cs.LG(2025-10-01)

📊 共 9 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) (5 🔗2) 支柱二:RL算法与架构 (RL & Architecture) (2) 支柱八:物理动画 (Physics-based Animation) (1) 支柱一:机器人控制 (Robot Control) (1)

🔬 支柱九:具身大模型 (Embodied Foundation Models) (5 篇)

#题目一句话要点标签🔗
1 Combining Large Language Models and Gradient-Free Optimization for Automatic Control Policy Synthesis 结合大语言模型与无梯度优化,实现自动控制策略生成 large language model
2 Automated Structured Radiology Report Generation with Rich Clinical Context 提出C-SRRG,利用丰富临床上下文自动生成结构化放射报告,提升报告质量。 large language model multimodal
3 Learning a Zeroth-Order Optimizer for Fine-Tuning LLMs 提出ZO Fine-tuner,一种学习型零阶优化器,用于高效微调大语言模型。 large language model foundation model
4 AbsTopK: Rethinking Sparse Autoencoders For Bidirectional Features 提出AbsTopK以解决稀疏自编码器的双向特征表示问题 large language model
5 Microsaccade-Inspired Probing: Positional Encoding Perturbations Reveal LLM Misbehaviours 微眼跳启发式探测:位置编码扰动揭示大语言模型的不良行为 large language model

🔬 支柱二:RL算法与架构 (RL & Architecture) (2 篇)

#题目一句话要点标签🔗
6 Plug-and-Play Prompt Refinement via Latent Feedback for Diffusion Model Alignment 提出PromptLoop:一种基于隐空间反馈的即插即用提示优化扩散模型对齐框架 reinforcement learning large language model multimodal
7 Can Mamba Learn In Context with Outliers? A Theoretical Generalization Analysis 首次理论分析Mamba模型ICL泛化能力,解决含离群点的二元分类问题 Mamba linear attention

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

#题目一句话要点标签🔗
8 Robust Spatiotemporally Contiguous Anomaly Detection Using Tensor Decomposition 提出基于张量分解的鲁棒时空连续异常检测方法,适用于视频监控等领域。 spatiotemporal

🔬 支柱一:机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗
9 End-to-End Training of High-Dimensional Optimal Control with Implicit Hamiltonians via Jacobian-Free Backpropagation 提出基于隐式哈密顿量的端到端高维最优控制方法,通过无雅可比反向传播实现高效训练。 trajectory optimization

⬅️ 返回 cs.LG 首页 · 🏠 返回主页