Practical insights on the effect of different encodings, ans\"atze and measurements in quantum and hybrid convolutional neural networks

TL;DR

This paper systematically evaluates how data encoding, variational ansätze, and measurement choices affect the performance of quantum and hybrid convolutional neural networks in satellite image classification, revealing data encoding as the most influential factor.

Contribution

It provides a comprehensive analysis of design choices in PQCs for quantum and hybrid CNNs, highlighting the dominant role of data encoding strategies.

Findings

Data encoding strategy significantly impacts hybrid model accuracy, with over 30% variation.

Variational ansätze and measurement basis have a minor effect, under 5% accuracy change.

In purely quantum models, measurement and data-to-amplitude mapping are most influential.

Abstract

This study investigates the design choices of parameterized quantum circuits (PQCs) within quantum and hybrid convolutional neural network (HQNN and QCNN) architectures, applied to the task of satellite image classification using the EuroSAT dataset. We systematically evaluate the performance implications of data encoding techniques, variational ans\"atze, and measurement in approx. 500 distinct model configurations. Our analysis reveals a clear hierarchy of influence on model performance. For hybrid architectures, which were benchmarked against their direct classical equivalents (e.g. the same architecture with the PQCs removed), the data encoding strategy is the dominant factor, with validation accuracy varying over 30% for distinct embeddings. In contrast, the selection of variational ans\"atze and measurement basis had a comparatively marginal effect, with validation accuracy variations remaining below 5%. For purely quantum models, restricted to amplitude encoding, performance was most dependent on the measurement protocol and the data-to-amplitude mapping. The measurement strategy varied the validation accuracy by up to 30% and the encoding mapping by around 8 percentage points.