by Allcock, Jonathan, Hsieh, Chang-Yu, Kerenidis, Iordanis and Zhang, Shengyu
Abstract:
Quantum machine learning has the potential for broad industrial applications, and the development of quantum algorithms for improving the performance of neural networks is of particular interest given the central role they play in machine learning today. We present quantum algorithms for training and evaluating feedforward neural networks based on the canonical classical feedforward and backpropagation algorithms. Our algorithms rely on an efficient quantum subroutine for approximating inner products between vectors in a robust way, and on implicitly storing intermediate values in quantum random access memory for fast retrieval at later stages. The running times of our algorithms can be quadratically faster in the size of the network than their standard classical counterparts since they depend linearly on the number of neurons in the network, and not on the number of connections between neurons. Furthermore, networks trained by our quantum algorithm may have an intrinsic resilience to overfitting, as the algorithm naturally mimics the effects of classical techniques used to regularize networks. Our algorithms can also be used as the basis for new quantum-inspired classical algorithms with the same dependence on the network dimensions as their quantum counterparts but with quadratic overhead in other parameters that makes them relatively impractical.
Reference:
Quantum Algorithms for Feedforward Neural Networks (Allcock, Jonathan, Hsieh, Chang-Yu, Kerenidis, Iordanis and Zhang, Shengyu), In ACM Transactions on Quantum Computing, Association for Computing Machinery, volume 1, 2020.
Bibtex Entry:
@article{10.1145/3411466,
abstract = {Quantum machine learning has the potential for broad industrial applications, and the development of quantum algorithms for improving the performance of neural networks is of particular interest given the central role they play in machine learning today. We present quantum algorithms for training and evaluating feedforward neural networks based on the canonical classical feedforward and backpropagation algorithms. Our algorithms rely on an efficient quantum subroutine for approximating inner products between vectors in a robust way, and on implicitly storing intermediate values in quantum random access memory for fast retrieval at later stages. The running times of our algorithms can be quadratically faster in the size of the network than their standard classical counterparts since they depend linearly on the number of neurons in the network, and not on the number of connections between neurons. Furthermore, networks trained by our quantum algorithm may have an intrinsic resilience to overfitting, as the algorithm naturally mimics the effects of classical techniques used to regularize networks. Our algorithms can also be used as the basis for new quantum-inspired classical algorithms with the same dependence on the network dimensions as their quantum counterparts but with quadratic overhead in other parameters that makes them relatively impractical.},
address = {New York, NY, USA},
articleno = {6},
author = {Allcock, Jonathan and Hsieh, Chang-Yu and Kerenidis, Iordanis and Zhang, Shengyu},
date-added = {2021-03-09 21:05:05 +0100},
date-modified = {2021-03-09 21:05:05 +0100},
doi = {10.1145/3411466},
issn = {2643-6809},
issue_date = {December 2020},
journal = {ACM Transactions on Quantum Computing},
keywords = {Quantum algorithms, qRAM},
month = oct,
number = {1},
numpages = {24},
publisher = {Association for Computing Machinery},
title = {Quantum Algorithms for Feedforward Neural Networks},
url = {https://doi.org/10.1145/3411466},
volume = {1},
year = {2020},
bdsk-url-1 = {https://doi.org/10.1145/3411466}}