A Cognitively-Inspired Neural Architecture for Visual Abstract Reasoning Using Contrastive Perceptual and Conceptual Processing

TL;DR

This paper presents CPCNet, a neural architecture inspired by human cognition that iteratively integrates perceptual and conceptual processing for visual abstract reasoning, achieving superior accuracy on Raven's Matrices datasets.

Contribution

Introduces CPCNet, a novel contrastive network modeling iterative perceptual and conceptual reasoning, and proposes a more balanced RAVEN dataset variant.

Findings

CPCNet outperforms previous models on RAVEN dataset.

CPCNet uses less inductive bias than prior models.

AB-RAVEN dataset reduces class imbalance issues.

Abstract

We introduce a new neural architecture for solving visual abstract reasoning tasks inspired by human cognition, specifically by observations that human abstract reasoning often interleaves perceptual and conceptual processing as part of a flexible, iterative, and dynamic cognitive process. Inspired by this principle, our architecture models visual abstract reasoning as an iterative, self-contrasting learning process that pursues consistency between perceptual and conceptual processing of visual stimuli. We explain how this new Contrastive Perceptual-Conceptual Network (CPCNet) works using matrix reasoning problems in the style of the well-known Raven's Progressive Matrices intelligence test. Experiments on the machine learning dataset RAVEN show that CPCNet achieves higher accuracy than all previously published models while also using the weakest inductive bias. We also point out a substantial and previously unremarked class imbalance in the original RAVEN dataset, and we propose a new variant of RAVEN -- AB-RAVEN -- that is more balanced in terms of abstract concepts.