework for Time Series Forecasting
标题：傅里叶基映射：时间序列预测的时频学习框架
链接：https://arxiv.org/abs/2507.09445

【9】XiChen: An observation-scalable fully AI-driven global weather forecasting system with 4D variational knowledge
标题：XiChen：一个可观测扩展的完全人工智能驱动的全球天气预报系统，具有4D变分知识
链接：https://arxiv.org/abs/2507.09202

【10】Queue up for takeoff: a transferable deep learning framework for flight delay prediction
标题：排队起飞：用于航班延误预测的可转移深度学习框架
链接：https://arxiv.org/abs/2507.09084

【11】ToxBench: A Binding Affinity Prediction Benchmark with AB-FEP-Calculated Labels for Human Estrogen Receptor Alpha
标题：ToxBench：使用AB-BEP计算的人类雌激素受体Alpha标记物的结合亲和力预测基准
链接：https://arxiv.org/abs/2507.08966

【12】e-Profits: A Business-Aligned Evaluation Metric for Profit-Sensitive Customer Churn Prediction
标题：e-Profits：一种面向业务的利润敏感型客户流失预测评估指标
链接：https://arxiv.org/abs/2507.08860

【13】Foundation models for time series forecasting: Application in conformal prediction
标题：时间序列预测的基础模型：在保形预测中的应用
链接：https://arxiv.org/abs/2507.08858

【14】A Hybrid Machine Learning Framework for Optimizing Crop Selection via Agronomic and Economic Forecasting
标题：通过农艺和经济预测优化作物选择的混合机器学习框架
链接：https://arxiv.org/abs/2507.08832

【15】Machine-Precision Prediction of Low-Dimensional Chaotic Systems
标题：低维混乱系统的机器精度预测
链接：https://arxiv.org/abs/2507.09652

【16】Physics-informed machine learning: A mathematical framework with applications to time series forecasting
标题：基于物理的机器学习：应用于时间序列预测的数学框架
链接：https://arxiv.org/abs/2507.08906

其他神经网络|深度学习|模型|建模(45篇)

【1】Disentangling Neural Disjunctive Normal Form Models
标题：解开神经析取范式模型
链接：https://arxiv.org/abs/2507.10546

【2】Split Happens: Combating Advanced Threats with Split Learning and Function Secret Sharing
标题：拆分发生：通过拆分学习和函数秘密共享对抗高级威胁
链接：https://arxiv.org/abs/2507.10494

【3】Overcoming catastrophic forgetting in neural networks
标题：克服神经网络中的灾难性遗忘
链接：https://arxiv.org/abs/2507.10485

【4】Multiple Choice Learning of Low Rank Adapters for Language Modeling
标题：语言建模低级别适配器的多项选择学习
链接：https://arxiv.org/abs/2507.10419

【5】Stochastic Operator Network: A Stochastic Maximum Principle Based Approach to Operator Learning
标题：随机运营商网络：基于随机最大原则的运营商学习方法
链接：https://arxiv.org/abs/2507.10401

【6】Anticipating the Selectivity of Cyclization Reaction Pathways with Neural Network Potentials
标题：利用神经网络潜力预测环化反应途径的选择性
链接：https://arxiv.org/abs/2507.10400

【7】Test-Time Canonicalization by Foundation Models for Robust Perception
标题：通过稳健感知的基础模型实现测试时规范化
链接：https://arxiv.org/abs/2507.10375

【8】Enhanced DeepONet for 1-D consolidation operator learning: an architectural investigation
标题：用于1-D合并操作员学习的增强DeepONet：架构调查
链接：https://arxiv.org/abs/2507.10368

【9】Parallel Sampling of Diffusion Models on $SO(3)$
标题：$SO（3）$上扩散模型的并行抽样
链接：https://arxiv.org/abs/2507.10347

【10】Some Super-approximation Rates of ReLU Neural Networks for Korobov Functions
标题：Korobov函数的ReLU神经网络的某些超逼近率
链接：https://arxiv.org/abs/2507.10345

【11】Deep Recurrence for Dynamical Segmentation Models
标题：动态分割模型的深度回归
链接：https://arxiv.org/abs/2507.10143

【12】(Almost) Free Modality Stitching of Foundation Models
标题：基础模型的（几乎）自由模态拼接
链接：https://arxiv.org/abs/2507.10015

【13】Effects of structural properties of neural networks on machine learning performance
标题：神经网络结构特性对机器学习性能的影响
链接：https://arxiv.org/abs/2507.10005

【14】Rethinking Inductive Bias in Geographically Neural Network Weighted Regression
标题：重新思考地理神经网络加权回归中的归纳偏差
链接：https://arxiv.org/abs/2507.09958

