Learner-aware Teaching: Inverse Reinforcement Learning with Preferences and Constraints
This addresses the challenge of effective teaching in AI systems where learners have personal preferences, which is incremental as it builds on existing IRL frameworks.
The paper tackles the problem of teaching in inverse reinforcement learning when the learner has its own preferences, such as biases or constraints, by comparing learner-agnostic teaching (ignoring preferences) with learner-aware teaching (accounting for preferences). It shows that learner-aware teaching algorithms achieve significant performance improvements over learner-agnostic teaching.
Inverse reinforcement learning (IRL) enables an agent to learn complex behavior by observing demonstrations from a (near-)optimal policy. The typical assumption is that the learner's goal is to match the teacher's demonstrated behavior. In this paper, we consider the setting where the learner has its own preferences that it additionally takes into consideration. These preferences can for example capture behavioral biases, mismatched worldviews, or physical constraints. We study two teaching approaches: learner-agnostic teaching, where the teacher provides demonstrations from an optimal policy ignoring the learner's preferences, and learner-aware teaching, where the teacher accounts for the learner's preferences. We design learner-aware teaching algorithms and show that significant performance improvements can be achieved over learner-agnostic teaching.