Environment Generation for Zero-Shot Compositional Reinforcement Learning

TL;DR

This paper introduces CoDE, a method for automatically generating compositional environments to train RL agents, enabling better learning, robustness, and zero-shot generalization to complex, multi-step tasks.

Contribution

The paper proposes a novel environment generation algorithm for compositional tasks, along with two benchmark frameworks, improving RL training and generalization in complex scenarios.

Findings

CoDE achieves 4x higher success rate than baselines.

Successfully trains agents on real web navigation tasks.

Generates environments with multiple pages or rooms for complex task learning.

Abstract

Many real-world problems are compositional - solving them requires completing interdependent sub-tasks, either in series or in parallel, that can be represented as a dependency graph. Deep reinforcement learning (RL) agents often struggle to learn such complex tasks due to the long time horizons and sparse rewards. To address this problem, we present Compositional Design of Environments (CoDE), which trains a Generator agent to automatically build a series of compositional tasks tailored to the RL agent's current skill level. This automatic curriculum not only enables the agent to learn more complex tasks than it could have otherwise, but also selects tasks where the agent's performance is weak, enhancing its robustness and ability to generalize zero-shot to unseen tasks at test-time. We analyze why current environment generation techniques are insufficient for the problem of generating compositional tasks, and propose a new algorithm that addresses these issues. Our results assess learning and generalization across multiple compositional tasks, including the real-world problem of learning to navigate and interact with web pages. We learn to generate environments composed of multiple pages or rooms, and train RL agents capable of completing wide-range of complex tasks in those environments. We contribute two new benchmark frameworks for generating compositional tasks, compositional MiniGrid and gMiniWoB for web navigation.CoDE yields 4x higher success rate than the strongest baseline, and demonstrates strong performance of real websites learned on 3500 primitive tasks.