cs.LG（2025-04-30）

📊 共 13 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (7 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (7 篇)

#题目一句话要点标签🔗⭐
1 When Reasoning Beats Scale: A 1.5B Reasoning Model Outranks 13B LLMs as Discriminator 提出基于推理的判别器，15亿参数模型在文本到SQL任务上超越130亿参数LLM。 large language model chain-of-thought
2 Generative Machine Learning in Adaptive Control of Dynamic Manufacturing Processes: A Review 综述性论文：探讨生成式机器学习在动态制造过程自适应控制中的应用 multimodal
3 Insulin Resistance Prediction From Wearables and Routine Blood Biomarkers 结合可穿戴设备与血液生物标志物，利用深度学习预测胰岛素抵抗 large language model
4 Traceback of Poisoning Attacks to Retrieval-Augmented Generation 提出RAGForensics以解决检索增强生成系统的中毒攻击问题 large language model ✅
5 MatMMFuse: Multi-Modal Fusion model for Material Property Prediction MatMMFuse：融合晶体图和文本信息，提升材料属性预测精度。 large language model
6 Towards Film-Making Production Dialogue, Narration, Monologue Adaptive Moving Dubbing Benchmarks 提出TA-Dubbing基准，自适应电影制作中对话、旁白和独白的移动配音评估。 large language model
7 COSMOS: Predictable and Cost-Effective Adaptation of LLMs 提出COSMOS以高效预测LLMs的适应性与成本 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
8 FAST-Q: Fast-track Exploration with Adversarially Balanced State Representations for Counterfactual Action Estimation in Offline Reinforcement Learning FAST-Q通过对抗平衡状态表示加速离线强化学习中的反事实动作探索，提升游戏平台用户价值。 reinforcement learning offline reinforcement learning conservative q-learning
9 Dragonfly: a modular deep reinforcement learning library Dragonfly：一个模块化的深度强化学习库，易于实验和开发 reinforcement learning deep reinforcement learning
10 Recursive KL Divergence Optimization: A Dynamic Framework for Representation Learning 提出递归KL散度优化框架RKDO，提升表征学习效率与稳定性。 representation learning contrastive learning
11 Synergy-CLIP: Extending CLIP with Multi-modal Integration for Robust Representation Learning Synergy-CLIP：通过多模态融合扩展CLIP，实现鲁棒的表征学习 representation learning
12 ABG-NAS: Adaptive Bayesian Genetic Neural Architecture Search for Graph Representation Learning 提出ABG-NAS，用于自适应贝叶斯遗传图神经网络架构搜索，提升图表示学习效果。 representation learning ✅

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

#题目一句话要点标签🔗⭐
13 OpenAVS: Training-Free Open-Vocabulary Audio Visual Segmentation with Foundational Models 提出OpenAVS，利用多媒体基础模型实现免训练的开放词汇音视频分割 open-vocabulary open vocabulary foundation model

⬅️ 返回 cs.LG 首页 · 🏠 返回主页