【15】Iceberg: Enhancing HLS Modeling with Synthetic Data
标题：Iceberg：利用合成数据增强HLS建模
链接：https://arxiv.org/abs/2507.09948

【16】Algorithm Development in Neural Networks: Insights from the Streaming Parity Task
标题：神经网络中的算法开发：流媒体对等任务的见解
链接：https://arxiv.org/abs/2507.09897

【17】AdaBrain-Bench: Benchmarking Brain Foundation Models for Brain-Computer Interface Applications
标题：AdaBrain-Bench：脑机接口应用的脑基础模型基准
链接：https://arxiv.org/abs/2507.09882

【18】Task Priors: Enhancing Model Evaluation by Considering the Entire Space of Downstream Tasks
标题：任务优先级：通过考虑下游任务的整个空间来增强模型评估
链接：https://arxiv.org/abs/2507.09871

【19】Compressed Computation: Dense Circuits in a Toy Model of the Universal-AND Problem
标题：压缩计算：普适与问题玩具模型中的密集电路
链接：https://arxiv.org/abs/2507.09816

【20】Physics-informed neural networks for high-dimensional solutions and snaking bifurcations in nonlinear lattices
标题：基于物理知识的神经网络用于非线性网格中的多维解和蛇形分叉
链接：https://arxiv.org/abs/2507.09782

【21】Networked Information Aggregation via Machine Learning
标题：通过机器学习聚合网络信息
链接：https://arxiv.org/abs/2507.09683

【22】Toward Developing Machine-Learning-Aided Tools for the Thermomechanical Monitoring of Nuclear Reactor Components
标题：开发用于核反应堆部件热机械监测的机器学习辅助工具
链接：https://arxiv.org/abs/2507.09443

【23】Dynamic Sparse Causal-Attention Temporal Networks for Interpretable Causality Discovery in Multivariate Time Series
标题：用于多元时间序列中可解释因果关系发现的动态稀疏因果关系-注意时间网络
链接：https://arxiv.org/abs/2507.09439

【24】A Random Matrix Theory Perspective on the Learning Dynamics of Multi-head Latent Attention
标题：随机矩阵理论视角下多头隐性注意力的学习动力学
链接：https://arxiv.org/abs/2507.09394

【25】Unified Linear Parametric Map Modeling and Perception-aware Trajectory Planning for Mobile Robotics
标题：移动机器人的统一线性参数地图建模和感知轨迹规划
链接：https://arxiv.org/abs/2507.09340

【26】Optimizing Basis Function Selection in Constructive Wavelet Neural Networks and Its Applications
标题：构造性子波神经网络中基函数的优化选择及其应用
链接 ：https://arxiv.org/abs/2507.09213

【27】Capturing Unseen Spatial Extremes Through Knowledge-Informed Generative Modeling
标题：通过知识知情生成建模捕捉不可见的空间极端
链接：https://arxiv.org/abs/2507.09211

【28】Mind the Gap: Preserving and Compensating for the Modality Gap in CLIP-Based Continual Learning
标题：注意差距：保留和补偿基于CLIP的持续学习中的模式差距
链接：https://arxiv.org/abs/2507.09118

【29】Imitation Learning in Continuous Action Spaces: Mitigating Compounding Error without Interaction
标题：连续动作空间中的模仿学习：在没有交互的情况下减轻复合误差
链接：https://arxiv.org/abs/2507.09061

【30】Can Contrastive Learning Improve Class-Imbalanced Diffusion Model?
标题：对比学习能否改善班级不平衡扩散模型？
链接：https://arxiv.org/abs/2507.09052

【31】Confounder-Free Continual Learning via Recursive Feature Normalization
标题：通过回归特征规范化实现无混淆连续学习
链接：https://arxiv.org/abs/2507.09031

【32】Model Parallelism With Subnetwork Data Parallelism
标题：利用子网络数据并行主义模型并行主义
链接：https://arxiv.org/abs/2507.09029

【33】Enhancing RLHF with Human Gaze Modeling
标题：通过人体凝视建模增强RL HF
链接：https://arxiv.org/abs/2507.09016

【34】Exploiting Leaderboards for Large-Scale Distribution of Malicious Models
标题：利用排行榜进行恶意模型的大规模分发
链接：https://arxiv.org/abs/2507.08983

【35】Simulating Three-dimensional Turbulence with Physics-informed Neural Networks
标题：利用物理知识的神经网络模拟三维湍流
链接：https://arxiv.org/abs/2507.08972

【36】Beyond Scores: Proximal Diffusion Models
标题：超越分数：近端扩散模型
链接：https://arxiv.org/abs/2507.08956

