🤖 AI Summary
To address noise suppression on NISQ hardware, this work proposes an end-to-end quantum circuit compilation method based on differentiable noise modeling and gradient-based optimization. The method introduces numerical gradient optimization (L-BFGS) into quantum compilation for the first time, constructing a noise-aware differentiable cost function that enables continuous gate-parameter optimization and automatic pruning of zero-angle gates. It employs a differentiable physical noise model, partially entangling parameterized gates (e.g., CRX, XX), and GPU acceleration, implemented as a low-latency, low-entanglement circuit compressor within Qiskit’s plugin architecture. Experiments on superconducting platforms show an average 35% reduction in error rates compared to state-of-the-art compilers, while maintaining compilation speed competitive with mainstream tools. The implementation is open-sourced as a general-purpose Qiskit optimizer.
📝 Abstract
We present COGNAC, a novel strategy for compiling quantum circuits based on numerical optimization algorithms from scientific computing. Observing that shorter-duration"partially entangling"gates tend to be less noisy than the typical"maximally entangling"gates, we use a simple and versatile noise model to construct a differentiable cost function. Standard gradient-based optimization algorithms running on a GPU can then quickly converge to a local optimum that closely approximates the target unitary. By reducing rotation angles to zero, COGNAC removes gates from a circuit, producing smaller quantum circuits. We have implemented this technique as a general-purpose Qiskit compiler plugin and compared performance with state-of-the-art optimizers on a variety of standard benchmarks. Testing our compiled circuits on superconducting quantum hardware, we find that COGNAC's optimizations produce circuits that are substantially less noisy than those produced by existing optimizers. These runtime performance gains come without a major compile-time cost, as COGNAC's parallelism allows it to retain a competitive optimization speed.
