Accelerating spiking neural networks using quantum algorithm with high success probability and high calculation accuracy
AI Breakdown
Get a structured breakdown of this paper — what it's about, the core idea, and key takeaways for the field.
Abstract
Utilizing quantum computers to deploy artificial neural networks (ANNs) will bring the potential of significant advancements in both speed and scale. In this paper, we propose a kind of quantum spike neural networks (SNNs) as well as comprehensively evaluate and give a detailed mathematical proof for the quantum SNNs, including its successful probability, calculation accuracy, and algorithm complexity. The proof shows the quantum SNNs' computational complexity that is log-polynomial in the data dimension. Furthermore, we provide a method to improve quantum SNNs' minimum successful probability to nearly 100%. Finally, we present the good performance of quantum SNNs for solving pattern recognition from the real-world.