【37】The Engineer's Dilemma: A Review of Establishing a Legal Framework for Integrating Machine Learning in Construction by Navigating Precedents and Industry Expectations
标题：工程师的困境：回顾通过借鉴先例和行业预期建立将机器学习集成到建筑中的法律框架
链接：https://arxiv.org/abs/2507.08908

【38】Physical Informed Neural Networks for modeling ocean pollutant
标题：用于海洋污染物建模的物理信息神经网络
链接：https://arxiv.org/abs/2507.08834

【39】View Invariant Learning for Vision-Language Navigation in Continuous Environments
标题：查看连续环境中视觉语言导航的不变学习
链接：https://arxiv.org/abs/2507.08831

【40】Dynamical stability for dense patterns in discrete attractor neural networks
标题：离散吸引子神经网络密集模式的动态稳定性
链接：https://arxiv.org/abs/2507.10383

【41】MF-GLaM: A multifidelity stochastic emulator using generalized lambda models
标题：MF-GLaM：使用广义Lambda模型的多保真随机模拟器
链接：https://arxiv.org/abs/2507.10303

【42】Sequence-Model-Guided Measurement Selection for Quantum State Learning
标题：量子状态学习的序列模型引导测量选择
链接：https://arxiv.org/abs/2507.09891

【43】Physics-Based Machine Learning Closures and Wall Models for Hypersonic Transition-Continuum Boundary Layer Predictions
标题：基于物理的机器学习闭合和壁模型用于高超音速过渡-连续边界层预测
链接：https://arxiv.org/abs/2507.08986

【44】The Bayesian Approach to Continual Learning: An Overview
标题：持续学习的Bayesian方法：概述
链接：https://arxiv.org/abs/2507.08922

【45】DiffNMR: Diffusion Models for Nuclear Magnetic Resonance Spectra Elucidation
标题：迪夫核磁共振光谱解析的扩散模型
链接：https://arxiv.org/abs/2507.08854

其他(44篇)

【1】Quantize-then-Rectify: Efficient VQ-VAE Training
标题：量化然后纠正：高效的VQ-VAE训练
链接：https://arxiv.org/abs/2507.10547

【2】National level satellite-based crop field inventories in smallholder landscapes
标题：国家一级小农景观卫星农田清查
链接：https://arxiv.org/abs/2507.10499

【3】The Target Polish: A New Approach to Outlier-Resistant Non-Negative Matrix and Tensor Factorization
标题：目标波兰语：抗异常值非负矩阵和张量分解的新方法
链接：https://arxiv.org/abs/2507.10484

【4】Logic layer Prompt Control Injection (LPCI): A Novel Security Vulnerability Class in Agentic Systems
标题：逻辑层提示控制注入（LCI）：统计系统中的一种新型安全漏洞类
链接：https://arxiv.org/abs/2507.10457

【5】FinTeam: A Multi-Agent Collaborative Intelligence System for Comprehensive Financial Scenarios
标题：FinTeam：针对综合金融场景的多智能体协作智能系统
链接：https://arxiv.org/abs/2507.10448

【6】Energy Efficiency in AI for 5G and Beyond: A DeepRx Case Study
标题：5G及其他领域人工智能的能源效率：DeepRx案例研究
链接：https://arxiv.org/abs/2507.10409

【7】On the Efficiency of Training Robust Decision Trees
标题：鲁棒决策树训练效率的研究
链接：https://arxiv.org/abs/2507.10048

【8】Compliance Minimization via Physics-Informed Gaussian Processes
标题：通过基于物理信息的高斯过程最小化合规性
链接：https://arxiv.org/abs/2507.09968

【9】MixLoRA-DSI: Dynamically Expandable Mixture-of-LoRA Experts for Rehearsal-Free Generative Retrieval over Dynamic Corpora
标题：MixLoRA-DSI：可动态扩展的Mixture-of-LoRA专家，用于动态库上的免排练生成式检索
链接：https://arxiv.org/abs/2507.09924

【10】Function Induction and Task Generalization: An Interpretability Study with Off-by-One Addition
标题：功能归纳和任务概括：带有差一添加的可解释性研究
链接：https://arxiv.org/abs/2507.09875

【11】A Pre-training Framework for Relational Data with Information-theoretic Principles
标题：具有信息论原理的关系数据预训练框架
链接：https://arxiv.org/abs/2507.09837

【12】Generative Cognitive Diagnosis
标题：生成性认知诊断
链接：https://arxiv.org/abs/2507.09831

【13】A Scalable and Efficient Signal Integration System for Job Matching
标题：用于职位匹配的可扩展且高效的信号集成系统
链接：https://arxiv.org/abs/2507.09797

