Optimizing continuous-variable functions with quantum annealing
Quantum annealing (QA) is a cutting-edge algorithm that leverages the unique properties of quantum computing to tackle complex combinatorial optimization problems (a class of mathematical problems dealing with discrete-variable functions). Quantum computers use the rules of quantum physics to solve such problems potentially faster than classical computers. In essence, they can explore multiple solutions to a problem simultaneously, giving them a significant speed advantage for certain tasks over classical computers. In particular, QA harnesses the phenomenon of “quantum tunneling,” where particles can “tunnel” through energy barriers without the requisite energy to cross over them, to find solutions for combinatorial optimization problems.
Up until now, QA has almost exclusively been used to solve discrete-variable functions (functions that have discrete-valued variables). The potential of QA for optimizing continuous-variable functions has remained largely unexplored.
Against this backdrop, a team of researchers from Japan, Shunta Arai and Hidetoshi Nishimori from Tokyo Institute of Technology (Tokyo Tech) and Hiroki Oshiyama of Tohoku University, has recently tested the continuous-variable optimization performance of QA on the D-Wave 2000Q quantum computer and compared the results with those by classical algorithms. Their findings were published in the journal Physical Review A.
“We systematically investigated whether QA has an advantage over classical algorithms by optimizing the Rastrigin function, a one-dimensional continuous function used as a standard for benchmarking optimization algorithms,” explains Prof. Nishimori. The team used a technique called “domain-wall encoding” to map a continuous variable to discrete Ising variables and performed two sets of benchmark tests.
The team first compared the performance of QA on the D-Wave 2000Q to that of several state-of-the-art optimization algorithms designed for continuous-variable functions, such as Nelder-Mead, conjugate gradient descent, basin-hopping, and differential evolution, all of them running on classical computers. They found that, for higher energy barriers, D-Wave performed as well or even better than the classical algorithms, albeit only for a limited time range. For longer execution times, the classical algorithms performed better, while D-Wave plateaued off.
In the second part of their study, the team compared the performance of D-Wave with classical discrete-variable optimization algorithms: simulated annealing (SA), simulated QA (SQA), and spin-vector Monte Carlo (SVM). They also included the time-evolving block decimation (TEBD) algorithm, which simulates noise-free coherent QA on a classical computer. Notably, the performance of all algorithms except TEBD, which outperformed the rest, was found to be dependent on the energy barrier height, a dependence natural for SA, SQA and SVM.
Crucially, the dependence of D-Wave QA on barrier height implies that its performance is affected by thermal noise in the hardware and can, therefore, be significantly improved by minimizing the noise and other hardware imperfections.
Prof. Nishimori highlights, “Classical algorithms would struggle to find a solution if the energy barrier becomes even higher than what we tested, while QA, if realized coherently, would be far less affected.” These results suggest that with optimized hardware, QA can significantly outperform state-of-the-art classical algorithms, even for optimizing continuous functions.
Overall, this study thus represents a significant step toward systematic and quantitative studies of continuous-variable optimization by QA compared to a range of well-established classical algorithms.
[Attachments] See images for this press release:
ELSE PRESS RELEASES FROM THIS DATE:
LOS ANGELES — City of Hope, one of the largest cancer research and treatment organizations in the United States, is opening a multicenter clinical trial to evaluate the efficacy of robotic-assisted, single-incision mastectomies. The minimally invasive procedure, which preserves the nipple and leaves only a small hidden scar on the side of the body, could potentially lead to significant improvements for breast surgery.
“City of Hope is once again taking the lead in investigating innovations, treatments and therapies that are making big leaps forward for patients with cancer. We’re participating ...
Oak Brook, IL – A Special Issue of SLAS Discovery, Emerging Drug Discovery Ecosystems, is new for September. Volume 28, Issue 6 features three perspectives, one original research article and one protocol that align with the Virginia Drug Discovery Consortium (VaDDC) and its efforts to enhance and promote drug discovery and development in the Commonwealth of Virginia.
