Application of Quantum Annealing to Training of Deep Neural Networks

TL;DR

This paper explores using quantum annealing to train Deep Belief Networks, demonstrating comparable or improved accuracy with fewer training iterations on a simplified MNIST dataset, suggesting potential for quantum-enhanced deep learning.

Contribution

It introduces a novel quantum annealing-based method for estimating model expectations in Restricted Boltzmann Machines, reducing training time compared to traditional methods.

Findings

Quantum sampling achieves similar or better accuracy than Contrastive Divergence.

Fewer training iterations are needed with quantum sampling.

Further research is required to generalize results to other datasets.

Abstract

In Deep Learning, a well-known approach for training a Deep Neural Network starts by training a generative Deep Belief Network model, typically using Contrastive Divergence (CD), then fine-tuning the weights using backpropagation or other discriminative techniques. However, the generative training can be time-consuming due to the slow mixing of Gibbs sampling. We investigated an alternative approach that estimates model expectations of Restricted Boltzmann Machines using samples from a D-Wave quantum annealing machine. We tested this method on a coarse-grained version of the MNIST data set. In our tests we found that the quantum sampling-based training approach achieves comparable or better accuracy with significantly fewer iterations of generative training than conventional CD-based training. Further investigation is needed to determine whether similar improvements can be achieved for other data sets, and to what extent these improvements can be attributed to quantum effects.