Advances in Neural Information Processing Systems 37 / 测试间自进化

DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning

作者：Hao Bai, Yifei Zhou, Mert Cemri, Jiayi Pan, Alane Suhr, Sergey Levine, Aviral Kumar · 单位：UC Berkeley, Google DeepMind · 会议/期刊：Advances in Neural Information Processing Systems 37 · 方向：测试间自进化 · 发布日期：2024-06-14

自我进化智能体测试间自进化模型自进化测试时学习奖励学习
新标签打开 AlphaXiv新标签打开 Project
DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning figure
AlphaXiv 中文概览（可滚动查看）