QC Ware worked with the Roche Pharma Research and Early Development (pRED) group to use a quantum neural network architecture to scan retinal images to detect and diagnose the stage of diabetic retinopathy. The study was performed with up to six qubits on an IBM 27-qubit quantum processor along with classical simulations of the algorithms that used up to 100 qubits. The study indicated that the quantum algorithms would perform better than their classical counterparts because they required smaller models, were easier to train, and less resource intensive. Part of the advantage stems from the nature of the quantum algorithm which could analyze a data element from a more global context rather than a classical neural network which might focus only on its more local context. For additional details about this study, you can view a news release provided by QC Ware here and also the technical paper posted on arXiv here.

March 17, 2023