An exciting new error correction technique could dramatically shrink the size of quantum computers, making them over 200 times more compact.
Quantum computers can perform complex calculations much faster than classical computers. But their speed comes with a catch – more potential for errors. To generate just a single reliable “logical” qubit immune to errors, current systems require around 1,000 physical qubits. This translates to enormous infrastructure demands. Running advanced algorithms would need millions of physical qubits – impractical for real-world use.
Researchers have been seeking ways to slash the number of physical qubits needed per logical qubit. One approach uses “phase-flip” error correction, lowering requirements 60-fold for some algorithms. Another leverages “low-density parity check” (LDPC) codes to efficiently detect and fix errors.
Now, scientists from Alice & Bob and Inria have combined these two techniques in a novel way. Their method uses local qubit interactions and parallel logic gates, avoiding long-range qubit connectivity and complex error-checking operations previously needed.
This breakthrough could enable reliable logical qubits using today’s quantum hardware with as few as 1,500 physical qubits – a 200-fold improvement over prior approaches needing 20 million.
“Our approach makes quantum computers more realistic in time, cost and energy,” said Alice & Bob CEO Théau Peronnin. “With less than 100,000 physical qubits, we could run advanced algorithms out of reach before.”
The theoretical work advances LDPC codes through new gate designs tailored for existing quantum chips. By dramatically boosting error correction efficiency, it brings industry-relevant quantum computations closer to reality.
“By improving correction tenfold, Alice & Bob’s innovations could deliver practical logical qubits with current hardware,” added Jean-François Bobier of the Boston Consulting Group.
