Learning Object-Centric Video Models by Contrasting Sets

TL;DR

This paper proposes a set-based contrastive learning method with attention mechanisms for object-centric video modeling, improving object separation and prediction over previous slot-based contrastive approaches.

Contribution

It introduces a global set-based contrastive loss and attention-based encoders, enhancing object representation learning and interpretability in video models.

Findings

Outperforms previous contrastive methods in reconstruction quality

Achieves better future prediction accuracy

Provides interpretable object masks

Abstract

Contrastive, self-supervised learning of object representations recently emerged as an attractive alternative to reconstruction-based training. Prior approaches focus on contrasting individual object representations (slots) against one another. However, a fundamental problem with this approach is that the overall contrastive loss is the same for (i) representing a different object in each slot, as it is for (ii) (re-)representing the same object in all slots. Thus, this objective does not inherently push towards the emergence of object-centric representations in the slots. We address this problem by introducing a global, set-based contrastive loss: instead of contrasting individual slot representations against one another, we aggregate the representations and contrast the joined sets against one another. Additionally, we introduce attention-based encoders to this contrastive setup which simplifies training and provides interpretable object masks. Our results on two synthetic video datasets suggest that this approach compares favorably against previous contrastive methods in terms of reconstruction, future prediction and object separation performance.