arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2206.14802cs.ROcs.AIcs.LG
收藏

具有局部动力学模型的视觉前景

Visual Foresight With a Local Dynamics Model

Khoury College of Computer Sciences(胡里计算机学院);Northeastern University(东北大学)

Colin Kohler, Robert Platt

摘要:

无模型策略学习已被证明能够学习操作策略,可以使用单步操作原语解决长时间任务。然而,训练这些策略是一个耗时的过程,需要大量的数据。我们提出了本地动力学模型(LDM),有效地学习这些操作原语的状态转换函数。通过将LDM与无模型策略学习相结合,我们可以学习策略,这些策略可以使用一步前瞻规划来解决复杂的操作任务。我们表明,LDM是更采样效率和优于其他模型架构。当与规划相结合时,我们可以在模拟中的几个具有挑战性的操作任务上优于其他基于模型和无模型的策略。

英文摘要:

Model-free policy learning has been shown to be capable of learning manipulation policies which can solve long-time horizon tasks using single-step manipulation primitives. However, training these policies is a time-consuming process requiring large amounts of data. We propose the Local Dynamics Model (LDM) which efficiently learns the state-transition function for these manipulation primitives. By combining the LDM with model-free policy learning, we can learn policies which can solve complex manipulation tasks using one-step lookahead planning. We show that the LDM is both more sample-efficient and outperforms other model architectures. When combined with planning, we can outperform other model-based and model-free policies on several challenging manipulation tasks in simulation.

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。

↑