Object-centric architectures enable efficient causal representation learning

TL;DR

This paper introduces an object-centric architecture that combines weak supervision and causal learning to improve disentanglement of object properties in images, addressing limitations of traditional methods with non-injective generative functions.

Contribution

It modifies the Slot Attention architecture to incorporate weak supervision, enabling more data-efficient disentanglement of object properties in images.

Findings

Successfully disentangles object properties in image experiments

Requires fewer perturbations than Euclidean-based approaches

Addresses non-injectivity issues in object-based generative models

Abstract

Causal representation learning has showed a variety of settings in which we can disentangle latent variables with identifiability guarantees (up to some reasonable equivalence class). Common to all of these approaches is the assumption that (1) the latent variables are represented as $d$-dimensional vectors, and (2) that the observations are the output of some injective generative function of these latent variables. While these assumptions appear benign, we show that when the observations are of multiple objects, the generative function is no longer injective and disentanglement fails in practice. We can address this failure by combining recent developments in object-centric learning and causal representation learning. By modifying the Slot Attention architecture arXiv:2006.15055, we develop an object-centric architecture that leverages weak supervision from sparse perturbations to disentangle each object's properties. This approach is more data-efficient in the sense that it requires significantly fewer perturbations than a comparable approach that encodes to a Euclidean space and we show that this approach successfully disentangles the properties of a set of objects in a series of simple image-based disentanglement experiments.