Generative Quantum Machine Learning Methods for Cancer Detection using Augmented Datasets
Authors: Sophia Knutson, Nandika Goyal, Gennaro De Luca, Glen Uehara, Tanay Kamlesh Patel, Andreas Spanias
Organizations: School of Computer Science, Carnegie Mellon University, Pittsburgh, PA, USA · SenSIP Center, Arizona State University, Tempe, AZ, USA · School of Computing and Augmented Intelligence, Arizona State University, Tempe, AZ, USA
Cancer detection using AI methods and image databases faces several challenges, including limited and imbalanced image datasets, long algorithm training times, and high-complexity feature generation. In particular, many cancer image datasets contain very few malignant examples compared to benign ones, limiting the amount of minority-class data available for training. Although classical AI and especially deep learning methods have had some success in providing cancer diagnostics, Quantum Machine Learning (QML) methods may offer advantages through high-dimensional feature representations, generative data augmentation, and the exploitation of feature correlations through quantum properties such as entanglement. In this paper, we propose using a generative QML method on the compressed breast ultrasound image dataset (BUSI) to generate synthetic images and hence augmented datasets for classifier training. In particular, we train a convolutional neural network with real, synthetic, and a combination of real and synthetic images. We test the algorithm on real images and provide comparative results for all three training cases. Our results show that the classifier trained entirely on synthetic images achieved an AUC-ROC of 0.802 (80.2%), compared with 0.760 (76.0%) for the real-data baseline.
Figures & tables
Figure 1: Overview of the proposed Quantum Circuit Born Machine (QCBM) method for breast ultrasound data generation, augmentation and classification. The BUSI dataset along with a variational autoencoder (VAE) are used and classification is performed with a convolutional neural network (CNN).
Figure 2: Overview of the proposed pipelines. (a) Augmented method (training the classifier with real and synthetic data ) (b) Synthetic-only method (training the classifier on synthetic data only).
Figure 3: Image comparison; original shown on top versus VAE reconstructed images for the malignant class shown at the bottom.
Figure 4: Design of a 16-qubit IQP circuit used in QCBM.
Arm
AUC-ROC
AUC-PR
F1
Accuracy
Baseline
.760 ± .008
.643 ± .024
.070 ± .080
.743 ± .012
Class-weighted
.751 ± .006
.632 ± .026
.577 ± .014
.752 ± .026
Augmented
.752 ± .010
.604 ± .034
.221 ± .047
.754 ± .008
Synthetic-only
.802 ± .030
.645 ± .049
.484 ± .147
.694 ± .155
Table 1: Test-set performance (mean ± standard deviation across ten seeds). All arms share the same classifier, hyperparameters, and real test set.
Condition
Std. Dev.
Real malignant images
0.1115
QCBM
0.0070
QCBM ( T=2.0 , noise =0.05 )
0.0053
QCBM ( T=3.0 , noise =0.10 )
0.0056
GMM (binarized)
0.0065
GMM (continuous)
0.0052
Table 2: Bottleneck analysis based on predicted malignancy probability distributions.
Figure 5: Quantum hardware evaluation pipeline. The trained IQP QCBM is sampled using noiseless simulation, noisy emulation, and the IonQ Forte 1 processor. The resulting distributions are compared to characterize hardware-induced deviations.
BMW Group, Munich, Germany · Department of Computer Science and Institute of Applied Mathematics, University of British Columbia (UBC), Vancouver, Canada · Stewart Blusson Quantum Matter Institute (QMI), Vancouver, Canada +1