Google researchers have introduced a groundbreaking technique that could make quantum computing practical. They’re using artificial intelligence to tackle one of the biggest challenges in the field: achieving stable quantum states.
In a recent study published in Nature, scientists from Google DeepMind explained their new AI system, AlphaQubit. This system has shown impressive success in correcting the errors that have long plagued quantum computers.
According to Google, “Quantum computers have the potential to revolutionize drug discovery, material design, and fundamental physics—if we can get them to work reliably.” The problem? Quantum systems are incredibly fragile. Just a tiny disturbance—like heat or vibration—can throw them off, leading to errors that make calculations unreliable.
A paper released in March highlights the challenge: for practical use, quantum computers need an error rate of just one in a trillion operations (10^-12). However, current hardware has error rates between 10^-3 and 10^-2 per operation. This makes effective error correction essential.
Google points out that some problems that would take a conventional computer billions of years to solve could be tackled by a quantum computer in just hours. But these new processors are more prone to noise than traditional ones.
To make quantum computers more reliable—especially at scale—it's crucial to identify and correct these errors accurately.
That’s where AlphaQubit comes in. This AI system uses an advanced neural network architecture. It has achieved unprecedented accuracy in detecting and correcting quantum errors, showing 6% fewer errors than previous methods and 30% fewer than traditional techniques.
AlphaQubit maintained high accuracy across quantum systems ranging from 17 to 241 qubits. This suggests it can scale up to the larger systems needed for practical quantum computing.
Under the hood, AlphaQubit employs a two-stage approach. First, it trains on simulated quantum noise data to learn general patterns of quantum errors. Then, it adapts to real quantum hardware using a limited amount of experimental data.
This method allows AlphaQubit to handle complex real-world quantum noise effects, like cross-talk between qubits and leakage, where qubits exit their computational states.
However, it’s important to manage expectations. You won’t be building a quantum computer in your garage just yet. Despite its accuracy, AlphaQubit still faces significant hurdles before practical implementation. For instance, “Each consistency check in a fast superconducting quantum processor is measured a million times every second,” the researchers noted. While AlphaQubit is great at identifying errors, it’s currently too slow to correct them in real-time.
A spokesperson from DeepMind explained, “Training at larger code distances is more challenging because the examples are more complex.” This is crucial because the error rate scales exponentially with code distance. To achieve ultra-low error rates for fault-tolerant computation on large quantum circuits, they need to solve larger distances.
The research team is focusing on speed optimization, scalability, and integration as key areas for future development.
The relationship between AI and quantum computing is mutually beneficial. “We expect AI/ML and quantum computing to remain complementary approaches to computation,” the spokesperson said. AI can support the development of fault-tolerant quantum computers in various ways, like calibration and algorithm design. Meanwhile, researchers are also exploring quantum machine learning applications for quantum data.
This convergence could mark a significant turning point in computational science. As quantum computers become more reliable through AI-assisted error correction, they could also help develop more sophisticated AI systems. This creates a powerful cycle of technological advancement.
The long-awaited age of practical quantum computing may finally be on the horizon, though we’re not quite there yet.