The Cascaded Forward Algorithm for Neural Network Training

TL;DR

The paper introduces the Cascaded Forward (CaFo) algorithm, a new neural network training method that improves efficiency and accuracy by avoiding backpropagation and enabling independent block training.

Contribution

It proposes the CaFo framework, which directly outputs label distributions at each block, eliminating the need for negative samples and allowing parallel training.

Findings

Significant accuracy improvements on four image classification benchmarks.

Efficient training and testing due to no negative sample generation.

Blocks can be trained independently for parallel deployment.

Abstract

Backpropagation algorithm has been widely used as a mainstream learning procedure for neural networks in the past decade, and has played a significant role in the development of deep learning. However, there exist some limitations associated with this algorithm, such as getting stuck in local minima and experiencing vanishing/exploding gradients, which have led to questions about its biological plausibility. To address these limitations, alternative algorithms to backpropagation have been preliminarily explored, with the Forward-Forward (FF) algorithm being one of the most well-known. In this paper we propose a new learning framework for neural networks, namely Cascaded Forward (CaFo) algorithm, which does not rely on BP optimization as that in FF. Unlike FF, our framework directly outputs label distributions at each cascaded block, which does not require generation of additional negative samples and thus leads to a more efficient process at both training and testing. Moreover, in our framework each block can be trained independently, so it can be easily deployed into parallel acceleration systems. The proposed method is evaluated on four public image classification benchmarks, and the experimental results illustrate significant improvement in prediction accuracy in comparison with the baseline.