cs.LG（2025-12-05）

📊 共 23 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (14) 支柱二：RL算法与架构 (RL & Architecture) (6) 支柱一：机器人控制 (Robot Control) (1) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (14 篇)

#题目一句话要点标签🔗⭐
1 TS-HINT: Enhancing Semiconductor Time Series Regression Using Attention Hints From Large Language Model Reasoning TS-HINT：利用大语言模型推理提示增强半导体时序回归 large language model foundation model chain-of-thought
2 Taxonomy-Adaptive Moderation Model with Robust Guardrails for Large Language Models 提出Roblox Guard 1.0，增强LLM系统输入输出安全性的分类自适应审核模型 large language model chain-of-thought
3 Scaling and Transferability of Annealing Strategies in Large Language Model Training 提出一种可迁移的学习率退火策略优化框架，提升大语言模型训练效率。 large language model
4 Poodle: Seamlessly Scaling Down Large Language Models with Just-in-Time Model Replacement Poodle：即时模型替换，无缝缩减大语言模型规模 large language model
5 The Forgotten Shield: Safety Grafting in Parameter-Space for Medical MLLMs 提出参数空间安全嫁接方法，提升医学多模态大语言模型的安全性。 large language model multimodal
6 Physics-Informed Neural Koopman Machine for Interpretable Longitudinal Personalized Alzheimer's Disease Forecasting 提出神经Koopman机(NKM)，用于可解释的阿尔茨海默病纵向个性化预测。 multimodal
7 MaxShapley: Towards Incentive-compatible Generative Search with Fair Context Attribution 提出MaxShapley算法，用于检索增强生成搜索中的激励兼容和公平内容归因。 large language model
8 Impugan: Learning Conditional Generative Models for Robust Data Imputation Impugan：一种用于鲁棒数据插补的条件生成对抗网络模型 multimodal
9 KQ-SVD: Compressing the KV Cache with Provable Guarantees on Attention Fidelity KQ-SVD：通过优化Attention矩阵低秩分解压缩KV缓存，提升LLM推理效率 large language model
10 Mitigating Catastrophic Forgetting in Mathematical Reasoning Finetuning through Mixed Training 提出混合训练策略，缓解数学推理微调中的灾难性遗忘问题 large language model
11 Bootstrapping Fuzzers for Compilers of Low-Resource Language Dialects Using Language Models Germinator利用语言模型为低资源语言编译器自动生成fuzzer，提升测试效率。 large language model
12 Feasibility of AI-Assisted Programming for End-User Development 探索AI辅助编程在终端用户开发中的可行性，以替代或补充低代码/无代码平台。 large language model
13 RevoNAD: Reflective Evolutionary Exploration for Neural Architecture Design RevoNAD：一种反射式进化探索的神经架构设计方法，提升架构搜索的可靠性和部署性。 large language model
14 When Forgetting Builds Reliability: LLM Unlearning for Reliable Hardware Code Generation 提出面向硬件代码生成的LLM遗忘框架，提升代码可靠性 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
15 Entropy Ratio Clipping as a Soft Global Constraint for Stable Reinforcement Learning 提出熵率裁剪(ERC)机制，稳定强化学习训练，提升大语言模型后训练效果。 reinforcement learning PPO large language model
16 Quantifying Memory Use in Reinforcement Learning with Temporal Range 提出Temporal Range指标，量化强化学习策略对历史观测的记忆依赖。 reinforcement learning SSM
17 Average-reward reinforcement learning in semi-Markov decision processes via relative value iteration 提出基于相对值迭代的平均奖励SMDP强化学习算法，并证明其收敛性。 reinforcement learning
18 JaxWildfire: A GPU-Accelerated Wildfire Simulator for Reinforcement Learning 提出JaxWildfire，一种GPU加速的野火模拟器，用于强化学习。 reinforcement learning
19 Utility Boundary of Dataset Distillation: Scaling and Configuration-Coverage Laws 提出配置-动态-误差分析框架，揭示数据集蒸馏的缩放律和配置覆盖律。 distillation
20 FieldSeer I: Physics-Guided World Models for Long-Horizon Electromagnetic Dynamics under Partial Observability FieldSeer I：基于物理引导的世界模型，用于部分观测下长时程电磁动力学预测 world model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
21 Hierarchical Task Offloading and Trajectory Optimization in Low-Altitude Intelligent Networks Via Auction and Diffusion-based MARL 提出基于拍卖和扩散MARL的低空智能网络分层任务卸载与轨迹优化方案 trajectory optimization reinforcement learning

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
22 Over-the-Air Semantic Alignment with Stacked Intelligent Metasurfaces 提出基于堆叠智能超表面的无线语义对齐框架，解决异构模型潜在表示失准问题。 semantic mapping semantic map

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

#题目一句话要点标签🔗⭐
23 Designing an Optimal Sensor Network via Minimizing Information Loss 提出一种基于物理模拟和贝叶斯实验设计的传感器网络优化方法，最小化信息损失。 spatiotemporal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页