arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.04013cs.LGcs.AIcs.RO
收藏

Zero-shot Sim2Real跨环境适应

Zero-shot Sim2Real Adaptation Across Environments

Applied Artificial Intelligence Institute(应用人工智能研究院);Deakin University(迪肯大学)

Buddhika Laknath Semage, Thommen George Karimpanal, Santu Rana, Svetha Venkatesh

摘要:

基于仿真的学习通常为机器人技术中的强化学习应用提供了一种具有成本效益的手段。然而,仿真器通常无法准确地复制真实世界的动态,因此弥合sim2real差距是基于仿真的学习中的一个重要问题。目前的解决方案,以弥合sim2real差距涉及混合模拟器,增强神经残差模型。不幸的是,它们需要针对每个单独的环境配置的单独的残差模型(即,诸如质量、摩擦等的环境变量的固定设置),因此不能快速转移到新的环境。为了解决这个问题,我们提出了一个反向动作转换(RAT)的政策,学习模仿在现实世界中的模拟政策。一旦从单个环境中学习,RAT就可以部署在通用策略网络之上,以实现对新环境的zero-shot适应。我们经验性地评估我们的方法在一组连续的控制任务,并观察其优势,作为一个Few-Shot和zero-shot学习竞争基线。

英文摘要:

Simulation based learning often provides a cost-efficient recourse to reinforcement learning applications in robotics. However, simulators are generally incapable of accurately replicating real-world dynamics, and thus bridging the sim2real gap is an important problem in simulation based learning. Current solutions to bridge the sim2real gap involve hybrid simulators that are augmented with neural residual models. Unfortunately, they require a separate residual model for each individual environment configuration (i.e., a fixed setting of environment variables such as mass, friction etc.), and thus are not transferable to new environments quickly. To address this issue, we propose a Reverse Action Transformation (RAT) policy which learns to imitate simulated policies in the real-world. Once learnt from a single environment, RAT can then be deployed on top of a Universal Policy Network to achieve zero-shot adaptation to new environments. We empirically evaluate our approach in a set of continuous control tasks and observe its advantage as a few-shot and zero-shot learner over competing baselines.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