Cloud-Based Quantum Computing Gains Traction Across SectorsQuantum Computing Resources With Reasonable Pricing Will Accelerate Adoption
The traveling salesman problem stands as a challenging algorithmic problem with extensive implications. The problem revolves around determining the shortest route between a set of points, presenting complexity as the number of destinations increases. The practical application of the TSP extends to logistics-related issues for businesses to optimize routes.
Although machine learning is employed to address TSP, it falls short of delivering optimal results. On the other hand, while quantum computing can deliver efficiency, there’s a caveat: Building and maintaining scalable and stable quantum hardware is prohibitively expensive, deterring the development of real-world applications.
Currently, the application scope for real-world business problems remains narrow. However, major cloud providers such as AWS, Google, IBM and Microsoft are working to change this equation by delivering quantum computing as a service over the cloud.
Quantum Computing Adoption: Cloud as a Cost-Effective Catalyst
Due to its deep computation characteristics, cloud is considered ideal for quantum computing. The integration of these two technologies could help solve intricate issues in domains such as materials science simulations, cryptography and optimization. Quantum computing spans a wide range of applications, from improving climate modeling to enhancing drug discovery and optimizing supply chains.
Quantum computing as a service is gaining popularity, particularly with cloud services easing access to quantum hardware and simulators. According to Kanishka Agiwal, head of service lines, India and South Asia, AWS, accessibility is a big challenge. "Getting a physical quantum computer or even a timeslot on one to drive experiments or projects is extremely difficult as very few of them are available," he said.
With access over cloud, the development and implementation of business use cases are expected to move faster. "Cost and scale are the two key factors driving adoption of cloud-based quantum computing," Agiwal said.
Enterprises Embark on Quantum-Computing-as-a-Service
In 2019, AWS launched Amazon Braket, a fully managed quantum computing service that allows customers to access multiple quantum computing hardware on a pay-as-you-go basis using an AWS console. It also provides a fully managed simulator that enables the users to test their algorithms and experiments on a simulator before deploying them on a quantum computer.
Early adopters, such as Boeing, collaborated with AWS for Amazon Braket to explore quantum algorithms' potential in advanced material research. "Boeing is investigating how disruptive quantum computing, sensing and networking technologies can enhance products and services for our customers," said Charles Toups, VP and GM, disruptive computing and networks, Boeing. Boeing foresees quantum computing’s potential to address aerospace industry’s challenges around fundamental materials science research, complex system optimization and secure communications.
In 2022, automotive company Volkswagen also deployed Amazon Braket to evaluate the performance of a quantum algorithm for the binary paint shop problem, BPSP, with applications extending to scheduling optimization and fraud detection. Rowen Wu, staff product manager, Q-CTRL, describes BPSP as a standard optimization problem that is described in the context of an automotive paint shop but can also be applied to various other industries such as printing and textile manufacturing.
De Beers, employing Amazon Braket for artificial diamonds, showcases the diverse applications of quantum computing. AWS has also introduced a production public use case for quantum key distribution, applicable to specific projects.
According to Agiwal, various experiments are underway, utilizing quantum computing over the cloud to address practical business applications, including TSP.
Another use case is modelling ammonia as a molecule in the fertilizer industry. Given its widespread use, producers seek to know how ammonia disintegrates over a period. Attempting to model this behavior in a traditional computing environment will be time-consuming and resource-intensive. Quantum computers can do it in a few seconds.
Quantum Computing in Enterprise Applications
QpiAI, a company building AI and quantum-based solutions, has put into production the portfolio optimization use case for the financial services industry and running proofs of concept with several banks. "The problem statement has been addressed and now we are ready to integrate the solution into the workflow," said Dr. Nagendra Nagaraja, founder, CEO and chairman of QpiAI. Dr. Nagaraja lists risk management and risk reporting as the other key application besides portfolio optimization.
The solution was built using quantum computing on cloud as part of Quantum Computing Applications Lab known as QCAL (a collaboration between AWS and Ministry of Electronics and Information Technology, Government of India). Under the initiative, researchers, academia, startups and developers submit their proposals for quantum computing. The proposals that are accepted receive credits from AWS, access to Amazon Braket, and a support system of partners and subject matter experts.
The company is also experimenting with quantum computing in the cloud for credit card fraud detection. As the number of features increases from 100 to 5,000, the training time only rises modestly, going from 20 seconds to 40 seconds. In contrast, when compared to a classical computer, the corresponding increase in training time is nearly 50 times.
National Institute of Technology (NIT), Andhra Pradesh, a leading academic institution in India and part of QCAL, is attempting to solve diagnostic problems in healthcare. It is building quantum probabilistic models over cloud to help the diagnosis of difficult diseases. Traditional diagnosis faces challenges as symptoms can't be isolated, leading to significant disagreement among experts. Dr. Karthick Seshadri, assistant professor at NIT, underscored the need to consider symptom interactions, a task made possible by quantum probabilistic models. These models offer tractable inferences, providing likely prognosis, diagnosis and recommended actions.
Jayasri Dontabhaktuni, associate professor of physics at Mahindra University, also a part of the QCAL program, is developing quantum ML-based models via cloud for use in inverse material design and diagnostics.
Cloud-Based Quantum Computing: A Glimpse Into the Future
In terms of enterprise applications, quantum computing via the cloud could support various computationally intensive tasks, such as last-mile optimization, encryption and other complex problems. The technology holds promise across multiple domains including logistics, cybersecurity, predictive equipment maintenance and weather prediction. However, Dr. Nagaraja cautions that everything may not be solved through quantum computing. "Some hard-to-solve problems, which can give exponential results in terms of impact, are suitable for quantum computer. We need to go after those," he said.
Cloud-based quantum computing is currently in a nascent stage and the service model is anticipated to complement existing compute and AI services offered by cloud providers. Integrating quantum computing with cloud platforms points toward a future where these technologies could revolutionize various industries by solving previously intractable problems.