Slot Contrastive Networks: A Contrastive Approach for Representing Objects

TL;DR

This paper introduces Slot Contrastive Networks, a novel unsupervised method that leverages object motion and a contrastive loss to improve object representation in video data, evaluated on Atari games.

Contribution

The paper proposes a contrastive learning approach that focuses on motion cues for unsupervised object representation, avoiding pixel-space losses common in prior static image methods.

Findings

Effective in capturing moving objects in video sequences

Achieves higher diversity in slot representations compared to baselines

Demonstrates strong performance on Atari game data

Abstract

Unsupervised extraction of objects from low-level visual data is an important goal for further progress in machine learning. Existing approaches for representing objects without labels use structured generative models with static images. These methods focus a large amount of their capacity on reconstructing unimportant background pixels, missing low contrast or small objects. Conversely, we present a new method that avoids losses in pixel space and over-reliance on the limited signal a static image provides. Our approach takes advantage of objects' motion by learning a discriminative, time-contrastive loss in the space of slot representations, attempting to force each slot to not only capture entities that move, but capture distinct objects from the other slots. Moreover, we introduce a new quantitative evaluation metric to measure how "diverse" a set of slot vectors are, and use it to evaluate our model on 20 Atari games.