GLIB: Efficient Exploration for Relational Model-Based Reinforcement Learning via Goal-Literal Babbling
This addresses the problem of exploration efficiency in relational model-based RL for researchers and practitioners, offering a novel approach with proven advantages.
The paper tackles efficient exploration for relational model-based reinforcement learning without extrinsic rewards by proposing goal-literal babbling (GLIB), which samples relational conjunctive goals and plans to achieve them. The method shows strong experimental outperformance over existing methods in prediction and planning tasks, with theoretical guarantees of convergence to the ground truth model.
We address the problem of efficient exploration for transition model learning in the relational model-based reinforcement learning setting without extrinsic goals or rewards. Inspired by human curiosity, we propose goal-literal babbling (GLIB), a simple and general method for exploration in such problems. GLIB samples relational conjunctive goals that can be understood as specific, targeted effects that the agent would like to achieve in the world, and plans to achieve these goals using the transition model being learned. We provide theoretical guarantees showing that exploration with GLIB will converge almost surely to the ground truth model. Experimentally, we find GLIB to strongly outperform existing methods in both prediction and planning on a range of tasks, encompassing standard PDDL and PPDDL planning benchmarks and a robotic manipulation task implemented in the PyBullet physics simulator. Video: https://youtu.be/F6lmrPT6TOY Code: https://git.io/JIsTB