Terra Quantum has published a peer-reviewed paper detailing a new quantum error correction method called QMM-Enhanced Error Correction. The method, which is based on a concept called the Quantum Memory Matrix (QMM), is a hardware-validated, measurement-free approach for suppressing quantum errors. The technology was validated on IBM’s superconducting processors.
The QMM layer is designed to function as a lightweight, unitary “booster” that enhances fidelity without mid-circuit measurements or added two-qubit gates. When combined with a repetition code, logical fidelity is reported to increase to 94%, representing a 32% gain achieved without the addition of CX gates. Simulations suggest that three QMM layers can achieve error rates comparable to those of a distance-3 surface code while requiring ten times fewer qubits. In hybrid workloads such as variational quantum classifiers, QMM is said to reduce training loss by 35% and halve run-to-run performance variance.
This approach is positioned as a paradigm for Noisy Intermediate-Scale Quantum (NISQ) processors where conventional error correction methods may be impractical. It enables deployable error suppression on existing machines without requiring architectural changes and addresses challenges across platforms where mid-circuit measurements are infeasible. The method is intended to enable more performance per qubit and provide a path toward scalable, fault-tolerant quantum computing without requiring a redesign of the stack.
Read the full announcement here and the paper in Wiley Advanced Quantum Technologies here.
August 15, 2025
Leave A Comment