Effective Compression of Quantum Braided Circuits Aided by ZX-Calculus

TL;DR

This paper introduces a ZX-calculus based method for compressing quantum braided circuits, significantly reducing their size and overhead, thus enhancing the feasibility of surface-code quantum computation.

Contribution

It presents a novel approach using ZX-calculus for scalable compression of braided quantum circuits, outperforming previous diagram-based methods.

Findings

Achieved up to 77% compression on Clifford+T circuits.

Demonstrated the method's advantage over traditional diagram-based approaches.

Showed that braided circuits have comparable overheads to lattice-surgery circuits.

Abstract

Mapping a quantum algorithm to any practical large-scale quantum computer will require a sequence of compilations and optimizations. At the level of fault-tolerant encoding, one likely requirement of this process is the translation into a topological circuit, for which braided circuits represent one candidate model. Given the large overhead associated with encoded circuits, it is paramount to reduce their size in terms of computation time and qubit number through circuit compression. While these optimizations have typically been performed in the language of three-dimensional diagrams, such a representation does not allow an efficient, general, and scalable approach to reduction or verification. We propose the use of the ZX-calculus as an intermediate language for braided circuit compression, demonstrating advantage by comparing results using this approach with those previously obtained for the compression of A- and Y-state distillation circuits. We then provide a benchmark of our method against a small set of Clifford+T circuits, yielding compression percentages of 77%. Our results suggest that the overheads of braided, defect-based circuits are comparable to those of their lattice-surgery counterparts, restoring the potential of this model for surface-code quantum computation.