Quantum AI

Quantum computing is expected to be the springboard for a huge leap forward in artificial intelligence (AI). A new interdisciplinary research field called quantum machine learning (QML) has emerged at the intersection of these two key technologies. Explaining how the two are connected, Prof. Christian Bauckhage from the Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS in Sankt Augustin says, “From a mathematical perspective, a lot of AI problems are really problems of combinatorial optimization, which serves to do things like calculate the best delivery routes. If these problems are highly complex with lots of variables, it is very difficult, if not impossible, to find the best solution in a reasonable length of time using today’s digital computers. A quantum computer, by contrast, would be able to solve them in a flash.”

Quantum computers can process large datasets in a single step and identify patterns in the data that conventional computers are unable to detect. Moreover, they might be better capable to cope with by incomplete or corrupt data. In a few years, this fusion of quantum computing and artificial intelligence could have repercussions that echo in practically every area of our lives. QML could help enhance logistics, improve power grid management and optimize investment portfolios in the finance sector. It could also expedite training for large neural networks. Bauckhage and his team at the Fraunhofer Cluster of Excellence Cognitive Internet Technologies are working on a QML project to develop quantum algorithms for problems of combinatorial optimization that are fundamental to machine learning and AI. “We want to be at the forefront of the coming revolution in quantum computing and provide industry with solutions as quickly as possible.”

Fraunhofer is pursuing a similar goal in a project called PlanQK. Together with 14 partners, Fraunhofer FOKUS is seeking to develop a platform for quantum artificial intelligence. This will create a joint forum for AI and quantum computing specialists, developers, users, customers, service providers and consultants to share their knowledge of QML algorithms and applications. One example of a scenario with practical implications is banking fraud, which QML could help detect and even predict. In September 2019, this project won a competition for economically impactful innovations in artificial intelligence sponsored by Germany’s Federal Ministry for Economic Affairs and Energy (BMWi).