Researchers from Google Quantum AI have made a groundbreaking discovery, demonstrating that today’s noisy intermediate-scale quantum (NISQ) computers can execute benchmark calculations that would take classical computers years to solve. Published in Nature, the study utilized the random circuit sampling benchmark, recognized as one of the most challenging benchmarks for quantum computers.
These findings indicate that quantum computers, even with their current noise levels, hold the potential to outperform classical computers for specific tasks—a pivotal advancement in the journey toward practical quantum computing applications.
The research leveraged Google’s 67-qubit Sycamore quantum chip, revealing the existence of a "stable computationally complex phase" achievable with present-day quantum processors. This means that quantum computers can perform calculations that exceed the capabilities of classical supercomputers, despite the interference from background noise. The experiment is part of a broader investigation into how quantum processors can manage complex computations amidst such noise.
Sergio Boixo, principal scientist at Google Quantum AI, noted, “The question that has been investigated lately by some prominent researchers and publications is whether you can find a phase where noisy quantum computers can fundamentally outperform supercomputers.”
Boixo explained that the experiments demonstrated a transition between two phases in quantum computing. In the “low-noise phase,” the benchmark proved sufficiently complex for the quantum computer to outperform classical counterparts.
Furthermore, the research validated the random circuit sampling benchmark that has been employed in experiments since 2019, proving it is beyond the reach of classical supercomputers. Boixo also highlighted that the findings support the theoretical “Neven's Law,” which posits that quantum computing power is improving at a doubly exponential rate compared to conventional computing.
However, while the random circuit sampling benchmark is significant, it currently lacks real-world applications. Future endeavors will focus on refining this benchmark for practical uses, paving the way for more applicable quantum computing technologies.
