arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2312.05044cs.LGcs.AI
收藏

面向目标条件策略的逆向学习

Backward Learning for Goal-Conditioned Policies

Technical University of Dortmund(多特蒙德工业大学)

Marc Höftmann, Jan Robine, Stefan Harmeling

AI总结:

该研究针对无奖励强化学习下的目标条件策略学习问题,提出先构建逆向世界模型、生成逆向轨迹再经最短路径优化、最后模仿学习的多步方法,在64×64像素鸟瞰图迷宫环境中验证了方法可稳定抵达多目标。

摘要:

我们能否在无奖励的情况下通过强化学习学习策略?能否仅通过尝试抵达目标状态来学习策略?我们给出了肯定答案,提出一套多步骤流程:首先学习时间逆向的世界模型,其次生成抵达目标的逆向轨迹,第三用最短路径搜索算法优化这些序列,最后通过模仿学习训练神经网络策略。我们在观测为64×64像素鸟瞰图的确定性迷宫环境中评估该方法,结果表明它能稳定抵达多个目标。

英文摘要:

Can we learn policies in reinforcement learning without rewards? Can we learn a policy just by trying to reach a goal state? We answer these questions positively by proposing a multi-step procedure that first learns a world model that goes backward in time, secondly generates goal-reaching backward trajectories, thirdly improves those sequences using shortest path finding algorithms, and finally trains a neural network policy by imitation learning. We evaluate our method on a deterministic maze environment where the observations are $64\times 64$ pixel bird's eye images and can show that it consistently reaches several goals.

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。

↑