Learning to Learn with Quantum Optimization via Quantum Neural Networks

TL;DR

This paper presents a quantum meta-learning approach using Quantum LSTM networks to optimize QAOA parameters, enabling faster convergence and better solutions for large combinatorial problems on NISQ devices.

Contribution

It introduces a novel quantum neural network-based optimizer trained on small instances that generalizes to larger problems, improving quantum optimization efficiency.

Findings

QLSTM optimizer reduces convergence iterations.

Achieves higher approximation ratios than classical methods.

Demonstrates scalability on complex quantum optimization problems.

Abstract

Quantum Approximate Optimization Algorithms (QAOA) promise efficient solutions to classically intractable combinatorial optimization problems by harnessing shallow-depth quantum circuits. Yet, their performance and scalability often hinge on effective parameter optimization, which remains nontrivial due to rugged energy landscapes and hardware noise. In this work, we introduce a quantum meta-learning framework that combines quantum neural networks, specifically Quantum Long Short-Term Memory (QLSTM) architectures, with QAOA. By training the QLSTM optimizer on smaller graph instances, our approach rapidly generalizes to larger, more complex problems, substantially reducing the number of iterations required for convergence. Through comprehensive benchmarks on Max-Cut and Sherrington-Kirkpatrick model instances, we demonstrate that QLSTM-based optimizers converge faster and achieve higher approximation ratios compared to classical baselines, thereby offering a robust pathway toward scalable quantum optimization in the NISQ era.