AutoQML publishes methodology for linking automated machine learning and quantum computers

The AutoQML project, funded by the the Federal Ministry for Economic Affairs and Climate Action (BMWK), has published a detailed description of the framework that will seamlessly integrate quantum machine learning (QML) algorithms into automated machine learning (AutoML). AutoQML is intended to enable industrial companies to use quantum computers and machine learning for practical applications without having to rely on in-depth specialist knowledge.

Logo AutoQML
© Projekt AutoQML

The use of both artificial intelligence (AI) and quantum computers (QC) promises companies increased efficiency and the possibility of new business models. However, this still requires specific technical knowledge and time-consuming work steps. In order to drive innovation and save resources, it is therefore necessary to simplify the complex tasks of AI-based data evaluation and modelling and accelerate them through the use of quantum computers. The framework described by the project addresses these challenges through the combination of automated machine learning (AutoML) and quantum algorithms and opens up new horizons for achieving performance, speed and complexity advantages in an industrial context.

The AutoQML framework will act as an interface by integrating quantum machine learning (QML) algorithms into established AutoML libraries and allowing automated tasks to be executed on real quantum computers. Hyperparameter optimisation and combinatorial algorithm selection can then also be accelerated using quantum computing. The consortium has now described the framework in detail in a study that can be downloaded from the project website. The first programme libraries have also already been created and made available as open source - more will follow.

"By automating and integrating quantum methods into classic ML processes, the AutoQML framework gives companies low-threshold access to (Q)ML methods without having to rely on in-depth expert knowledge," explains project manager Dr Christian Tutschku from Fraunhofer IAO. "Specialists can also use the toolings to generate (and adapt) initial drafts or to assess the suitability of QML methods for specific problem instances." The first quantum solutions for selected use cases in the application areas of production and automotive have already been implemented and simulated in the project. All of this promises to make it much easier to start using AI solutions and quantum computers for industrial applications. The planned integration into the platform developed in the PlanQK project, which is also funded by the BMWK, should also make the framework widely accessible for companies.

With the publication of the framework, the AutoQML consortium has taken a major step towards its goal of removing barriers to the use of artificial intelligence and quantum computing and introducing experts from small and medium-sized companies in the manufacturing industry to the world of data analysis and modelling. In the remaining project duration, the project members are now working on fully implementing the described solutions and making them available to application companies. By automating processes that normally require in-depth knowledge and a great deal of time and manpower, AutoQML promises to make a significant contribution to accelerating innovation while increasing efficiency in research and industry.

Further Informationen