Pre-Conference Talk by LI Wenjun | Improving Environment Novelty Quantification for Effective Unsupervised Environment Design

Unsupervised Environment Design (UED) formalizes the problem of autocur- ricula through interactive training between a teacher agent and a student agent. The teacher generates new training environments with high learning potential, curating an adaptive curriculum that strengthens the student’s ability to handle unseen scenarios. Existing UED methods mainly rely on regret, a metric that measures the difference between the agent’s optimal and actual performance, to guide curriculum design. Regret-driven methods generate curricula that progres- sively increase environment complexity for the student but overlook environment novelty–a critical element for enhancing an agent’s generalizability. Measuring environment novelty is especially challenging due to the underspecified nature of environment parameters in UED, and existing approaches face significant limi- tations. To address this, this paper introduces the Coverage-based Evaluation of Novelty In Environment (CENIE) framework. CENIE proposes a scalable, domain- agnostic, and curriculum-aware approach to quantifying environment novelty by leveraging the student’s state-action space coverage from previous curriculum experiences. We then propose an implementation of CENIE that models this cov- erage and measures environment novelty using Gaussian Mixture Models. By integrating both regret and novelty as complementary objectives for curriculum design, CENIE facilitates effective exploration across the state-action space while progressively increasing curriculum complexity. Empirical evaluations demonstrate that augmenting existing regret-based UED algorithms with CENIE achieves state- of-the-art performance across multiple benchmarks, underscoring the effectiveness of novelty-driven autocurricula for robust generalization.

This is a Pre-Conference talk for The Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS 2024).

LI Wenjun is a Ph.D. candidate in Computer Science at the SMU School of Computing and Information Systems, supervised by Professor Pradeep VARAKANTHAM. His research aims to design and build open-ended systems which continuously propose new tasks for RL agents to solve, ultimately producing generally capable agents.