References & Citations
Quantum Physics
Title: Quantum Deformed Neural Networks
(Submitted on 21 Oct 2020 (v1), last revised 25 Nov 2020 (this version, v2))
Abstract: We develop a new quantum neural network layer designed to run efficiently on a quantum computer but that can be simulated on a classical computer when restricted in the way it entangles input states. We first ask how a classical neural network architecture, both fully connected or convolutional, can be executed on a quantum computer using quantum phase estimation. We then deform the classical layer into a quantum design which entangles activations and weights into quantum superpositions. While the full model would need the exponential speedups delivered by a quantum computer, a restricted class of designs represent interesting new classical network layers that still use quantum features. We show that these quantum deformed neural networks can be trained and executed on normal data such as images, and even classically deliver modest improvements over standard architectures.
Submission history
From: Roberto Bondesan [view email][v1] Wed, 21 Oct 2020 09:46:12 GMT (59kb)
[v2] Wed, 25 Nov 2020 16:36:41 GMT (63kb)
Link back to: arXiv, form interface, contact.