Advancing Generalization Across a Variety of Abstract Visual Reasoning Tasks

TL;DR

This paper introduces PoNG, a new neural architecture with group convolution and normalization, that significantly improves out-of-distribution generalization across various abstract visual reasoning benchmarks.

Contribution

The paper proposes PoNG, a novel neural network architecture designed to enhance generalization in abstract visual reasoning tasks, especially in out-of-distribution scenarios.

Findings

PoNG outperforms existing methods on several AVR benchmarks.

Strong generalization capabilities demonstrated across synthetic and real-world images.

Effective in both i.i.d. and o.o.d. settings.

Abstract

The abstract visual reasoning (AVR) domain presents a diverse suite of analogy-based tasks devoted to studying model generalization. Recent years have brought dynamic progress in the field, particularly in i.i.d. scenarios, in which models are trained and evaluated on the same data distributions. Nevertheless, o.o.d. setups that assess model generalization to new test distributions remain challenging even for the most recent models. To advance generalization in AVR tasks, we present the Pathways of Normalized Group Convolution model (PoNG), a novel neural architecture that features group convolution, normalization, and a parallel design. We consider a wide set of AVR benchmarks, including Raven's Progressive Matrices and visual analogy problems with both synthetic and real-world images. The experiments demonstrate strong generalization capabilities of the proposed model, which in several settings outperforms the existing literature methods.