Simplifying quantum characterization through physical symmetries, machine learning, and a top-down approach

Abstract

Fault-tolerant quantum computation promises that complicated quantum algorithms can be performed using faulty components. Predicting the success probability of a fault-tolerant quantum computation is now understood when the component errors do not depend on time and space and are independently selected random Pauli operators of low weight. Our proposed novel theoretic methods focus on two tasks: characterizing and effectively canceling the time correlation in the noise and developing efficient protocols for detecting high-weight errors. We approach the problem from both the bottom-up and the top-down. The bottom-up approach uses the symmetry and physics of the quantum system to simplify characterization. It is a response to current characterization techniques that assume no knowledge about the error models. The top-down approach assumes the quantum system should be capable of implementing the fault-tolerant protocol. By running a set of verification and validation experiments, we can localize the faulty parts of the system that need extra characterization and calibration without testing all elements in detail. For both approaches, we will use machine learning to simplify models and track and predict how the errors are changing in time.

Document Details

Document Type
DoD Grant Award
Publication Date
Oct 22, 2020
Source ID
W911NF2110005

Entities

People

  • Kenneth R. Brown

Organizations

  • Army Contracting Command
  • Duke University
  • National Security Agency

Tags

Readers

  • Neural Network Machine Learning.
  • Parallel and Distributed Computing.
  • Theoretical Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • Quantum Computing
  • Space