Transfer Learning for Deep-Unfolded Combinatorial Optimization Solver with Quantum Annealer

TL;DR

This paper introduces a novel classical-quantum transfer learning approach that enhances a deep-unfolded quantum annealer-based combinatorial optimization solver, significantly improving its convergence speed and execution time.

Contribution

It proposes a new classical-quantum transfer learning method enabling the use of QA in trainable solvers, which was previously restricted to classical sampling.

Findings

Improved convergence speed of the quantum COP solver.

Reduced execution time compared to the original solver.

Effective integration of classical training with quantum annealing.

Abstract

Quantum annealing (QA) has attracted research interest as a sampler and combinatorial optimization problem (COP) solver. A recently proposed sampling-based solver for QA significantly reduces the required number of qubits, being capable of large COPs. In relation to this, a trainable sampling-based COP solver has been proposed that optimizes its internal parameters from a dataset by using a deep learning technique called deep unfolding. Although learning the internal parameters accelerates the convergence speed, the sampler in the trainable solver is restricted to using a classical sampler owing to the training cost. In this study, to utilize QA in the trainable solver, we propose classical-quantum transfer learning, where parameters are trained classically, and the trained parameters are used in the solver with QA. The results of numerical experiments demonstrate that the trainable quantum COP solver using classical-quantum transfer learning improves convergence speed and execution time over the original solver.