Hebbian Semi-Supervised Learning in a Sample Efficiency Setting

TL;DR

This paper introduces a semi-supervised training method for deep CNNs that combines Hebbian learning for internal layers with gradient descent for the final layer, improving sample efficiency especially with limited labeled data.

Contribution

It presents a novel semi-supervised approach that pre-trains internal layers with Hebbian learning, reducing the need for labeled data in training deep neural networks.

Findings

Outperforms end-to-end supervised training in low-label regimes

Better results than VAE-based semi-supervised learning with limited labels

Effective in various object recognition datasets

Abstract

We propose to address the issue of sample efficiency, in Deep Convolutional Neural Networks (DCNN), with a semi-supervised training strategy that combines Hebbian learning with gradient descent: all internal layers (both convolutional and fully connected) are pre-trained using an unsupervised approach based on Hebbian learning, and the last fully connected layer (the classification layer) is trained using Stochastic Gradient Descent (SGD). In fact, as Hebbian learning is an unsupervised learning method, its potential lies in the possibility of training the internal layers of a DCNN without labels. Only the final fully connected layer has to be trained with labeled examples. We performed experiments on various object recognition datasets, in different regimes of sample efficiency, comparing our semi-supervised (Hebbian for internal layers + SGD for the final fully connected layer) approach with end-to-end supervised backprop training, and with semi-supervised learning based on Variational Auto-Encoder (VAE). The results show that, in regimes where the number of available labeled samples is low, our semi-supervised approach outperforms the other approaches in almost all the cases.