Unsupervised Causal Generative Understanding of Images

Anciukevičius, T., Fox-Roberts, P., Rosten, E. and Henderson, P. (2022) Unsupervised Causal Generative Understanding of Images. Workshop on Spurious Correlations, Invariance and Stability at the International Conference on Machine Learning (ICML 2022), Baltimore, MD, USA, 22 Jul 2022.

Abstract

We present a novel causal generative model for unsupervised object-centric 3D scene understanding that generalizes robustly to out-of-distribution images. This model is trained to reconstruct multi-view images via a latent representation describing the shapes, colours and positions of the 3D objects they show. We then propose an inference algorithm that can infer this latent representation given a single out-of-distribution image as input. We conduct extensive experiments applying our approach to test datasets that have zero probability under the training distribution. Our approach significantly out-performs baselines that do not capture the true causal image generation process.

• Item Type: Conference or Workshop Item
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Henderson, Dr Paul
• Authors: Anciukevičius, T., Fox-Roberts, P., Rosten, E., and Henderson, P.
• College/School: College of Science and Engineering > School of Computing Science
• Copyright Holders: Copyright © 2022 by the author(s).
• Publisher Policy: Reproduced with the permission of the authors

Related URLs

• Organisation