Hierarchical Programmatic Reinforcement Learning via Learning to Compose Programs

TL;DR

This paper introduces a hierarchical reinforcement learning framework that learns to compose programs from an embedding space, enabling the creation of more complex, interpretable policies that generalize better to new scenarios.

Contribution

It proposes a novel hierarchical programmatic RL approach that learns to compose programs, overcoming dataset limitations and improving credit assignment for complex behaviors.

Findings

Outperforms baselines in the Karel domain

Learns to compose programs for complex behaviors

Addresses dataset distribution limitations

Abstract

Aiming to produce reinforcement learning (RL) policies that are human-interpretable and can generalize better to novel scenarios, Trivedi et al. (2021) present a method (LEAPS) that first learns a program embedding space to continuously parameterize diverse programs from a pre-generated program dataset, and then searches for a task-solving program in the learned program embedding space when given a task. Despite the encouraging results, the program policies that LEAPS can produce are limited by the distribution of the program dataset. Furthermore, during searching, LEAPS evaluates each candidate program solely based on its return, failing to precisely reward correct parts of programs and penalize incorrect parts. To address these issues, we propose to learn a meta-policy that composes a series of programs sampled from the learned program embedding space. By learning to compose programs, our proposed hierarchical programmatic reinforcement learning (HPRL) framework can produce program policies that describe out-of-distributionally complex behaviors and directly assign credits to programs that induce desired behaviors. The experimental results in the Karel domain show that our proposed framework outperforms baselines. The ablation studies confirm the limitations of LEAPS and justify our design choices.