【14】Leveraging Distribution Matching to Make Approximate Machine Unlearning Faster
标题：利用分布匹配使机器更快地去学习
链接：https://arxiv.org/abs/2507.09786

【15】Knowing When to Quit: Probabilistic Early Exits for Speech Separation
标题：知道何时退出：可能因言语分离而提前退出
链接：https://arxiv.org/abs/2507.09768

【16】MB-RIRs: a Synthetic Room Impulse Response Dataset with Frequency-Dependent Absorption Coefficients
标题：MB-RIR：具有频率相关吸收系数的合成房间脉冲响应数据集
链接：https://arxiv.org/abs/2507.09750

【17】Universal Physics Simulation: A Foundational Diffusion Approach
标题：宇宙物理模拟：一种基础扩散方法
链接：https://arxiv.org/abs/2507.09733

【18】Continental scale habitat modelling with artificial intelligence and multimodal earth observation
标题：利用人工智能和多模式地球观测进行大陆规模栖息地建模
链接：https://arxiv.org/abs/2507.09732

【19】Phase transition of the Sinkhorn-Knopp algorithm
标题 ：Sinkhorn-Knopp算法的相变
链接：https://arxiv.org/abs/2507.09711

【20】EPT-2 Technical Report
标题：EPT-2技术报告
链接：https://arxiv.org/abs/2507.09703

【21】Cultivating Pluralism In Algorithmic Monoculture: The Community Alignment Dataset
标题：在单一文化中培养多元化：社区对齐数据集
链接：https://arxiv.org/abs/2507.09650

【22】Disentanglement and Assessment of Shortcuts in Ophthalmological Retinal Imaging Exams
标题：眼科视网膜成像检查中的解开和捷径评估
链接：https://arxiv.org/abs/2507.09640

【23】Incentive-Aware Dynamic Resource Allocation under Long-Term Cost Constraints
标题：长期成本约束下激励意识的动态资源配置
链接：https://arxiv.org/abs/2507.09473

【24】GreenCrossingAI: A Camera Trap/Computer Vision Pipeline for Environmental Science Research Groups
标题：GreenCrossingAI：环境科学研究小组的摄像机陷阱/计算机视觉管道
链接：https://arxiv.org/abs/2507.09410

【25】Supercharging Floorplan Localization with Semantic Rays
标题：基于语义射线的平面图定位
链接：https://arxiv.org/abs/2507.09291

【26】ClaritySpeech: Dementia Obfuscation in Speech
标题：ClaritySpeech：言语中的痴呆症混淆
链接：https://arxiv.org/abs/2507.09282

【27】Warm Starts Accelerate Generative Modelling
标题：温暖的开始加速生成建模
链接：https://arxiv.org/abs/2507.09212

【28】A Study of Value-Aware Eigenoptions
标题：价值感知特征期权研究
链接：https://arxiv.org/abs/2507.09127

【29】Continuous-Time Signal Decomposition: An Implicit Neural Generalization of PCA and ICA
标题：连续时间信号分解：PCA和ICA的隐式神经推广
链接：https://arxiv.org/abs/2507.09091

【30】SetupBench: Assessing Software Engineering Agents' Ability to Bootstrap Development Environments
标题：SetupBench：评估软件工程代理引导开发环境的能力
链接：https://arxiv.org/abs/2507.09063

【31】Last Layer Hamiltonian Monte Carlo
标题：最后一层汉密尔顿蒙特卡洛
链接：https://arxiv.org/abs/2507.08905

【32】GUIDE: Towards Scalable Advising for Research Ideas
标题：指南：迈向可扩展的研究想法咨询
链接：https://arxiv.org/abs/2507.08870

【33】Underrepresentation, Label Bias, and Proxies: Towards Data Bias Profiles for the EU AI Act and Beyond
标题：代表性不足、标签偏见和代理：欧盟人工智能法案及其他法案的数据偏见概况
链接：https://arxiv.org/abs/2507.08866

【34】On the under-reaching phenomenon in message-passing neural PDE solvers: revisiting the CFL condition
标题：关于消息传递神经DTE解算器中的不足现象：重温IPL条件
链接：https://arxiv.org/abs/2507.08861

【35】Can We Predict Your Next Move Without Breaking Your Privacy?
标题：我们可以在不侵犯您隐私的情况下预测您的下一步行动吗？
链接：https://arxiv.org/abs/2507.08843

【36】LoRA Is Slower Than You Think
标题：LoRA比你想象的要慢
链接：https://arxiv.org/abs/2507.08833

【37】An Algorithm for Identifying Interpretable Subgroups With Elevated Treatment Effects
标题：识别具有更高治疗效果的可解释亚组的算法
链接：https://arxiv.org/abs/2507.09494