Interaction-Grounded Learning
This addresses a novel setting for reinforcement learning in applications like prosthetic arms, where no explicit rewards are available, though it appears incremental as it builds on existing RL frameworks with new assumptions.
The paper tackles the problem of learning to interact with an environment without explicit reward signals, proposing Interaction-Grounded Learning where an agent must infer a latent reward from multidimensional feedback to optimize policies. It provides theoretical guarantees and a proof-of-concept empirical evaluation to demonstrate the approach's effectiveness.
Consider a prosthetic arm, learning to adapt to its user's control signals. We propose Interaction-Grounded Learning for this novel setting, in which a learner's goal is to interact with the environment with no grounding or explicit reward to optimize its policies. Such a problem evades common RL solutions which require an explicit reward. The learning agent observes a multidimensional context vector, takes an action, and then observes a multidimensional feedback vector. This multidimensional feedback vector has no explicit reward information. In order to succeed, the algorithm must learn how to evaluate the feedback vector to discover a latent reward signal, with which it can ground its policies without supervision. We show that in an Interaction-Grounded Learning setting, with certain natural assumptions, a learner can discover the latent reward and ground its policy for successful interaction. We provide theoretical guarantees and a proof-of-concept empirical evaluation to demonstrate the effectiveness of our proposed approach.