Self-Supervised Relational Reasoning for Representation Learning

TL;DR

This paper introduces a self-supervised relational reasoning approach that enhances representation learning by discriminating entity relationships, leading to improved performance in downstream tasks like classification and image retrieval.

Contribution

The work presents a novel self-supervised relational reasoning framework that leverages intra- and inter-entity discrimination to produce richer neural representations, outperforming existing methods.

Findings

Outperforms state-of-the-art by 3% in accuracy

Achieves an average of 14% improvement over competitors

Links effectiveness to mutual information maximization

Abstract

In self-supervised learning, a system is tasked with achieving a surrogate objective by defining alternative targets on a set of unlabeled data. The aim is to build useful representations that can be used in downstream tasks, without costly manual annotation. In this work, we propose a novel self-supervised formulation of relational reasoning that allows a learner to bootstrap a signal from information implicit in unlabeled data. Training a relation head to discriminate how entities relate to themselves (intra-reasoning) and other entities (inter-reasoning), results in rich and descriptive representations in the underlying neural network backbone, which can be used in downstream tasks such as classification and image retrieval. We evaluate the proposed method following a rigorous experimental procedure, using standard datasets, protocols, and backbones. Self-supervised relational reasoning outperforms the best competitor in all conditions by an average 14% in accuracy, and the most recent state-of-the-art model by 3%. We link the effectiveness of the method to the maximization of a Bernoulli log-likelihood, which can be considered as a proxy for maximizing the mutual information, resulting in a more efficient objective with respect to the commonly used contrastive losses.