Reducing Mid-Circuit Measurements via Probabilistic Circuits

TL;DR

This paper introduces a static circuit optimization method for quantum circuits that replaces some mid-circuit measurements with randomized gates, reducing hardware demands while maintaining circuit functionality.

Contribution

It presents a novel polynomial-time optimization algorithm that precomputes measurement outcome probabilities to eliminate certain mid-circuit measurements in quantum circuits.

Findings

Reduces the need for classical feedback in quantum hardware.

Maintains circuit equivalence while simplifying measurement requirements.

Efficiently scales with circuit size and complexity.

Abstract

Mid-circuit measurements and measurement-controlled gates are supported by an increasing number of quantum hardware platforms and will become more relevant as an essential building block for quantum error correction. However, mid-circuit measurements impose significant demands on the quantum hardware due to the required signal analysis and classical feedback loop. This work presents a static circuit optimization algorithm that can substitute some of these measurements with an equivalent circuit with randomized gate applications. Our method uses ideas from constant propagation to classically precompute measurement outcome probabilities. Our proposed optimization is efficient, as its runtime scales polynomially on the number of qubits and gates of the circuit.