betaFeedback

WorksPhysics and Astronomy

CAVQS: Context-Aware Variational Quantum Circuit Synthesis via Unitary Redistribution and Dynamic Reconstruction

Abstract

Variational quantum circuit synthesis jointly searches circuit structure and continuous parameters to reduce costly two-qubit gates, but block-based workflows typically instantiate blocks independently, limiting simplification across block boundaries. Moreover, fully unitary synthesis does not exploit opportunities created by permitted output measurements and classical feedforward. We present CAVQS, a context-aware variational synthesis framework that combines unitary redistribution and dynamic reconstruction. First, CAVQS uses QR decomposition to transfer unitary factors between overlapping blocks, preserving the overall transformation while reshaping local synthesis targets to enable cross-block reductions. Second, for blocks that precede permitted terminal measurements, CAVQS constructs ancilla-free dynamic implementations with a CNOT-budgeted pre-measurement search and analytic branch-unitary construction, thereby minimizing expected CNOT cost. Across random-circuit targets, CAVQS reduces mean two-qubit gate counts by 28.79%-42.52%, 2.05%-52.62%, and 30.51%-74.64% relative to QFAST, SQUANDER, and Fast-USYN, respectively, on common successful instances. On 4–6-qubit quantum algorithm benchmarks, the corresponding reductions are 27.27%-36.77%, 20.65%-69.10%, and 36.31%-71.88%, respectively.

Paper details

vlnc:W_qwa7q6gh
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
arXiv nonexclusive license
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…