Quantum circuit optimizer uses machine learning for better efficiency
Bridge of $Ψ$'s: Quantum Circuit Optimization with Schrödinger Bridges
Machine Learning
Summary
Quantum circuits are instructions for quantum computers, and making them shorter and simpler helps computers work faster and with fewer errors. The authors developed a new method called BOPS that learns how to rewrite these circuits into simpler versions by training on examples, instead of following fixed rules. They designed a special machine learning model inspired by Schrödinger bridges to do this. Tested on tough circuits, BOPS cut the number of steps and layers in half compared to existing methods.
What this means in practice
- •For quantum hardware engineers: Reduce gate counts and depth in quantum circuits to lower hardware execution cost and error rates.
- •For quantum software developers: Automatically optimize quantum programs using learned models improving over standard rule-based compilers.
Authors
Lino S. Hofstetter, Lia Yeh, Prakash Murali
Abstract
Quantum circuit optimization replaces a circuit with an equivalent one of fewer gates and lower depth, reducing execution cost and error rate. We ask whether a generative model can learn this transformation directly from examples, rather than selecting from a fixed rewrite library or rigid algebraic routines. We present Bridge of $Ψ$'s (BOPS), a generative model based on Schrödinger bridges, using a custom denoiser architecture, that learns a transformation from a source circuit into an equivalent optimized circuit. We train it on data constructed to be hard for existing optimizers, by applying rewrite rules backwards so that each input has a known lower-cost target. On held-out 8 qubits $\times$ 64 depth Clifford+$T$ circuits, BOPS reduces gate count by $2.46\times$ and depth by $2.45\times$ in geometric mean, outperforming all nine baseline optimizers. This constitutes the first generative model bridging quantum circuits and frontier machine learning methods, opening up the quantum compilation stack to learned optimization along multiple axes.