TACO: Learning task decomposition via temporal alignment for control

Conference item

Many advanced Learning from Demonstration (LfD) methods consider the decomposition of complex, real-world tasks into simpler sub-tasks. By reusing the corresponding sub-policies within and between tasks, they provide training data for each policy from different high-level tasks and compose them to perform novel ones. However, most existing approaches to modular LfD focus either on learning a single high-level task or depend on domain knowledge and temporal segmentation. By contrast, we propose a weakly supervised, domain-agnostic approach based on task sketches, which include only the sequence of sub-tasks performed in each demonstration. Our approach simultaneously aligns the sketches with the observed demonstrations and learns the required sub-policies, which improves generalisation in comparison to separate optimisation procedures. We evaluate the approach on multiple domains, including a simulated 3D robot arm control task using purely image-based observations. The approach performs commensurately with fully supervised approaches, while requiring significantly less annotation effort, and significantly outperforms methods which separate segmentation and imitation.

Authors: Shiarlis, K; Wulfmeier, M; Salter, S; Whiteson, S; Posner, H

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Journal of Machine Learning Research
• Host title: International Conference on Machine Learning
• Journal: Thirty-fifth International Conference on Machine Learning (ICML 2018)
• Publication date: 2018-07-03
• Acceptance date: 2018-06-12