Homomorphism Autoencoder -- Learning Group Structured Representations from Observed Transitions

arXiv2207.12067MaRDI QIDQ6405934FDOQ6405934

Authors: Hamza Keurti, Hsiao-Ru Pan, Michel Besserve, Benjamin F. Grewe, Bernhard Schölkopf 

Publication date: 25 July 2022

Abstract: How can agents learn internal models that veridically represent interactions with the real world is a largely open question. As machine learning is moving towards representations containing not just observational but also interventional knowledge, we study this problem using tools from representation learning and group theory. We propose methods enabling an agent acting upon the world to learn internal representations of sensory information that are consistent with actions that modify it. We use an autoencoder equipped with a group representation acting on its latent space, trained using an equivariance-derived loss in order to enforce a suitable homomorphism property on the group representation. In contrast to existing work, our approach does not require prior knowledge of the group and does not restrict the set of actions the agent can perform. We motivate our method theoretically, and show empirically that it can learn a group representation of the actions, thereby capturing the structure of the set of transformations applied to the environment. We further show that this allows agents to predict the effect of sequences of future actions with improved accuracy.

Has companion code repository: https://github.com/hamzakeurti/homomorphismvae

This page was built for publication: Homomorphism Autoencoder -- Learning Group Structured Representations from Observed Transitions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6405934)