cs.LG（2025-02-11）

📊 共 11 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

#	题目	一句话要点	标签	🔗
1	ADMN: A Layer-Wise Adaptive Multimodal Network for Dynamic Input Noise and Compute Resources	提出ADMN，一种层级自适应多模态网络，解决动态输入噪声和计算资源约束问题。	multimodal
2	DarwinLM: Evolutionary Structured Pruning of Large Language Models	DarwinLM：通过演化结构化剪枝压缩大型语言模型，提升推理效率。	large language model	✅
3	Automated Capability Discovery via Foundation Model Self-Exploration	提出自动化能力发现框架以评估基础模型的多样化能力	foundation model	✅
4	CIRCUIT: A Benchmark for Circuit Interpretation and Reasoning Capabilities of LLMs	提出CIRCUIT基准数据集，评估LLM在电路推理方面的能力	large language model
5	Towards Efficient Optimizer Design for LLM via Structured Fisher Approximation with a Low-Rank Extension	提出基于结构化Fisher近似和低秩扩展的高效LLM优化器设计方法	large language model
6	EvoFlow: Evolving Diverse Agentic Workflows On The Fly	EvoFlow：动态演化多样化Agent工作流，实现异构LLM协同。	large language model

#	题目	一句话要点	标签	🔗
7	DrugImproverGPT: A Large Language Model for Drug Optimization with Fine-Tuning via Structured Policy Optimization	DrugImproverGPT：基于结构化策略优化微调的大语言模型，用于药物优化	reinforcement learning large language model	✅
8	Towards Training One-Step Diffusion Models Without Distillation	提出无需教师模型监督的单步扩散模型训练方法，性能超越蒸馏方法。	distillation
9	Advancing Autonomous VLM Agents via Variational Subgoal-Conditioned Reinforcement Learning	提出VSC-RL，通过变分子目标强化学习提升自主VLM Agent在复杂任务中的效率。	reinforcement learning
10	LASP-2: Rethinking Sequence Parallelism for Linear Attention and Its Hybrid	LASP-2：重新设计线性注意力序列并行，提升超长序列训练效率	linear attention	✅
11	Model Selection for Off-policy Evaluation: New Algorithms and Experimental Protocol	针对离线策略评估的模型选择，提出新算法与实验协议。	reinforcement learning offline reinforcement learning