Naples Team Cuts CNOT Gates in Clifford Circuits
A research group in Naples has developed AlphaClifford, a reinforcement learning framework for Clifford circuit synthesis that produces circuits with fewer total gates and CNOT gates than established creation methods. The framework also beats comparable systems when tuning circuits for specific quantum computer architectures. It represents circuit relationships through algebraic properties.
OutlookPlausible
AlphaClifford's trained policies could be integrated into existing quantum compilers such as Qiskit or TKET to automatically reduce CNOT counts in Clifford subroutines used for error correction and randomized benchmarking on near-term devices.