IBM unveils quantum computing breakthrough that checks its own results

The team introduced a technique called doped Clifford sampling (DCS), which combines error detection with a built-in way to verify the accuracy of a quantum computation. The researchers say the method addresses two major challenges in quantum computing: keeping errors under control as circuits become larger and confirming that the results are trustworthy.

Quantum computers are expected to solve some problems much faster than today's most powerful classical supercomputers, but proving that they have done so reliably has remained difficult. Existing approaches often require strong assumptions about hardware performance or become impractical as computations grow more complex.

The new method uses specially designed quantum circuits that can detect many errors during computation. The researchers first verify the accuracy of a simpler version of the circuit, then add carefully placed T gates, which increase the computation's complexity while preserving the circuit's error detection features. This allows them to estimate a reliable lower bound for the accuracy of the final, much harder computation without directly simulating it.

To demonstrate the approach, the team ran an experiment on an IBM superconducting quantum processor using a circuit with 70 logical qubits, 97 physical qubits, a circuit depth of 70, and 468 T gates. The experiment reduced effective gate error rates by about 10 times after filtering out runs where errors were detected. The researchers established a 95% confidence lower bound of 0.284 for the fidelity of the final quantum state and collected 2,051 accepted samples in about 16 minutes.

According to the paper, the tested circuit would be difficult to reproduce using current tensor-network and stabilizer-based classical simulation methods, although the authors note that future advances in simulation algorithms or computing hardware could change that assessment. Their primary goal, they say, was to demonstrate that quantum computations believed to be classically hard can also be executed with high confidence in their correctness.

The researchers said the work represents a step toward scalable quantum computing by combining stronger error suppression with practical verification techniques. They added that future advances in hardware could allow similar methods to support fault-tolerant quantum computing, while fully device-independent verification remains an important challenge for the field.

Earlier, Qazinform News Agency reported that a new Microsoft quantum chip boosted qubit reliability 1,000 times.