Quantum Computing in Infinite Dimensions
Online recommendations from Netflix. The biometrics registered by a wearable device. Applications like these are possible through machine learning, a subfield of computer science that Physics Professor George Siopsis and his colleagues have delved into on a quantum level. They’ve developed mathematical processes that take the concept into infinite dimensions, with the potential to speed up computers far beyond conventional machines and make data storage and processing cheaper and more energy efficient. The findings appear in a recent issue of Physical Review Letters.
Machine learning is the idea that computers can perform specific tasks without actually being programmed to do them. They “learn,” as it were, via calculations and mathematical rules that help them adapt when they encounter new data. As Siopsis and his colleagues point out in “Quantum Machine Learning over Infinite Dimensions,” machine learning in the quantum realm has up until now relied on discrete variables, which have a finite number of values. They have developed machine learning algorithms based on continuous variables that work in infinite dimensions. Further, they outline an all-photonic implementation of these algorithms, taking advantage of photonics (currently available technology) to support a quantum computing framework with exponential speedups over classical algorithms.
The research is certainly timely. As Wired reported on March 6, Google plans to produce a quantum computer in the next five years, with IBM announcing plans to offer commercial quantum machines within the year. Siopsis pointed out that quantum machine learning could also require less energy as it makes it possible to store more information per quantum mode, and at a lower cost.
The paper’s authors besides Siopsis include Hoi-Kwan Lau (Institute of Theoretical Physics, Ulm University, Germany), Raphael Pooser (Oak Ridge National Laboratory Quantum Information Science Group and a Joint Faculty Assistant Professor with UT Physics), and Christian Weedbrook (Xanadu Quantum Technologies, Inc., Toronto).
- “Quantum Machine Learning over Infinite Dimensions” in Physical Review Letters
- “Physicists Extend Quantum Machine Learning to Infinite Dimensions” (from phys.org)
- “UT Physicist and Team Study Quantum Computing; IBM Announces First Commercial Service” (from University Herald)
- Background information on machine learning kindly provided with permission by SAS