LASER World of Photonics, the world's leading trade fair for optics and photonics, offers a central forum for all players in the photonics industry from April 26 - 29, 2022. We invite you to visit the Fraunhofer Institute for Cognitive Systems IKS at the booth of the Fraunhofer-Gesellschaft in the exhibition area »World of Quantum« and learn more about our research and innovative solutions for future topics in the field of quantum computing.
Artificial intelligence is becoming increasingly important in medical diagnostics. In the medical context, however, it is particularly important that accurate and reliable decisions are made. This usually requires large training data sets in the training of machine learning methods - but in this field, often only small data sets with 100 or 1000 images are available.
Quantum computing-based algorithms could help to enable training with accurate results even with such small datasets, thus significantly increasing accuracy compared to conventional training methods for tasks in medical diagnostics.
Researchers from the Fraunhofer Institute for Cognitive Systems IKS will present a so-called hybrid quantum-classical convolutional neural network at the joint booth of the »World of Quantum«. Conventional convolutional neural networks are widely used in image recognition, but require large amounts of data to train, especially for very complex tasks. In contrast, at the demonstrator we show how a quantum computing-based variant of this can learn a typical problem from medical diagnostics.
The example considers the identification of breast cancer in ultrasound images of the breast. Using training curves and images, it can be clearly demonstrated how the quantum computer-aided algorithm works in practice. The presented algorithm is already applicable for NISQ quantum computers.