Characterization and Supression of Environmental Noise in Quantum Systems Using Physics-Informed Neural Networks

Abstract

This research project aims to develop advanced numerical methods and machine-learning techniques to characterize andsuppress environmental noise in quantum systems. Building on previous research funded by FAPESP, the project will utilize data-drivenapproaches to identify and parametrize noise in quantum logic gates. The project will focus on programming control pulses using Qiskiton IBM's quantum computing platforms, followed by collecting performance data and applying physics-informed learning to analyzenoise patterns. These techniques will integrate physical laws and domain knowledge to classify noise types and identify correlations,optimizing control pulses to supress noise effects. The ultimate goal is to enhance the fidelity of quantum operations, contributing tothe advancement of quantum computing beyond the NISQ (Noisy Intermediate-Scale Quantum) era. The research will involve regular progress reviews, documentation, and dissemination of findings through publications and reports.