cs.LG（2026-04-23）

📊 共 20 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (11 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗1) 支柱一：机器人控制 (Robot Control) (1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (11 篇)

#题目一句话要点标签🔗⭐
1 ARFBench: Benchmarking Time Series Question Answering Ability for Software Incident Response ARFBench：软件事件响应时序问答能力评测基准 foundation model multimodal ✅
2 Supervised Learning Has a Necessary Geometric Blind Spot: Theory, Consequences, and Minimal Repair 监督学习存在几何盲点：提出TDI指标并用PMH方法修复 foundation model
3 Focus Session: Hardware and Software Techniques for Accelerating Multimodal Foundation Models 提出软硬件协同优化方法，加速多模态基础模型在医疗和代码生成任务中的应用。 foundation model multimodal
4 Mochi: Aligning Pre-training and Inference for Efficient Graph Foundation Models via Meta-Learning Mochi：通过元学习对齐预训练与推理，实现高效图基础模型 foundation model
5 Toward Efficient Membership Inference Attacks against Federated Large Language Models: A Projection Residual Approach 提出ProjRes以解决联邦大语言模型的成员推断攻击问题 large language model
6 Tempered Sequential Monte Carlo for Trajectory and Policy Optimization with Differentiable Dynamics 提出基于退火序贯蒙特卡洛的轨迹与策略优化方法，适用于可微分动力学系统。 multimodal
7 PrivUn: Unveiling Latent Ripple Effects and Shallow Forgetting in Privacy Unlearning PrivUn：揭示隐私卸载中的潜在涟漪效应和浅层遗忘问题 large language model
8 Audio Video Verbal Analysis (AVVA) for Capturing Classroom Dialogues 提出AVVA框架，用于捕捉和分析课堂对话中的多模态交互信息。 multimodal
9 Low-Rank Adaptation Redux for Large Models 综述LoRA：以信号处理视角解析大模型参数高效微调方法 foundation model
10 A-THENA: Early Intrusion Detection for IoT with Time-Aware Hybrid Encoding and Network-Specific Augmentation A-THENA：基于时间感知混合编码和网络特定增强的物联网早期入侵检测系统 TAMP
11 Decoupled Travel Planning with Behavior Forest 提出基于行为森林的解耦旅行规划方法，提升LLM在复杂约束下的规划能力。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (7 篇)

#题目一句话要点标签🔗⭐
12 Learning Dynamic Representations and Policies from Multimodal Clinical Time-Series with Informative Missingness 提出利用信息缺失性的多模态临床时间序列动态表征学习框架，提升治疗策略学习和预后预测。 policy learning representation learning multimodal
13 Understanding and Mitigating Spurious Signal Amplification in Test-Time Reinforcement Learning for Math Reasoning 提出DDRL框架，解决测试时强化学习中数学推理的伪标签噪声放大问题 reinforcement learning large language model ✅
14 Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding 提出基于强化学习的Propose-then-Critic协同进化框架，用于提升GUI界面元素定位精度。 reinforcement learning
15 Insect-inspired modular architectures as inductive biases for reinforcement learning 提出昆虫启发式模块化强化学习架构，解决复杂导航任务中动态行为竞争问题。 reinforcement learning PPO
16 Null-Space Flow Matching for MIMO Channel Estimation in Latency-Constrained Systems 提出空域流匹配方法，解决低延迟约束下的MIMO信道估计问题 flow matching
17 Dynamical Priors as a Training Objective in Reinforcement Learning DP-RL：通过动态先验作为强化学习的训练目标，提升决策时序一致性 reinforcement learning
18 CAP: Controllable Alignment Prompting for Unlearning in LLMs 提出CAP框架，通过可控对齐提示实现LLM的知识遗忘。 reinforcement learning large language model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
19 PermaFrost-Attack: Stealth Pretraining Seeding(SPS) for planting Logic Landmines During LLM Training 提出PermaFrost-Attack，通过隐蔽预训练注入逻辑炸弹攻击大语言模型 manipulation large language model foundation model

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

#题目一句话要点标签🔗⭐
20 A Scale-Adaptive Framework for Joint Spatiotemporal Super-Resolution with Diffusion Models 提出基于扩散模型的尺度自适应时空超分辨率框架，解决气候应用中多尺度问题。 spatiotemporal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页