Equivariant Quantum Neural Networks for Image Clasification

NeurIPS 2024 Workshop MLNCP Submission47 Authors

Keywords: Quantum Machine Learning, Equivariant Quantum Neural Networks, Image classification, Quantum computing

TL;DR: An equivariant quantum neural network model for image classification, invariant under rotations and reflections.

Abstract: We present a novel Equivariant Quantum Neural Network (EQNN) that takes advantage of common symmetries in image data, specifically rotational and reflective symmetries. By embedding these properties into the design of the quantum neural network, we can greatly reduce the number of parameters that need to be trained, simplifying the model and enhancing its efficiency. This technique allows for improved learning on smaller datasets and better generalization. We test the model’s performance on standard image classification datasets and benchmark it against other quantum approaches.

Submission Number: 47