Rigetti Computing Inc., a California-based quantum computing company and pioneer in full-stack quantum-classical computing, has recently announced their current progress on the performance of a typical machine learning model for identifying breast cancer and pneumonia through hybridization using a Rigetti quantum co-processor.
Rigetti recently described a proof of concept for a quantum machine learning (QML) application to improve the performance of a standard AI model for identifying cases of breast cancer and pneumonia using MedMNIST dataset collection.
Convolutional neural networks (CNNs) are a well-established machine learning method that frequently displays cutting-edge performance on various machine learning problems, particularly in image processing and classification. Initially presented in 2019, quantum convolutional (“quanvolutional”) neural network layers function on input data similarly to randomized classical convolutional layers. These randomized quantum circuits provide additional properties to the data that may be impossible for classical computers to accomplish, thereby providing a practical technique to include quantum computers in machine learning and artificial intelligence (AI) pipelines.
In Rigetti experiments, Rigetti’s quanvolutional neural network method improved the performance of a standard machine learning model for identifying breast cancer and pneumonia. While the proof of concept does not demonstrate quantum advantage, it has already been considered an essential step toward analyzing and developing the potential application of quantum-enhanced image classification categorization in real-world scenarios.
These QML applications are intended to be available primarily on the Strangeworks platform. The quantum kernel and quantum convolutional “quanvolutional” neural network algorithms are meant to accelerate the development of applications related to regression and classification problems. They are explicitly tuned for Rigetti quantum computers.
Read more about it here.