CURLing the Dream: Contrastive Representations for World Modeling in Reinforcement Learning
Abstract
In this work, we present Curled-Dreamer, a novel reinforcement learning algorithm that integrates contrastive learning into the DreamerV3 framework to enhance performance in visual reinforcement learning tasks. By incorporating the contrastive loss from the CURL algorithm and a reconstruction loss from autoencoder, Curled-Dreamer achieves significant improvements in various DeepMind Control Suite tasks. Our extensive experiments demonstrate that Curled-Dreamer consistently outperforms state-of-the-art algorithms, achieving higher mean and median scores across a diverse set of tasks. The results indicate that the proposed approach not only accelerates learning but also enhances the robustness of the learned policies. This work highlights the potential of combining different learning paradigms to achieve superior performance in reinforcement learning applications.
Cite
@article{arxiv.2408.05781,
title = {CURLing the Dream: Contrastive Representations for World Modeling in Reinforcement Learning},
author = {Victor Augusto Kich and Jair Augusto Bottega and Raul Steinmetz and Ricardo Bedin Grando and Ayano Yorozu and Akihisa Ohya},
journal= {arXiv preprint arXiv:2408.05781},
year = {2024}
}
Comments
Paper accepted for 24th International Conference on Control, Automation and Systems (ICCAS)