Efficient Supervision for Robot Learning via Imitation, Simulation, and Adaptation
This work addresses data efficiency challenges in robot learning, which is incremental as it builds on existing methods without introducing a new paradigm.
The paper tackles the problem of inefficient data collection and curation pipelines in robot learning by targeting imitation learning, domain adaptation, and simulation transfer to increase efficiency through more powerful data sources and better information extraction.
Recent successes in machine learning have led to a shift in the design of autonomous systems, improving performance on existing tasks and rendering new applications possible. Data-focused approaches gain relevance across diverse, intricate applications when developing data collection and curation pipelines becomes more effective than manual behaviour design. The following work aims at increasing the efficiency of this pipeline in two principal ways: by utilising more powerful sources of informative data and by extracting additional information from existing data. In particular, we target three orthogonal fronts: imitation learning, domain adaptation, and transfer from simulation.