Unsupervised Environment Design (UED) is a paradigm that automatically generates a curriculum of training environments, enabling agents trained in these environments to develop general capabilities, i.e., achieving good zero-shot transfer performance. However, existing UED approaches focus primarily on the random generation of environments for open-ended agent training. This is impractical in resource-limited scenarios where there is a constraint on the number of environments that can be generated. In this paper, we introduce a hierarchical MDP framework for environment design under resource constraints. It consists of an upper-level RL teacher agent that generates suitable training environments for a lower-level student agent. The RL teacher can leverage previously discovered environment structures and generate environments at the frontier of the student's capabilities by observing the student policy's representation. Additionally, to alleviate the time-consuming process of collecting the experience of the upper-level teacher, we utilize recent advances in generative modeling to synthesize a trajectory dataset for training the teacher agent. Our method significantly reduces the resource-intensive interactions between agents and environments, and empirical experiments across various domains demonstrate the effectiveness of our approach.
Keywords: hierarchical MDP, environment design, general capability, generative model
TL;DR: In this paper, we introduce a hierarchical MDP framework for environment design under resource constraints, and we utilize recent advances in generative modeling to synthesize a trajectory dataset for training the teacher agent.
Abstract:
Supplementary Material: zip
Primary Area: Reinforcement learning
Submission Number: 11014
Loading