cs.LG（2025-02-12）

📊 共 25 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (13 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (10 🔗1) 支柱一：机器人控制 (Robot Control) (1) 支柱八：物理动画 (Physics-based Animation) (1 🔗1)

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

#题目一句话要点标签🔗⭐
1 I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models ThinkDiff：通过对齐视觉-语言模型，赋予扩散模型多模态上下文推理能力 large language model multimodal ✅
2 Continuous Cardiac Arrest Prediction in ICU using PPG Foundation Model 提出基于PPG预训练模型的FEAN网络，用于ICU中连续心搏骤停预测。 foundation model
3 Mathematical Reasoning in Large Language Models: Assessing Logical and Arithmetic Errors across Wide Numerical Ranges 提出GSM-Ranges数据集与新评估方法，评估LLM在不同数值范围下的数学推理能力 large language model
4 Quality over Quantity: Boosting Data Efficiency Through Ensembled Multimodal Data Curation EcoDatum：通过集成多模态数据清洗算子提升数据效率，解决网络爬取数据集的质量问题。 multimodal
5 LLM4GNAS: A Large Language Model Based Toolkit for Graph Neural Architecture Search LLM4GNAS：基于大语言模型的图神经网络架构搜索工具包 large language model
6 Spectral Journey: How Transformers Predict the Shortest Path Transformer学习最短路径：揭示谱分解与路径规划的关联 large language model
7 Commercial LLM Agents Are Already Vulnerable to Simple Yet Dangerous Attacks 揭示商业LLM Agent的简单而危险的攻击漏洞，无需机器学习知识 large language model
8 The Paradox of Stochasticity: Limited Creativity and Computational Decoupling in Temperature-Varied LLM Outputs of Structured Fictional Data 研究表明：温度对LLM生成结构化虚构数据的影响有限，模型架构是性能关键 large language model
9 Trustworthy GNNs with LLMs: A Systematic Review and Taxonomy 综述：利用大语言模型提升图神经网络可信度，提出系统分类法 large language model
10 Self-Evaluation for Job-Shop Scheduling 提出基于自评估的Job-Shop调度方法，超越现有技术水平 large language model
11 Democratizing AI: Open-source Scalable LLM Training on GPU-based Supercomputers AxoNN：开源可扩展LLM训练框架，实现GPU超算上的高效训练。 large language model
12 One Example Shown, Many Concepts Known! Counterexample-Driven Conceptual Reasoning in Mathematical LLMs 提出CounterMATH基准，提升数学LLM基于反例的概念推理能力 large language model
13 LowRA: Accurate and Efficient LoRA Fine-Tuning of LLMs under 2 Bits LowRA：在低于2比特下实现LLM的精确高效LoRA微调 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-Based User Scheduling for Collaborative Perception 提出基于深度强化学习的V2X用户调度算法SchedCP，用于协同感知。 reinforcement learning deep reinforcement learning
15 From Layers to States: A State Space Model Perspective to Deep Neural Network Layer Dynamics 提出S6LA模块，利用选择性状态空间模型提升深度神经网络层聚合能力，改进图像分类与检测。 SSM state space model
16 A Survey on Data-Centric AI: Tabular Learning from Reinforcement Learning and Generative AI Perspective 综述：数据中心AI视角下表格数据的强化学习与生成式学习 reinforcement learning
17 Necessary and Sufficient Oracles: Toward a Computational Taxonomy For Reinforcement Learning 针对强化学习，提出必要且充分的监督学习Oracle计算分类方法 reinforcement learning
18 Distillation Scaling Laws 提出蒸馏缩放定律，优化师生模型计算资源分配以提升学生模型性能。 distillation
19 LLM Pretraining with Continuous Concepts 提出CoCoMix框架，融合离散token预测与连续概念，提升LLM预训练效率与可解释性。 distillation large language model
20 One-Shot Federated Learning with Classifier-Free Diffusion Models 提出OSCAR，一种基于无分类器扩散模型的单次联邦学习方法，降低通信和计算开销。 distillation foundation model
21 Closer through commonality: Enhancing hypergraph contrastive learning with shared groups 提出HyFi：通过共享群组增强超图对比学习，提升节点分类性能 contrastive learning ✅
22 Enhanced Load Forecasting with GAT-LSTM: Leveraging Grid and Temporal Features 提出GAT-LSTM模型，融合电网和时序特征，提升电力负荷预测精度 MAE spatial relationship
23 Towards Principled Unsupervised Multi-Agent Reinforcement Learning 提出一种基于混合熵最大化的可扩展非监督多智能体强化学习方法 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
24 LDC-MTL: Balancing Multi-Task Learning through Scalable Loss Discrepancy Control 提出LDC-MTL以解决多任务学习中的损失不平衡问题 manipulation

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

#题目一句话要点标签🔗⭐
25 TANTE: Time-Adaptive Operator Learning via Neural Taylor Expansion 提出TANTE框架以解决时间依赖PDE的适应性问题 spatiotemporal ✅

⬅️ 返回 cs.LG 首页 · 🏠 返回主页