Quantum computational advantage with a programmable photonic processor

Area of Science:

• Quantum Computing
• Quantum Optics
• Computational Physics

Background:

• Quantum computers offer computational advantage over classical systems for specific tasks.
• Previous photonic quantum processors lacked full gate programmability and were susceptible to spoofing.
• Demonstrating quantum advantage requires outperforming classical algorithms on well-defined problems.

Purpose of the Study:

• To demonstrate quantum computational advantage using a dynamically programmable photonic processor.
• To validate the capabilities of the Borealis quantum processor for Gaussian Boson Sampling.
• To establish photonics as a viable platform for practical quantum computation.

Main Methods:

• Utilized Borealis, a photonic processor with dynamic programmability across all gates.
• Executed Gaussian Boson Sampling (GBS) on 216 entangled squeezed modes with 3D connectivity.
• Employed a time-multiplexed, photon-number-resolving architecture.

Main Results:

• Achieved quantum computational advantage, with Borealis outperforming classical simulations by over 50 million times in runtime.
• Generated GBS experiments with up to 219 photons and a mean photon number of 125.
• Demonstrated a runtime of 36 microseconds for Borealis, compared to an estimated 9,000+ years for classical supercomputers.

Conclusions:

• Borealis establishes a new benchmark for photonic quantum computing, showcasing dynamic programmability and computational advantage.
• The results validate key technological aspects of photonics for building practical quantum computers.
• This work represents a significant step towards realizing the potential of quantum computation.