Google Quantum AI's Willow Chip and Quantum Advantage

In an advancement in quantum computing, researchers from Google, MIT, Stanford, and Caltech reported in Nature on October 22 a "verifiable display of quantum advantage" using Google's Willow quantum processor. They demonstrated that Willow outperformed supercomputers in solving specific problems.

Understanding Quantum Interference

• Quantum particles can behave like waves, and their probability wave functions can interfere with each other.
• Constructive interference amplifies probabilities of correct answers, while destructive interference cancels out incorrect ones.

Decoded Quantum Interferometry (DQI) Algorithm

• Utilizes quantum Fourier transform to control wave-like nature of quantum bits.
• For the optimal polynomial intersection problem, DQI provides faster solutions than classical computers.

Measuring Information Scrambling in Quantum Systems

• Information initially concentrated in one quantum bit spreads across all bits in a quantum system.
• Researchers likened this to blue dye spreading in a swimming pool, becoming uniformly distributed and hidden in complex interactions.
• A novel experiment involving sound wave interference was used to measure this scrambling.
• Quantum circuits simulated on supercomputers would take over three years, whereas Willow completed tasks in two hours.

Challenges and Next Steps

• The researchers haven't mathematically proven the inherent difficulty for classical computers to solve the same problems.
• Future research should independently solve unsolved problems using the quantum method.
• Applications of these findings are still prospective, with practical scientific discoveries yet to materialize.
• Improvements in error correction and scaling reliable quantum bits are necessary for broader applications.

Previous Experiments and Implications

• In 2019, Google attempted random circuit sampling with the Sycamore processor.
• The problem solved by Willow was meaningful and verifiable against classical or other quantum computers.
• An early application may be in Hamiltonian learning, comparing experimental data with simulations to infer unknown parameters.

The findings build on principles developed by Nobel laureates in physics, with Michel Devoret, a laureate, being key to Google's quantum hardware developments. These studies indicate a significant step forward but highlight the long road ahead in practical quantum computing applications.