cs.LG（2024-06-19）

📊 共 8 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

#	题目	一句话要点	标签	🔗
1	Game of LLMs: Discovering Structural Constructs in Activities using Large Language Models	利用大语言模型发现活动中的结构单元，提升智能家居场景下活动识别性能	large language model
2	PRESTO: Progressive Pretraining Enhances Synthetic Chemistry Outcomes	PRESTO：渐进式预训练提升合成化学任务性能	large language model multimodal	✅
3	SDQ: Sparse Decomposed Quantization for LLM Inference	提出SDQ：一种稀疏分解量化方法，用于加速LLM推理并降低内存占用。	large language model
4	Prose-to-P4: Leveraging High Level Languages	利用大型语言模型，实现自然语言到P4数据平面代码的自动生成。	large language model
5	BoA: Attention-aware Post-training Quantization without Backpropagation	提出BoA：一种无需反向传播的注意力感知后训练量化方法	large language model	✅

#	题目	一句话要点	标签	🔗
6	Can Low-Rank Knowledge Distillation in LLMs be Useful for Microelectronic Reasoning?	提出LoRA-KD方案，探索LLM在微电子推理中的应用潜力	distillation large language model
7	Global Human-guided Counterfactual Explanations for Molecular Properties via Reinforcement Learning	提出RLHEX模型，通过强化学习生成符合人类认知的分子性质全局反事实解释。	reinforcement learning PPO	✅
8	Infinite-Horizon Reinforcement Learning with Multinomial Logistic Function Approximation	提出基于多项逻辑函数逼近的无限期强化学习算法	reinforcement learning