A biologically plausible neural network for Slow Feature Analysis

TL;DR

This paper introduces Bio-SFA, a biologically plausible neural network algorithm for Slow Feature Analysis, capable of online learning and local synaptic updates, validated on naturalistic stimuli.

Contribution

The work develops the first biologically plausible, online SFA algorithm that can be mapped onto neural networks with local updates.

Findings

Bio-SFA successfully extracts slow features from natural stimuli.

The algorithm aligns with properties of visual and hippocampal cells.

It demonstrates potential for modeling brain learning mechanisms.

Abstract

Learning latent features from time series data is an important problem in both machine learning and brain function. One approach, called Slow Feature Analysis (SFA), leverages the slowness of many salient features relative to the rapidly varying input signals. Furthermore, when trained on naturalistic stimuli, SFA reproduces interesting properties of cells in the primary visual cortex and hippocampus, suggesting that the brain uses temporal slowness as a computational principle for learning latent features. However, despite the potential relevance of SFA for modeling brain function, there is currently no SFA algorithm with a biologically plausible neural network implementation, by which we mean an algorithm operates in the online setting and can be mapped onto a neural network with local synaptic updates. In this work, starting from an SFA objective, we derive an SFA algorithm, called Bio-SFA, with a biologically plausible neural network implementation. We validate Bio-SFA on naturalistic stimuli.