Intrinsically Motivated Multimodal Structure Learning
This work addresses robot learning of object affordances for planning tasks, but it appears incremental as it applies existing structure learning techniques to multimodal representations.
The paper tackles the problem of learning transition dynamics for robot interactions with semi-permanent structures by developing an intrinsically motivated multimodal structure learning method, resulting in models that predict state distribution changes and show performance improvements as transition actions increase in experiments on a bimanual mobile manipulator.
We present a long-term intrinsically motivated structure learning method for modeling transition dynamics during controlled interactions between a robot and semi-permanent structures in the world. In particular, we discuss how partially-observable state is represented using distributions over a Markovian state and build models of objects that predict how state distributions change in response to interactions with such objects. These structures serve as the basis for a number of possible future tasks defined as Markov Decision Processes (MDPs). The approach is an example of a structure learning technique applied to a multimodal affordance representation that yields a population of forward models for use in planning. We evaluate the approach using experiments on a bimanual mobile manipulator (uBot-6) that show the performance of model acquisition as the number of transition actions increases.