Learning body-affordances to simplify action spaces
This addresses the problem of simplifying action spaces for embodied agents, but it appears incremental as it builds on recent machine learning approaches with conceptual simplifications.
The paper tackles the challenge of controlling embodied agents with many degrees of freedom by proposing a method to discover and interpolate between high-level actions or body-affordances, providing a low-dimensional interface for high-dimensional action policies, though no concrete numerical results are reported.
Controlling embodied agents with many actuated degrees of freedom is a challenging task. We propose a method that can discover and interpolate between context dependent high-level actions or body-affordances. These provide an abstract, low-dimensional interface indexing high-dimensional and time- extended action policies. Our method is related to recent ap- proaches in the machine learning literature but is conceptually simpler and easier to implement. More specifically our method requires the choice of a n-dimensional target sensor space that is endowed with a distance metric. The method then learns an also n-dimensional embedding of possibly reactive body-affordances that spread as far as possible throughout the target sensor space.