Being able to solve a task in diverse ways makes agents more robust to task variations and less prone to local optima. In this context, constrained diversity optimization has become a useful reinforcement learning (RL) framework for training a set of diverse agents in parallel. However, existing constrained-diversity RL methods often under-explore in complex tasks such as robot manipulation, resulting in limited behavioral diversity. We address this with a two-stage curriculum that introduces a spline-based trajectory prior as an inductive bias to produce diverse, high-reward behaviors in an initial stage, and then distills these behaviors into reactive, step-wise policies in a second stage. In our empirical evaluation, we provide novel insights into challenges of diversity-targeted training and show that our curriculum increases the diversity of learned skills while maintaining high task performance.
@article{arxiv.2506.01568,
title = {Trajectory First: A Curriculum for Discovering Diverse Policies},
author = {Cornelius V. Braun and Sayantan Auddy and Marc Toussaint},
journal= {arXiv preprint arXiv:2506.01568},
year = {2026}
}
Comments
Accepted into the Inductive Biases in Reinforcement Learning Workshop at RLC 2025