Optimal Goal-Reaching Reinforcement Learning via Quasimetric Learning
This work addresses goal-reaching RL for robotics and control applications, representing an incremental advance with a novel method for a known bottleneck.
The paper tackled the problem of learning optimal value functions in goal-reaching reinforcement learning by introducing Quasimetric Reinforcement Learning (QRL), which leverages quasimetric models and demonstrated improved sample efficiency and performance on benchmarks.
In goal-reaching reinforcement learning (RL), the optimal value function has a particular geometry, called quasimetric structure. This paper introduces Quasimetric Reinforcement Learning (QRL), a new RL method that utilizes quasimetric models to learn optimal value functions. Distinct from prior approaches, the QRL objective is specifically designed for quasimetrics, and provides strong theoretical recovery guarantees. Empirically, we conduct thorough analyses on a discretized MountainCar environment, identifying properties of QRL and its advantages over alternatives. On offline and online goal-reaching benchmarks, QRL also demonstrates improved sample efficiency and performance, across both state-based and image-based observations.