Handwritten Digit Recognition with a Committee of Deep Neural Nets on GPUs

TL;DR

This paper demonstrates that using a committee of deep neural networks trained on GPUs can significantly improve handwritten digit recognition accuracy on the MNIST benchmark, achieving a new record error rate of 0.31%.

Contribution

It introduces a novel ensemble approach with deep MLPs trained on GPUs, surpassing previous complex methods in accuracy.

Findings

Achieved 0.31% error rate on MNIST.

Ensemble of deep MLPs outperforms previous methods.

GPU acceleration enables efficient training of deep models.

Abstract

The competitive MNIST handwritten digit recognition benchmark has a long history of broken records since 1998. The most recent substantial improvement by others dates back 7 years (error rate 0.4%) . Recently we were able to significantly improve this result, using graphics cards to greatly speed up training of simple but deep MLPs, which achieved 0.35%, outperforming all the previous more complex methods. Here we report another substantial improvement: 0.31% obtained using a committee of MLPs.