cs.LG（2024-02-06）

📊 共 35 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (16 🔗3) 支柱九：具身大模型 (Embodied Foundation Models) (13 🔗2) 支柱一：机器人控制 (Robot Control) (4) 支柱五：交互与反应 (Interaction & Reaction) (1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱二：RL算法与架构 (RL & Architecture) (16 篇)

#题目一句话要点标签🔗⭐
1 Entropy-regularized Diffusion Policy with Q-Ensembles for Offline Reinforcement Learning 提出熵正则化扩散策略与Q集成以提升离线强化学习性能 reinforcement learning offline RL offline reinforcement learning ✅
2 Learning to Generate Explainable Stock Predictions using Self-Reflective Large Language Models 提出SEP框架以解决股票预测解释性不足的问题 PPO large language model
3 SEABO: A Simple Search-Based Method for Offline Imitation Learning 提出SEABO以解决离线模仿学习中的奖励函数设计问题 reinforcement learning offline RL offline reinforcement learning ✅
4 The Hedgehog & the Porcupine: Expressive Linear Attentions with Softmax Mimicry 提出Hedgehog以解决线性注意力性能不足问题 linear attention large language model
5 Return-Aligned Decision Transformer 提出Return-Aligned Decision Transformer以解决目标回报对齐问题 reinforcement learning offline reinforcement learning decision transformer ✅
6 Informed Reinforcement Learning for Situation-Aware Traffic Rule Exceptions 提出知情强化学习以解决交通规则例外问题 reinforcement learning reward design
7 Can Mamba Learn How to Learn? A Comparative Study on In-Context Learning Tasks 提出MambaFormer以提升状态空间模型的上下文学习能力 Mamba SSM
8 Averaging $n$-step Returns Reduces Variance in Reinforcement Learning 提出复合回报以降低强化学习中的方差问题 reinforcement learning PPO
9 CAMBranch: Contrastive Learning with Augmented MILPs for Branching 提出CAMBranch以解决混合整数线性规划中的分支决策问题 imitation learning contrastive learning
10 Efficient Availability Attacks against Supervised and Contrastive Learning Simultaneously 提出高效可用性攻击以同时针对监督学习和对比学习 contrastive learning
11 In-context learning agents are asymmetric belief updaters 研究大型语言模型的不对称信念更新机制 reinforcement learning large language model
12 AdaFlow: Imitation Learning with Variance-Adaptive Flow-Based Policies 提出AdaFlow以解决多模态决策的高效模仿学习问题 imitation learning
13 No-Regret Reinforcement Learning in Smooth MDPs 提出ν-平滑性假设以解决连续状态空间中的无悔强化学习问题 reinforcement learning
14 Reinforcement Learning from Bagged Reward 提出基于袋装奖励的强化学习方法以解决奖励信号设计难题 reinforcement learning
15 Improved Generalization of Weight Space Networks via Augmentations 提出数据增强方法以解决深权重空间网络的过拟合问题 contrastive learning NeRF
16 MusicRL: Aligning Music Generation to Human Preferences 提出MusicRL以解决音乐生成中的人类偏好对齐问题 reinforcement learning RLHF

🔬 支柱九：具身大模型 (Embodied Foundation Models) (13 篇)

#题目一句话要点标签🔗⭐
17 RAP: Retrieval-Augmented Planning with Contextual Memory for Multimodal LLM Agents 提出检索增强规划框架以解决多模态LLM代理的决策问题 large language model multimodal
18 A quantitative analysis of knowledge-learning preferences in large language models in molecular science 提出ChEBI-20-MM基准以量化大语言模型在分子科学中的知识学习偏好 large language model
19 Discovering and Reasoning of Causality in the Hidden World with Large Language Models 提出COAT框架以自动发现隐藏因果变量 large language model
20 Large Language Models to Enhance Bayesian Optimization 提出LLAMBO以增强贝叶斯优化的效果 large language model
21 MOMENT: A Family of Open Time-series Foundation Models 提出MOMENT以解决时间序列分析中的数据稀缺问题 foundation model
22 Lens: A Knowledge-Guided Foundation Model for Network Traffic 提出Lens模型以解决网络流量分类与生成问题 foundation model
23 Understanding the Effect of Noise in LLM Training Data with Algorithmic Chains of Thought 提出TInt框架以研究LLM训练数据噪声对性能的影响 large language model chain-of-thought
24 HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal 提出HarmBench框架以解决自动化红队评估标准化问题 large language model ✅
25 Fine-Tuned Language Models Generate Stable Inorganic Materials as Text 提出微调语言模型以生成稳定无机材料 large language model
26 Provably learning a multi-head attention layer 提出可证明学习多头注意力层以解决学习效率问题 large language model
27 Batch Universal Prediction 提出批量普遍预测以评估大语言模型性能 large language model
28 BiLLM: Pushing the Limit of Post-Training Quantization for LLMs 提出BiLLM以解决大语言模型后训练量化性能不足问题 large language model ✅
29 ReLU$^2$ Wins: Discovering Efficient Activation Functions for Sparse LLMs 提出ReLU$^2$以提高稀疏大语言模型的计算效率 large language model

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
30 Reinforcement Learning with Ensemble Model Predictive Safety Certification 提出集成模型预测安全认证以解决强化学习安全性问题 model predictive control reinforcement learning deep reinforcement learning
31 Transductive Reward Inference on Graph 提出图上转导奖励推断方法以解决离线强化学习中的奖励估计问题 locomotion manipulation reinforcement learning
32 SUB-PLAY: Adversarial Policies against Partially Observed Multi-Agent Reinforcement Learning Systems 提出SUB-PLAY以解决部分可观察多智能体强化学习中的安全威胁问题 manipulation reinforcement learning
33 BotSSCL: Social Bot Detection with Self-Supervised Contrastive Learning 提出BotSSCL以解决社交机器人检测中的挑战 manipulation contrastive learning

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#题目一句话要点标签🔗⭐
34 Exploring higher-order neural network node interactions with total correlation 提出Local CorEx以解决高阶变量交互问题 HOI

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

#题目一句话要点标签🔗⭐
35 Modeling Spatio-temporal Dynamical Systems with Neural Discrete Learning and Levels-of-Experts 提出通用专家模块以解决时空动态系统建模问题 optical flow

⬅️ 返回 cs.LG 首页 · 🏠 返回主页