Accenture said it has been granted a US patent for a “quantum computing machine learning module” that trains artificial intelligence (AI) models to determine when and how computational tasks would be best handled by quantum computing versus classical computing methods, and route them to the appropriate option.
In doing so, the module could help organisations understand where quantum computing can have the most impact within their businesses, and when classical computing may still be the best option for a given task.
Accenture’s new patent — US Patent No. 10,275,721 — describes a solution to address the tradeoff between the benefits of using quantum computing versus classical computing resources.
By determining when and how to use the power of quantum computing, such a system can help perform computational tasks in the most efficient and cost-effective way possible.
Additionally, the module has the ability to learn to prioritise certain tasks. As more advanced and efficient quantum and classical systems are introduced over time, the quantum computing machine learning module can adapt accordingly.
“Quantum computing has enormous potential, offering truly groundbreaking capabilities to rapidly solve businesses’ most difficult computational challenges,” said Marc Carrel-Billiard, senior managing director at Accenture Labs. “And determining when to employ quantum — as opposed to, or in tandem with classical computing — is critical to realizing this potential.”
The new patent adds to Accenture’s global intellectual property portfolio, including a patent awarded in 2018 for a “multi-state quantum optimization engine.” It also builds on years of Accenture’s quantum investments, partnerships and R&D efforts.
The patent for the quantum computing machine learning module was co-invented by several Accenture associates — Carl Dukatz, principal director in Accenture’s quantum computing practice; Daniel Garrison, managing director at Accenture Digital; and Lascelles Forrester, managing director at Accenture Digital.
