Q-CTRL: error correction technology improves the success rate of the quantum algorithm by 9000 times
Today, Q-CTRL, an Australian quantum company, announced the results of hardware benchmark experiments, which proved that its automatic error correction technology has increased the probability of quantum computing algorithm's success on real hardware by 9000 times, more than 25 times that reported by Q-CTRL in November last year.
With these improvements, Q-CTRL is narrowing the performance gap between the promise of growing quantum computers on the market and the actual computing power they provide to end-users.
Q-CTRL will announce the technology behind the experiment at the March annual meeting of the American Physical Society (APS) (the world's largest Physics Conference) in Chicago from March 14 to 18. Quantum computing researchers and developers can obtain the underlying software tools of hardware stability, error reduction, and automation through the technology of Q-CTRL (including its flagship product Boulder Opal).
The success rate of the quantum algorithm has increased 9000 times
At present, most quantum computers are very error-prone, so they can only run the shortest algorithms; This challenge hinders the provision of useful quantum computing to end-users. Through its research activities, Q-CTRL has identified ways to use artificial intelligence and automation, which can not only significantly improve the errors perplexing the isolation of quantum logic elements (building blocks of quantum applications) but also improve the errors that only appear in the execution of the complete algorithm.
Q-CTRL measures how its software tools improve important industry indicators when executed on actual hardware, including the algorithm benchmark pursued by the American quantum Economic Development Alliance (QED-c). Q-CTRL compares their tools with publicly available state-of-the-art circuit compilation and measurement error mitigation strategies.
The first is the Bernstein Vazirani algorithm, which is a key benchmark established by QED-c in its recent paper on "application-oriented performance benchmark".
As shown in the figure below, compared with other competitor methods, the Q-CTRL method provides the best performance from 5 qubits to 16 qubits, and the larger the scale of a quantum computer, the higher the degree of performance improvement and grows exponentially. When the scale is increased to 16 qubits, Q-CTRL achieves an advantage of 9000 times over the competitor's method.
Bernstein Vazirani algorithm; The horizontal axis is the number of quantum bits and the vertical axis is the success rate of the quantum algorithm.
Then, Q-CTRL studies Grover quantum search algorithm, a needle in a haystack algorithm for unstructured search. In a five-qubit example, Q-CTRL's software changes the result in binary mode, from a "completely wrong" solution to a qualitatively correct solution. Although in this case, the probability of success is only increased by about 20-30 times, the qualitative value provided to end-users is obvious.
When the Grover algorithm is executed, the probability of Q-CTRL getting the correct answer is 20-30 times higher than that of the competitor method.
These results are achieved through Q-CTRL's new zero-configuration toolset fire opal for quantum computing end-users (based on the functions of the flagship product Boulder Opal).
Relevant functions will be launched on Amazon bracket
Q-CTRL said that they have achieved 9000 times improvement in quantum algorithms in the benchmark test, which has brought real utility benefits to customer algorithms. For example, researchers from European quantum start-up Beit, who are early users of Q-CTRL technology.
Paulina Mazurek, CEO of Beit, said: "Beit has been pushing the limits of quantum algorithms, but it faces the same obstacles as everyone else in terms of hardware performance. Fire-Opal has opened up a new frontier in our research, brought quantum advantages closer, and implemented a basic quantum algorithm with better results than classical algorithms. Therefore, to some extent, this software has fundamentally changed The hardware achieves the results that the previous benchmark considered impossible. This left a deep impression on us. "
Zapata computing, another partner of Q-CTRL, has integrated the fire opal prototype into their platform orquestra. At the Q2B conference in December 2021, Q-CTRL showed how this integration can combine Zapata's proprietary algorithms and methods (robust amplitude estimation, RAE) to improve the performance of today's quantum backend. The results show that fire opal seems to reduce the estimation error by about 11 times, while Zapata's Rae improves the estimation accuracy by about 5.5 times.
These results are achieved entirely through traditional cloud access to commercial quantum computers, without special access to hardware. This proves that these functions can be provided to any user who has an Internet connection and wants to get more functions from today's quantum computers.
AWS recently announced that fire opal will soon be launched on Amazon bracket to provide access to various quantum computing backend.
Professor Michael J. biercuk, founder and CEO of Q-CTRL, said: "Our benchmark experiments show that today's quantum computers have hidden performance inside and can be made available through the correct error correction software tools - without any changes to the hardware. We are pleased to provide this technology to researchers, end-users, and manufacturers around the world to accelerate the realization of quantum advantages and bring real-world applications closer to realization. ”
Q-CTRL will hold a seminar at APS. Participants will learn how to use software tools including boulder opal to improve and automate their quantum hardware, save time and increase research results.
About Q-CTRL
The computing power of quantum computing is expected to bring transformative capabilities to drug discovery, enterprise logistics, finance, and other application fields. However, the underlying hardware is extremely unstable and fragile, which prevents these machines from realizing their full potential. Q-CTRL focuses on providing hardware-independent and error correction enterprise software to provide useful quantum computing for enterprises.
