Evolving Hybrid Quantum-Classical Architectures for Image Classification

2026-10-02

AI Summary

This study addresses the challenge that quantum circuits in hybrid quantum-classical networks typically rely on manual design, limiting their adaptability to image classification tasks. To overcome this, the project extends the EXAQC framework by employing evolutionary algorithms to automatically search for parameterized quantum circuits (PQCs) as intermediate modules, constructing a hybrid architecture that integrates angle encoding with classical feature extraction layers. The core contribution lies in transcending the limitations of conventionally handcrafted ansätze through automated optimization of quantum circuit structures, thereby significantly enhancing model compactness. Experimental results demonstrate that the proposed method achieves 85.68% accuracy on CIFAR-10 using 25 times fewer parameters and 98.42% accuracy on MNIST, validating the efficiency and superiority of automatically discovered quantum circuits.

Abstract

Hybrid quantum classical neural networks integrate parameterized quantum circuits (PQCs) with established deep learning architectures, but their performance depends strongly on the choice of quantum circuit architecture, a choice that remains largely manual. Most existing approaches rely on hand-designed or fixed circuit ansätze, requiring circuit structure, gate composition, and qubit connectivity to be specified in advance with no guarantee that they suit the task. This limitation is especially acute in image classification, where quantum circuits must transform features extracted by classical networks while remaining compact enough for practical training, requirements that generic, task-agnostic ansätze are unlikely to satisfy simultaneously. We extend EXAQC, an evolutionary framework for automated quantum circuit discovery, to image classification. EXAQC evolves PQCs as intermediate processing modules while retaining classical feature-extraction and prediction layers. On MNIST, Fashion-MNIST, and CIFAR-10, EXAQC achieves 98.42%, 90.62%, and 85.47% accuracy, respectively, while using comparable gate counts to other quantum architecture-search methods. Against classical networks, evolved hybrid models maintain comparable accuracy with substantially fewer trainable parameters, reaching 85.68% on CIFAR-10 with over 25$\times$ fewer parameters than a 10-layer CNN. Encoding choice also matters: rotation-based encodings (RX, RY, U3) outperform amplitude encoding by 22-25 points on CIFAR-10. These results demonstrate that automated circuit discovery yields compact quantum modules that can replace larger classical components in vision architectures while retaining competitive accuracy.

Problem

• hybrid quantum-classical networks
• image classification
• quantum circuit architecture
• parameterized quantum circuits
• automated circuit discovery

Innovation

• Evolutionary Quantum Architecture Search
• Hybrid Quantum-Classical Neural Networks
• Parameterized Quantum Circuits
• Image Classification
• Automated Circuit Discovery