Rigetti Brings Quantum ML Apps to Strangeworks Platform
By Carolyn Mathas
Strangeworks, Inc. and Rigetti Computing, Inc. just announced that the Strangeworks platform will feature two new quantum machine learning (QML) applications from Rigetti. The quantum kernel and quantum convolutional “quanvolutional” neural network methods, specifically optimized for Rigetti quantum computers, target classification and development of regression applications.
According to whurley (William Hurley), Strangeworks founder and CEO, “We are excited about what Rigetti’s proof of concept could mean for image classification and beyond. This is early evidence of real-world quantum-powered applications resulting in more accurate and efficient work, while improving momentum, confidence and saving time and money across many industries. “
Access to Rigetti systems is made possible by tight integration of Rigetti’s quantum processing units and the platform. This integration delivers the higher performance of Rigetti systems, lowers overall program latency, and provides native Quil programming language support. Rigetti believes that QML applications can advance real-world research, including identifying diseases, conducting climate and weather modeling, identifying manufacturing defects, and combating financial fraud.
Rigetti’s Quanvolutional Neural Network method improves image and video analysis by adding quantum-based features to an existing dataset used by classical neural networks. This facilitates the subsequent machine learning processing, which requires less data and fewer parameters to train the classical model. For example, medical imaging involves the large data sets and complex probability distributions that machine learning performs. Rigetti’s method improves the performance of a typical AI model for diagnosing breast cancer and pneumonia from the MedMNIST dataset collection. Previously, pure classical neural networks needed more than 800,000 parameters to train, quantum enhanced neural networks in comparison require only 200,000, or 75% fewer parameters.
Rigetti’s Quantum Kernel Method assesses similarities between points in a data set, valuable for a classification or regression model. By determining the similarities in the exponentially larger space offered by the quantum processing unit, the output can be used for anomaly detection. This method will be available to all users on the Strangeworks platform, while the convolutional neural network method will be available to select customers and partners.
Rigetti also joins the Strangeworks Backstage Pass program, offering up to $10,000 in sponsored credits to each approved user. Acceptance criteria involves an interview process to determine use cases and expectations, with a priority on enterprise teams interested in leveraging quantum machine learning in their work. To apply for access to Rigetti through the Backstage Pass program, click here.
“The application of quantum capabilities to machine learning is obviously early days,” Whurley said. “What we see from Rigetti is a fantastic indication that quantum can be used to improve the machine learning process. It’s just the beginning. Challenges still include noise on existing quantum devices and the size of problems that can be approached. While we cannot say when these challenges will be overcome, we are very excited to help lead the way with Rigetti,” he added.
The Strangeworks platform is designed to remove the barriers to applying emerging computing technologies, such as quantum, to big problems, making them all as easy to access, integrate and manage as possible. Strangeworks will announce the general availability of its platform at SXSW 2023 as a pay-as-you-go service. Until then, early access is available through the Backstage Pass program.
More information is available in a press release posted on the Strangeworks website here.
December 13, 2022
