Systematic Visual Reasoning through Object-Centric Relational Abstraction

TL;DR

This paper introduces OCRA, a model that combines object-centric representations and relational abstraction to improve systematic visual reasoning and generalization on complex multi-object visual tasks.

Contribution

The paper presents OCRA, a novel model integrating object and relation representations, enhancing systematic reasoning and generalization beyond prior models.

Findings

OCRA achieves strong systematic generalization on complex visual tasks.

OCRA outperforms previous models on the CLEVR-ART dataset.

Object-centric relational abstraction improves visual reasoning capabilities.

Abstract

Human visual reasoning is characterized by an ability to identify abstract patterns from only a small number of examples, and to systematically generalize those patterns to novel inputs. This capacity depends in large part on our ability to represent complex visual inputs in terms of both objects and relations. Recent work in computer vision has introduced models with the capacity to extract object-centric representations, leading to the ability to process multi-object visual inputs, but falling short of the systematic generalization displayed by human reasoning. Other recent models have employed inductive biases for relational abstraction to achieve systematic generalization of learned abstract rules, but have generally assumed the presence of object-focused inputs. Here, we combine these two approaches, introducing Object-Centric Relational Abstraction (OCRA), a model that extracts explicit representations of both objects and abstract relations, and achieves strong systematic generalization in tasks (including a novel dataset, CLEVR-ART, with greater visual complexity) involving complex visual displays.