Inhibitory normalization of error signals improves learning in neural circuits

TL;DR

This paper investigates whether inhibitory normalization of error signals enhances learning in neural circuits, finding that it does so only when normalization affects both neural activity and learning signals, with implications for biological and artificial systems.

Contribution

It demonstrates that inhibition-mediated normalization improves learning only when applied to both neural activity and error signals in artificial neural networks.

Findings

Inhibition-mediated normalization alone does not improve learning.

Normalization of error signals significantly enhances performance.

Results suggest a dual role of inhibition in neural learning processes.

Abstract

Normalization is a critical operation in neural circuits. In the brain, there is evidence that normalization is implemented via inhibitory interneurons and allows neural populations to adjust to changes in the distribution of their inputs. In artificial neural networks (ANNs), normalization is used to improve learning in tasks that involve complex input distributions. However, it is unclear whether inhibition-mediated normalization in biological neural circuits also improves learning. Here, we explore this possibility using ANNs with separate excitatory and inhibitory populations trained on an image recognition task with variable luminosity. We find that inhibition-mediated normalization does not improve learning if normalization is applied only during inference. However, when this normalization is extended to include back-propagated errors, performance improves significantly. These results suggest that if inhibition-mediated normalization improves learning in the brain, it additionally requires the normalization of learning signals.