Navigation with QPHIL: Quantizing Planner for Hierarchical Implicit Q-Learning
This work addresses a key problem in offline RL for navigation, offering a novel method to improve performance in complex environments, though it builds on existing hierarchical and transformer-based techniques.
The paper tackles the challenge of mitigating incorrect policy updates due to noisy value estimates in offline reinforcement learning for complex navigation tasks by introducing a hierarchical transformer-based approach with a learned quantizer of the space, achieving state-of-the-art results.
Offline Reinforcement Learning (RL) has emerged as a powerful alternative to imitation learning for behavior modeling in various domains, particularly in complex navigation tasks. An existing challenge with Offline RL is the signal-to-noise ratio, i.e. how to mitigate incorrect policy updates due to errors in value estimates. Towards this, multiple works have demonstrated the advantage of hierarchical offline RL methods, which decouples high-level path planning from low-level path following. In this work, we present a novel hierarchical transformer-based approach leveraging a learned quantizer of the space. This quantization enables the training of a simpler zone-conditioned low-level policy and simplifies planning, which is reduced to discrete autoregressive prediction. Among other benefits, zone-level reasoning in planning enables explicit trajectory stitching rather than implicit stitching based on noisy value function estimates. By combining this transformer-based planner with recent advancements in offline RL, our proposed approach achieves state-of-the-art results in complex long-distance navigation environments.