Quantum machine learning for image classification

TL;DR

Two quantum machine learning models that leverage the principles of quantum mechanics for effective computations are introduced, enabling the execution of computations even in the noisy intermediate-scale quantum era, where circuits with a large number of qubits are currently infeasible.

Abstract

Image classification, a pivotal task in multiple industries, faces computational challenges due to the burgeoning volume of visual data. This research addresses these challenges by introducing two quantum machine learning models that leverage the principles of quantum mechanics for effective computations. Our first model, a hybrid quantum neural network with parallel quantum circuits, enables the execution of computations even in the noisy intermediate-scale quantum era, where circuits with a large number of qubits are currently infeasible. This model demonstrated a record-breaking classification accuracy of 99.21% on the full MNIST dataset, surpassing the performance of known quantum-classical models, while having eight times fewer parameters than its classical counterpart. Also, the results of testing this hybrid model on a Medical MNIST (classification accuracy over 99%), and on CIFAR-10 (classification accuracy over 82%), can serve as evidence of the generalizability of the model and highlights the efficiency of quantum layers in distinguishing common features of input data. Our second model introduces a hybrid quantum neural network with a Quanvolutional layer, reducing image resolution via a convolution process. The model matches the performance of its classical counterpart, having four times fewer trainable parameters, and outperforms a classical model with equal weight parameters. These models represent advancements in quantum machine learning research and illuminate the path towards more accurate image classification systems.

Figures (10)

• Figure 1: a) Examples of images from the MNIST dataset. b) Examples of ambiguous images from the MNIST dataset.
• Figure 2: Examples of images from the Medical MNIST dataset.
• Figure 3: Examples of images from the CIFAR-10 dataset.
• Figure 4: Architecture of the proposed HQNN-Parallel. The input data samples are transformed by a series of convolutional layers, that extract relevant features and reduce the dimensionality of the input. The output channels of the convolutional layers are then flattened into a single vector before being fed into the dense part of the HQNN-Parallel. The hybrid dense part contains a combination of classical and quantum layers. The quantum layers are implemented using parallel PQCs, which allow for simultaneous execution, reducing the total computation time. Quantum layers are depicted in the figure as blue rectangles, the top rectangle is a detailed version of subsequent quantum layers in the amount of $c$ circuits. The output of the last classical fully connected layer is a predicted digit in the range of 0 to 9.
• Figure 5: (a-b) Train and test results for the HQNN-Parallel and the CNN. The HQNN has a 99.21% accuracy on the test data and outperforms the CNN which has a 98.71% accuracy. The classical model has 8 times more variational parameters than the hybrid one. (c) Test accuracies of the HQNN-Parallel and its classical counterpart, the CNN.