cs.LG（2025-07-30）

📊 共 24 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (13) 支柱二：RL算法与架构 (RL & Architecture) (10) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (13 篇)

#题目一句话要点标签🔗⭐
1 Spec-VLA: Speculative Decoding for Vision-Language-Action Models with Relaxed Acceptance Spec-VLA：通过放宽接受条件加速视觉-语言-动作模型的推测解码 vision-language-action VLA large language model
2 Doctor Sun: A Bilingual Multimodal Large Language Model for Biomedical AI Doctor Sun：一种用于生物医学AI的双语多模态大型语言模型 large language model multimodal
3 A Foundation Model for Material Fracture Prediction 提出基于Transformer的材料断裂预测基础模型，提升泛化性和效率。 large language model foundation model multimodal
4 Investigating the Invertibility of Multimodal Latent Spaces: Limitations of Optimization-Based Methods 研究多模态隐空间的可逆性：优化方法的局限性分析 multimodal
5 Quantifying surprise in clinical care: Detecting highly informative events in electronic health records with foundation models 利用电子病历中的Foundation Model量化临床诊疗中的“意外”事件，从而检测高信息量事件。 foundation model
6 H2Tune: Federated Foundation Model Fine-Tuning with Hybrid Heterogeneity H2Tune：针对模型架构和任务双重异构的联邦基础模型微调框架 foundation model
7 Hybrid Hypergraph Networks for Multimodal Sequence Data Classification 提出混合超图网络HHN，用于建模多模态时序数据分类，提升长程依赖和跨模态交互。 multimodal
8 Multimodal Late Fusion Model for Problem-Solving Strategy Classification in a Machine Learning Game 提出多模态晚期融合模型，用于机器学习游戏中问题解决策略分类 multimodal
9 On the Sustainability of AI Inferences in the Edge 边缘AI推理可持续性研究：针对不同边缘设备和模型的性能与能耗权衡分析 large language model
10 KLLM: Fast LLM Inference with K-Means Quantization KLLM：基于K-Means量化的快速LLM推理加速器 large language model
11 Stop Evaluating AI with Human Tests, Develop Principled, AI-specific Tests instead 呼吁停止使用人类测试评估AI，转而开发AI专属的、基于原则的测试方法 large language model
12 Agentic Privacy-Preserving Machine Learning 提出Agentic-PPML框架，提升隐私保护大语言模型推理的实用性 large language model
13 Breaking Obfuscation: Cluster-Aware Graph with LLM-Aided Recovery for Malicious JavaScript Detection 提出DeCoda框架，结合LLM去混淆和聚类感知图学习，提升恶意JavaScript代码检测效果。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (10 篇)

#题目一句话要点标签🔗⭐
14 Deep Reinforcement Learning in Factor Investment 提出CAFPO，利用深度强化学习解决低频因子投资组合构建问题 reinforcement learning deep reinforcement learning DRL
15 Privileged Contrastive Pretraining for Multimodal Affect Modelling 提出Privileged Contrastive Pretraining框架，提升多模态情感模型在真实环境下的泛化能力。 contrastive learning privileged information multimodal
16 Planning for Cooler Cities: A Multimodal AI Framework for Predicting and Mitigating Urban Heat Stress through Urban Landscape Transformation 提出GSM-UTCI多模态AI框架，预测并缓解城市热应力，助力城市景观改造规划。 MAE multimodal
17 Efficient Differentially Private Fine-Tuning of LLMs via Reinforcement Learning 提出RLDP框架以优化大语言模型的差分隐私微调 reinforcement learning deep reinforcement learning SAC
18 Uni-Mol3: A Multi-Molecular Foundation Model for Advancing Organic Reaction Modeling Uni-Mol3：用于推进有机反应建模的多分子基础模型 representation learning foundation model
19 CS-SHRED: Enhancing SHRED for Robust Recovery of Spatiotemporal Dynamics 提出CS-SHRED以解决稀疏数据下时空动态重建问题 MAE sparse sensors spatiotemporal
20 G-Core: A Simple, Scalable and Balanced RLHF Trainer G-Core：一种简单、可扩展且均衡的RLHF训练框架，适用于大规模用户场景。 reinforcement learning RLHF large language model
21 A Bit of Freedom Goes a Long Way: Classical and Quantum Algorithms for Reinforcement Learning under a Generative Model 提出混合探索-生成式强化学习算法，量子算法在有限步MDP中突破经典regret界限。 reinforcement learning
22 RLVMR: Reinforcement Learning with Verifiable Meta-Reasoning Rewards for Robust Long-Horizon Agents RLVMR：基于可验证元推理奖励的强化学习，提升长时程Agent的鲁棒性 reinforcement learning
23 Resource-Efficient Automatic Software Vulnerability Assessment via Knowledge Distillation and Particle Swarm Optimization 提出基于知识蒸馏和粒子群优化的资源高效型软件漏洞自动评估框架 distillation

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
24 AI paradigm for solving differential equations: first-principles data generation and scale-dilation operator AI solver 提出基于第一性原理数据生成和尺度扩张算子的AI求解器，解决微分方程求解问题。 spatiotemporal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页