cs.LG（2025-06-21）

📊 共 15 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 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (4) 支柱一：机器人控制 (Robot Control) (1)

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

#题目一句话要点标签🔗⭐
1 An Interpretable Transformer-Based Foundation Model for Cross-Procedural Skill Assessment Using Raw fNIRS Signals 提出基于Transformer的可解释fNIRS基础模型，用于跨程序技能评估。 foundation model
2 PhysiX: A Foundation Model for Physics Simulations PhysiX：用于物理模拟的45亿参数自回归生成式基础模型 foundation model
3 A Comparative Study of Open-Source Libraries for Synthetic Tabular Data Generation: SDV vs. SynthCity 对比SDV与SynthCity：评估开源库在合成表格数据生成中的性能，并分析其统计相似性和预测效用。 large language model
4 Causal Spherical Hypergraph Networks for Modelling Social Uncertainty 提出Causal-SphHN，建模社会不确定性中的因果关系和高阶交互，提升预测精度。 multimodal
5 Safe Pruning LoRA: Robust Distance-Guided Pruning for Safety Alignment in Adaptation of LLMs 提出SPLoRA，通过稳健的距离引导剪枝提升LoRA适配大模型的安全性对齐。 large language model ✅
6 FaithfulSAE: Towards Capturing Faithful Features with Sparse Autoencoders without External Dataset Dependencies FaithfulSAE：利用模型自生成数据训练稀疏自编码器，提升特征忠实度 large language model
7 Online Multi-LLM Selection via Contextual Bandits under Unstructured Context Evolution 提出基于上下文Bandit的在线多LLM选择框架，解决非结构化上下文演化问题 large language model
8 Time-Prompt: Integrated Heterogeneous Prompts for Unlocking LLMs in Time Series Forecasting 提出Time-Prompt框架，利用异构提示解锁LLM在时间序列预测中的潜力 large language model
9 EQuARX: Efficient Quantized AllReduce in XLA for Distributed Machine Learning Acceleration 提出EQuARX，在XLA中实现高效量化AllReduce，加速分布式机器学习。 large language model
10 From Tiny Machine Learning to Tiny Deep Learning: A Survey 综述TinyML到TinyDL的演进：边缘AI的架构、优化与应用 foundation model

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

#题目一句话要点标签🔗⭐
11 Aligning Frozen LLMs by Reinforcement Learning: An Iterative Reweight-then-Optimize Approach 提出迭代重加权优化（IRO）框架，无需微调即可对齐冻结LLM。 reinforcement learning RLHF DPO
12 Research on Low-Latency Inference and Training Efficiency Optimization for Graph Neural Network and Large Language Model-Based Recommendation Systems 针对GNN与LLM混合推荐系统，提出低延迟推理与高效训练的优化方案 distillation large language model
13 Beyond instruction-conditioning, MoTE: Mixture of Task Experts for Multi-task Embedding Models 提出MoTE：用于多任务嵌入模型的混合任务专家方法，提升低容量模型性能。 representation learning contrastive learning
14 Predicting E-commerce Purchase Behavior using a DQN-Inspired Deep Learning Model for enhanced adaptability 提出DQN启发的深度学习模型，预测电商购买意图并提升适应性。 reinforcement learning predictive model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
15 Accelerating Residual Reinforcement Learning with Uncertainty Estimation 利用不确定性估计加速残差强化学习，提升样本效率并支持随机策略。 sim-to-real reinforcement learning

⬅️ 返回 cs.LG 首页 · 🏠 返回主页