Researchers from Google, MIT, Stanford, and Caltech have made a significant advancement in quantum computing, as reported in two papers published in Nature on October 22. The teams claim to have verified a clear instance of quantum advantage through their Willow quantum processor, demonstrating its ability to solve a specific problem more efficiently than existing supercomputers.
At the core of quantum computing is the principle of interference, akin to waves in a pond. When wave crests meet, they amplify each other, while crests and troughs can cancel each other out. This phenomenon is utilized by quantum computers, which manipulate the wave functions of particles, enhancing the probability of arriving at the correct solution while diminishing the wrong ones.
In one of the studies, the team introduced a novel quantum algorithm named Decoded Quantum Interferometry (DQI), aimed at optimization problems—tasks that require identifying the best solution among numerous possibilities. The DQI algorithm employs a quantum variant of the Fourier transform to harness the wave-like properties of quantum bits, orchestrating the interference of these waves to favor optimal outcomes. The researchers found that for the optimal polynomial intersection problem, DQI could provide high-quality approximations significantly faster than any classical computer.
The second study focused on the concept of scrambling, where information becomes dispersed in a complex quantum system. An analogy is drawn with dropping a small amount of dye into water; initially, the dye is localized, but over time it spreads throughout the pool, blending into the water and making it impossible to trace back to its original form. Similarly, in quantum systems, information stored in one quantum bit can become distributed across others during interactions, making retrieval challenging.
To address this, the researchers conducted an experiment akin to shouting in an echoing warehouse. When sound waves spread and bounce off surfaces, they become scrambled. If a bell is struck while the echoes still resonate, it alters the sound waves, leaving a faint imprint. By analyzing these altered echoes after reversing their paths, scientists can measure how much information has been scrambled within the system.
In their experiments, the researchers tackled highly intricate circuits, estimating that simulating these on the second fastest supercomputer would take over three years. In contrast, the Willow processor accomplished the same task in approximately two hours. However, while one paper presented a quantum algorithm that significantly outperformed classical computers, it has not been mathematically established that no classical computer could solve the same problem efficiently.
Future work is essential to confirm that certain problems remain inherently difficult for non-quantum computers, as well as to apply these methods to unresolved challenges in fields such as physics or chemistry. Although these studies represent important strides in quantum computing, their practical applications are still largely theoretical and depend on advancements in areas like error correction and scaling quantum bits.
In a previous experiment from 2019, researchers at Google attempted to address a problem known as random circuit sampling using their Sycamore processor. This involved predicting outcomes from a random program, but the results were harder to verify. In contrast, the recent studies tackled scientifically meaningful problems, yielding verifiable results. One promising application of their findings may be in Hamiltonian learning, which involves deducing unknown parameters of physical systems by comparing experimental data with simulated results.
The new research builds on the foundational work of recent Nobel laureates, including Michel Devoret, who is the chief scientist of quantum hardware at Google Quantum AI. Their contributions have been instrumental in advancing processors like Willow and addressing complex optimization challenges.
