LLS: Local Learning Rule for Deep Neural Networks Inspired by Neural Activity Synchronization

TL;DR

This paper introduces LLS, a local learning rule inspired by neural synchronization, enabling efficient deep neural network training with significantly reduced computational resources and memory, suitable for on-device applications.

Contribution

The paper proposes a novel local learning rule based on neural activity synchronization that matches backpropagation accuracy with fewer resources and no additional trainable parameters.

Findings

LLS achieves comparable accuracy to backpropagation.

LLS reduces MAC operations by up to 300 times.

LLS requires half the memory of traditional training methods.

Abstract

Training deep neural networks (DNNs) using traditional backpropagation (BP) presents challenges in terms of computational complexity and energy consumption, particularly for on-device learning where computational resources are limited. Various alternatives to BP, including random feedback alignment, forward-forward, and local classifiers, have been explored to address these challenges. These methods have their advantages, but they can encounter difficulties when dealing with intricate visual tasks or demand considerable computational resources. In this paper, we propose a novel Local Learning rule inspired by neural activity Synchronization phenomena (LLS) observed in the brain. LLS utilizes fixed periodic basis vectors to synchronize neuron activity within each layer, enabling efficient training without the need for additional trainable parameters. We demonstrate the effectiveness of LLS and its variations, LLS-M and LLS-MxM, on multiple image classification datasets, achieving accuracy comparable to BP with reduced computational complexity and minimal additional parameters. Specifically, LLS achieves comparable performance with up to $300 \times$ fewer multiply-accumulate (MAC) operations and half the memory requirements of BP. Furthermore, the performance of LLS on the Visual Wake Word (VWW) dataset highlights its suitability for on-device learning tasks, making it a promising candidate for edge hardware implementations.