Relationships from Entity Stream

TL;DR

This paper introduces a new architecture for relational reasoning that uses an attention-based entity stream, achieving comparable performance to the Relation Network while improving interpretability and reducing model complexity.

Contribution

It presents an attention-driven entity stream model that scales better and offers enhanced interpretability over the traditional Relation Network module.

Findings

Achieves similar accuracy to Relation Network on relational tasks.

Requires fewer parameters than the original RN module.

Provides better interpretability of relational reasoning processes.

Abstract

Relational reasoning is a central component of intelligent behavior, but has proven difficult for neural networks to learn. The Relation Network (RN) module was recently proposed by DeepMind to solve such problems, and demonstrated state-of-the-art results on a number of datasets. However, the RN module scales quadratically in the size of the input, since it calculates relationship factors between every patch in the visual field, including those that do not correspond to entities. In this paper, we describe an architecture that enables relationships to be determined from a stream of entities obtained by an attention mechanism over the input field. The model is trained end-to-end, and demonstrates equivalent performance with greater interpretability while requiring only a fraction of the model parameters of the original RN module.