"QIS Project Shows Novel Method for Privacy-Preserving Quantum ML"
Quantum computing promises to significantly advance computational capabilities. Programs such as the QIS@Perlmutter project at Lawrence Berkeley National Laboratory (Berkeley Lab) are gathering scientists to conduct the foundational research needed to support a quantum future. QIS@Perlmutter, established in 2021 to facilitate Quantum Information Science (QIS) research on the Perlmutter supercomputer at the National Energy Research Scientific Computing Center (NERSC), granted Perlmutter computing resources to 16 research teams in early 2022. Initial scientific results are beginning to emerge from these projects. In a recently published paper titled "Quantum machine learning with differential privacy," a QIS@Perlmutter research group presented findings from a Quantum Machine Learning (QML) project that explores techniques for preserving privacy within advanced quantum computing functions. QML brings the success of Machine Learning (ML) to quantum computers, with the quantum advantages of faster computation, convergence, and greater accuracy with fewer complexities. Differential privacy offers the probabilistic privacy guarantee on a trained ML model. Therefore, an attacker cannot easily reveal personal information from the training data. The objective is to protect training data from the training model. The researchers noted that this study marks the first proof-of-principle demonstration of privacy-preserving QML. This could ensure confidentiality and accurate learning on Noisy Intermediate-Scale Quantum (NISQ) technology. This article continues to discuss the new method for privacy-preserving QML.