论文HF Daily Papers · arXiv 2611.183979 天前
Gamer框架:基于键鼠轨迹预训练实现虚实环境长程实时交互
本文由 AI 根据公开内容整理,可能有误,请以原文为准:Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent(HF Daily Papers · arXiv 2611.18397)
核心内容
研究团队发布Gamer通用智能体框架,专为复杂交互环境中的长程任务设计。该模型基于数万小时设备级键盘鼠标交互轨迹进行持续预训练,覆盖模拟游戏与具身环境,使模型能够跨越物理与虚拟域学习环境动力学。通过稀疏思维指令微调降低冗余计算,并采用稳定的多轮强化学习方法解决有限上下文窗口、累积误差及稀疏奖励问题,在多个高难度3D开放世界游戏及2D基准测试中实现了媲美熟练人类玩家的性能表现。
技术架构采用端到端设计,融合感知、记忆管理与多步推理,共享训练部署管线以兼容2D图形界面与3D仿真世界。针对实时交互需求,构建了云边协同推理引擎,支持无界上下文执行与多环境并发服务,有效维持低在线推理延迟。实验表明,即便是重复性的键鼠操作轨迹,经规模化数据与算法优化后也能显著提升复杂技能学习效率,验证了利用数字交互数据训练跨域通用具身智能体的可行性。
关键看点
- 数据预训练:使用数万小时设备级键鼠交互轨迹进行持续预训练,涵盖模拟游戏与具身环境,赋予模型跨虚实域的环境动力学理解能力
- 算法优化:引入稀疏思维指令微调减少冗余计算,配合稳定多轮强化学习策略,有效克服长程交互中的上下文限制、误差累积与稀疏奖励难题
- 推理架构:部署云边协同推理引擎支持无界上下文与多环境并发,在保障低延迟实时交互的同时,实现3D开放世界中媲美人类玩家的操作效率
工程与产业观察
Gamer验证了海量人机交互轨迹作为具身智能预训练语料的有效性,其稀疏思维与云边推理架构为解决长程任务中的算力瓶颈与上下文遗忘提供了工程范式。该研究打通了数字世界操作经验向物理具身能力的迁移路径,对构建低成本、高泛化的通用机器人基础模型具有重要参考价值。
英文摘要 (Abstract)
We introduce Gamer, a generalist agent framework built for long-horizon tasks in complex interactive environments. Its end-to-end architecture combines perception, memory management, and multi-step reasoning. A shared training and deployment pipeline supports interaction with both conventional 2D graphical interfaces and complex 3D simulated worlds. We first continue pre-training on a dataset constructed from tens of thousands of hours of device-level keyboard and mouse interaction. These trajectories span simulated games and embodied environments, allowing the model to learn environmental dynamics across both physical and virtual domains. Sparse-thinking instruction tuning then substantially reduces redundant computation while preserving the reasoning needed for complex decisions. We train the policy for long-horizon interaction with a stable multi-turn reinforcement learning method that addresses finite context windows, accumulated rollout errors, and sparse rewards. For real-time interaction, a cloud-edge inference engine serves numerous simulated game environments concurrently, supports unbounded-context execution, and keeps online inference latency low. Across multiple challenging 3D open-world games, Gamer demonstrates solid generalization and operational efficiency comparable to skilled players. It also achieves competitive results on a range of conventional 2D game benchmarks. These results suggest that even repetitive keyboard and mouse trajectories can support substantial gains from scaling data and learning algorithms. Training on these trajectories helps Gamer learn complex skills more steadily in environments with both physical and digital properties. This approach offers a path toward generalist agents that can interact across digital and physical worlds.
Hugging Face Daily Papers 热度:83 upvotes 作者:Run Luo, Kuan Zhang, Qi Deng, Yukun Chen, Zhihao Yang, Yufei Zhang, Renrong Hu, Yi Zhang, Guiyu Ma, Tianyu Fan, Longze Chen, Hao Sun, Jiaming Li, Jiaming Yang, Xiaohao Liu, Tao Tan, Minzheng Wang, Han…
外部来源
本页事实对应的公开页面,以原文为准。(点击将离开本站)