Training a Restricted Boltzmann Machine for Classification by Labeling Model Samples

TL;DR

This paper introduces a novel semi-supervised training method for Restricted Boltzmann Machines on MNIST, where model samples are labeled manually, enabling efficient learning with fewer labeled data and automatic identification of challenging samples.

Contribution

It presents a new approach to training RBMs for classification by labeling model-generated samples, reducing manual labeling effort and improving semi-supervised learning performance.

Findings

Achieved competitive classification accuracy with fewer labeled samples.

Utilized video-like presentation of samples to human labelers for efficiency.

Automatically identified hard-to-classify examples to optimize manual labeling.

Abstract

We propose an alternative method for training a classification model. Using the MNIST set of handwritten digits and Restricted Boltzmann Machines, it is possible to reach a classification performance competitive to semi-supervised learning if we first train a model in an unsupervised fashion on unlabeled data only, and then manually add labels to model samples instead of training data samples with the help of a GUI. This approach can benefit from the fact that model samples can be presented to the human labeler in a video-like fashion, resulting in a higher number of labeled examples. Also, after some initial training, hard-to-classify examples can be distinguished from easy ones automatically, saving manual work.