Learning long-range spatial dependencies with horizontal gated-recurrent units

TL;DR

This paper introduces the hGRU, a recurrent neural network layer that learns long-range spatial dependencies, outperforming complex feedforward models in visual recognition tasks involving co-dependent features.

Contribution

The paper presents the hGRU, a novel recurrent layer that captures long-range spatial dependencies, showing superior performance over larger feedforward networks in recognition tasks.

Findings

hGRU matches or outperforms larger feedforward models

Single hGRU layer effectively learns long-range dependencies

Biological plausibility aligns with visual cortex data

Abstract

Progress in deep learning has spawned great successes in many engineering applications. As a prime example, convolutional neural networks, a type of feedforward neural networks, are now approaching -- and sometimes even surpassing -- human accuracy on a variety of visual recognition tasks. Here, however, we show that these neural networks and their recent extensions struggle in recognition tasks where co-dependent visual features must be detected over long spatial ranges. We introduce the horizontal gated-recurrent unit (hGRU) to learn intrinsic horizontal connections -- both within and across feature columns. We demonstrate that a single hGRU layer matches or outperforms all tested feedforward hierarchical baselines including state-of-the-art architectures which have orders of magnitude more free parameters. We further discuss the biological plausibility of the hGRU in comparison to anatomical data from the visual cortex as well as human behavioral data on a classic contour detection task.