Adaptive and Iteratively Improving Recurrent Lateral Connections

TL;DR

This paper introduces a dynamic recurrent lateral connection model inspired by primate visual cortex, which iteratively refines features and improves performance in visual recognition tasks without pretraining.

Contribution

It proposes a novel architecture with adaptive, iteratively improving lateral connections that enhance visual recognition performance over traditional feedforward models.

Findings

Outperforms existing architectures on ImageNet without pretraining

Leads to significant performance gains in visual action recognition

Demonstrates the effectiveness of dynamic lateral connections in neural models

Abstract

The current leading computer vision models are typically feed forward neural models, in which the output of one computational block is passed to the next one sequentially. This is in sharp contrast to the organization of the primate visual cortex, in which feedback and lateral connections are abundant. In this work, we propose a computational model for the role of lateral connections in a given block, in which the weights of the block vary dynamically as a function of its activations, and the input from the upstream blocks is iteratively reintroduced. We demonstrate how this novel architectural modification can lead to sizable gains in performance, when applied to visual action recognition without pretraining and that it outperforms the literature architectures with recurrent feedback processing on ImageNet.