An acute respiratory distress syndrome drug development collaboration stimulated by the Virginia Drug Discovery Consortium
Communities with more socioeconomic vulnerability experience longer-duration power outages than more advantaged communities, according to a study. Research has shown that environmental disasters hit economically and socially vulnerable communities hardest. Scott Ganz and colleagues assessed the unequal impacts caused by the procedures electric utilities follow to restore power to customers after extreme-weather related outages. Using data from eight Atlantic hurricanes that made landfall between January 2017 and October 2020, which knocked out power for a total of over 15 million customers in 588 counties in the Southeast, the authors find ...
The Genomics Platform at the Broad Institute of MIT and Harvard has been one of the world’s leading academic genome sequencing centers since the days of the Human Genome Project. For the last decade, these services have included clinical sequencing and other molecular assays through its wholly-owned subsidiary, the Clinical Research Sequencing Platform.
Now renamed Broad Clinical Labs (BCL), the lab is poised to further accelerate the power of ’omics technologies in clinical research, screening, and diagnostics.
BCL supports large-scale projects for which results need to be generated under a clinical quality system, such as analyses for clinical trials, biobank profiling, ...
UMass Amherst researchers have pushed forward the boundaries of biomedical engineering one hundredfold with a new method for DNA detection with unprecedented sensitivity.
“DNA detection is in the center of bioengineering,” says Jinglei Ping, lead author of the paper that appeared in Proceedings of the National Academy of Sciences. Ping is an assistant professor of mechanical and industrial engineering, an adjunct assistant professor in biomedical engineering and affiliated with the Center for Personalized Health Monitoring of the Institute for Applied Life Sciences. “Everyone ...
An individual diagnosed with type 2 diabetes at age 30 years could see their life expectancy fall by as much as 14 years, an international team of researchers has warned.
Even people who do not develop the condition until later in life – with a diagnosis at age 50 years – could see their life expectancy fall by up to six years, an analysis of data from 19 high-income countries found.
The researchers say the findings, published in The Lancet Diabetes & Endocrinology, highlight the urgent need to ...
Watching a loved one suffer the slow decline of dementia is hard enough.
But trying to find care for them that’s high-quality, safe, available and affordable is nearly as difficult, as millions of American families have found out the hard way in recent years.
That lack of options stems from a combination of workforce shortages, payment policies and a lack of consistency in dementia care delivery and regulation, according to experts from the University of Michigan and the University of California, San Francisco and their colleagues.
Now, fueled by new five-year funding expected to total $81 million from the NIH’s National Institute on Aging or ...
New research by the psychologists Lucía Vicente and Helena Matute from Deusto University in Bilbao, Spain, provides evidence that people can inherit artificial intelligence biases (systematic errors in AI outputs) in their decisions.
The astonishing results achieved by artificial intelligence systems, which can, for example, hold a conversation as a human does, have given this technology an image of high reliability. More and more professional fields are implementing AI-based tools to support the decision-making of specialists to minimise errors in their decisions. However, this technology is not without risks due to biases in AI results. ...
The decision to reach for a sugary beverage is heavily influenced by where you live, Friedman School of Nutrition Science and Policy researchers report in a new study published October 3 in the journal Nature Communications. While an analysis of the Global Dietary Database for the years 1990, 2005, and 2018 found overall consumption of sweetened drinks increased—by nearly 16% worldwide over the 28-year period studied—regional intake widely varied.
Sugary drinks are a public health concern because they have been widely associated ...
Australian researchers have developed a molecular-sized, more efficient version of a widely used electronic sensor, in a breakthrough that could bring widespread benefits.
Piezoresistors are commonly used to detect vibrations in electronics and automobiles, such as in smart phones for counting steps, and for airbag deployment in cars. They are also used in medical devices such as implantable pressure sensors, as well as in aviation and space travel.
In a nationwide initiative, researchers led by Dr Nadim Darwish from Curtin University, Professor Jeffrey ...
LAST 30 PRESS RELEASES:
[Press-News.org] Optimizing continuous-variable functions with quantum annealing