For several years, quantum computing has been at the heart of the most advanced scientific and technological research. Promising computing capabilities far beyond those of current supercomputers, this revolutionary technology faces many challenges. One of the main obstacles is linked to the errors inherent in qubits, these fundamental units of quantum computers. However, Google recently took a major step forward with Willow, a quantum processor capable of overcoming these limitations. This advance could be a game-changer by bringing quantum computing closer to practical, concrete applications.
What is quantum computing?
Classical computing is based on bits which represent binary states: 0 or 1. Quantum computing, on the other hand, uses qubits (or quantum bits) which can exist in several states simultaneously thanks to two fundamental properties of quantum physics: superposition and entanglement. This capacity allows quantum computers to carry out massively parallel calculations, which opens up new perspectives for solving complex problems.
However, these qubits are also extremely fragile. They are prone to frequent errors because they easily interact with their environment, thereby disrupting their quantum states. This fragility makes their use difficult over long periods or on a large scale, which limits their practical potential.
To address this challenge, researchers have developed quantum error correction, a process that relies on the creation of logical qubits. These are formed by combining several physical qubits, which allows errors to be detected and corrected. However, until recently, no machine had managed to cross a critical threshold where errors decrease exponentially as the machine grows.
Willow: a revolution in quantum error correction
WillowGoogle's new quantum processor, marks a major breakthrough by being the first in the world to cross this threshold, called error correction threshold. This key step, predicted as early as 1995 by researcher Peter Shor, is essential for quantum computing to become practical and scalable.
To achieve this, the processor uses logical qubits where each qubit is encoded by an array of physical qubits. This means that if a physical qubit fails, the data remains protected by the logical qubit. Google has also made significant improvements to the underlying technology, including:
- Improved calibration protocols to reduce initial errors.
- Machine learningused to identify and correct sources of errors.
- Increased consistency timeswhich extends the time qubits remain in a usable state.
Through these advances, Willow has demonstrated a unique ability to reduce errors exponentially when new qubits are added. This opens the way to building much larger and more reliable quantum computers.
Impressive performance
In rigorous testing, Willow demonstrated its power by performing in less than five minutes a calculation that would have taken ten seven billion years to the fastest supercomputer. For comparison, this time greatly exceeds the estimated age of the Universe. These impressive results, obtained using random circuit sampling testing (a standard method used to evaluate the performance of quantum computers), therefore highlight Willow's potential for applications that remain beyond the reach of current technologies.
The processor also displays coherence time of approximately 100 microseconds (length of time qubits can maintain their quantum state without being disturbed by the environment), five times longer than Google's previous quantum chip, Sycamore. This significant improvement made it possible to reduce logical error rates by a factor of 20, an essential step for quantum computers to outperform traditional supercomputers.
Towards a future closer to real applications
Beyond the demonstration of power, Willow opens the way to practical applications. The ultimate goal is to build quantum computers capable of solving real problems more quickly and efficiently than current supercomputers. Potential areas include:
- Molecular simulationused to discover new drugs or optimize materials.
- Network optimizationuseful for transport, logistics or energy systems.
- Basic researchwith quantum simulations that will provide a better understanding of the laws of physics.
To achieve this objective, Google is therefore working on the creation of even more reliable logical qubits, with a error rate less than one in a million. This will require combine approximately 1,457 physical qubits per logical qubita task that will depend on further advances in error correction and engineering.
