cs.LG（2024-11-04）

📊 共 18 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (10) 支柱二：RL算法与架构 (RL & Architecture) (4) 支柱一：机器人控制 (Robot Control) (3 🔗1) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (10 篇)

#题目一句话要点标签🔗⭐
1 Defining and Evaluating Physical Safety for Large Language Models 构建无人机物理安全基准，评估大型语言模型在机器人控制中的安全风险 large language model chain-of-thought
2 From Twitter to Reasoner: Understand Mobility Travel Modes and Sentiment Using Large Language Models 利用大型语言模型理解社交媒体中的出行方式和情感倾向 large language model
3 TableGPT2: A Large Multimodal Model with Tabular Data Integration TableGPT2：一个集成表格数据的大型多模态模型，提升表格数据处理能力。 multimodal
4 See it, Think it, Sorted: Large Multimodal Models are Few-shot Time Series Anomaly Analyzers 提出TAMA框架，利用大模型以少量样本实现时间序列异常检测与可解释分析。 multimodal
5 Sparsing Law: Towards Large Language Models with Greater Activation Sparsity 提出PPL-$p\%$稀疏度以提高大语言模型的激活稀疏性 large language model
6 SAFE: Slow and Fast Parameter-Efficient Tuning for Continual Learning with Pre-Trained Models 提出SAFE框架以解决持续学习中的遗忘问题 foundation model
7 Fair In-Context Learning via Latent Concept Variables 提出基于隐概念变量的公平ICL方法，缓解LLM在表格数据预测中的偏见。 large language model
8 GraphXAIN: Narratives to Explain Graph Neural Networks GraphXAIN：利用自然语言叙事解释图神经网络的预测结果 large language model
9 "Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization LLM量化精度-性能权衡研究：为Llama-3.1模型家族选择最优量化方案 large language model
10 Seq-VCR: Preventing Collapse in Intermediate Transformer Representations for Enhanced Reasoning 提出Seq-VCR，解决Transformer中间层表征坍塌问题，提升复杂推理能力 chain-of-thought

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
11 ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy ViTally Consistent：扩展细胞显微镜生物表征学习，构建大规模细胞表征基础模型 representation learning MAE foundation model
12 Revisiting K-mer Profile for Effective and Scalable Genome Representation Learning 重访K-mer谱，提出一种高效且可扩展的基因组表示学习方法，用于宏基因组分箱。 representation learning foundation model
13 LE-PDE++: Mamba for accelerating PDEs Simulations LE-PDE++：利用Mamba加速偏微分方程模拟，提升计算效率。 Mamba
14 Dynamic Weight Adjusting Deep Q-Networks for Real-Time Environmental Adaptation 提出交互式动态评估DQN，解决DQN在动态环境中适应性不足的问题 reinforcement learning deep reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
15 On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback 研究表明，针对用户反馈优化LLM可能导致其学习操纵和欺骗行为。 manipulation reinforcement learning
16 ManiBox: Enhancing Embodied Spatial Generalization via Scalable Simulation Data Generations ManiBox：通过可扩展的模拟数据生成增强具身空间泛化能力 manipulation sim2real zero-shot transfer ✅
17 Enhancing Table Representations with LLM-powered Synthetic Data Generation 提出基于LLM的合成数据生成方法，增强表格表示并提升表格推荐性能。 manipulation representation learning large language model

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
18 Adaptive Sparse Allocation with Mutual Choice & Feature Choice Sparse Autoencoders 提出自适应稀疏分配方法以解决特征选择问题 feature matching foundation model

⬅️ 返回 cs.LG 首页 · 🏠 返回主页