World models trained by contrastive learning are a compelling alternative to autoencoder-based world models, which learn by reconstructing pixel states. In this paper, we describe three cases where small changes in how we sample negative states in the contrastive loss lead to drastic changes in model performance. In previously studied Atari datasets, we show that leveraging time step correlations can double the performance of the Contrastive Structured World Model. We also collect a full version of the datasets to study contrastive learning under a more diverse set of experiences.
@article{arxiv.2107.11676,
title = {The Impact of Negative Sampling on Contrastive Structured World Models},
author = {Ondrej Biza and Elise van der Pol and Thomas Kipf},
journal= {arXiv preprint arXiv:2107.11676},
year = {2021}
}
Comments
This work appeared at the ICML 2021 Workshop: Self-Supervised Learning for Reasoning and Perception