CAx goes AIAx: Artificial Intelligence learns from experts
Stuttgart/Berlin. It is hoped that, in future, computers will not only make it easier for humans to calculate when designing, but that they will also learn to "think" like a designer. CAx, the abbreviation for the computer-aided processes of design and calculation such as Computer Aided Design (CAD) and Computer Aided Engineering (CAE), are continuing to evolve and, with AIAx, will reach the next level of digital product development. AIAx stands for Artificial Intelligence Aided x. The aim is for Artificial Intelligence to help manage the large, complex amounts of data involved in the digitalisation of the product development process, thereby relieving engineers of time-consuming routine activities. In the next three years, a project funded by the German Federal Ministry of Education and Research (BMBF) under the aegis of Daimler AG will carry out basic research.
Where numerical simulations are used in design and development, computers relieve humans of complex computational processes – a classic example of the efficient and practical division of labour between man and machine. However, human expertise is still required in order to assess the results, for assessment is often complex, with many criteria playing a role, the aim being to achieve several, sometimes competing goals. A key role is played by "soft" criteria, such as the experience, "gut feeling" or "human judgement" of experts. It is necessary to weigh up the situation so as to arrive at the best possible compromise. And, so far, that is something only humans can do.
With the development of Artificial Intelligence, it becomes conceivable to store expert knowledge in computers and apply their decision-making criteria to complex simulation results. The goal of the current AIAx research project is to combine different mathematical processes in order to develop a methodology to enable machines to learn how to make an expert assessment. Specifically, computers are to be fed with human experience to allow them to make worthwhile assessments.
Daimler already employs Artificial Intelligence in a variety of fields, including image recognition and image comprehension in the development of autonomous driving as well as speech recognition and predictive functions in the MBUX (Mercedes-Benz User Experience) infotainment system.
The BMBF-funded project is a further element of comprehensive Artificial Intelligence research, the purpose of which is to explore possible ways of using AI as a tool for products and services as well as for supporting the development and production process.
AIAx speeds up the optimisation of designs
In future, AIAx will use Artificial Intelligence to quickly evaluate simulation results and make an assessment, which can then be understood and validated by experts. Using the knowledge gained, engineers will be able to optimise their designs with even greater precision. To make this possible, AIAx must give clear reasons for the analysis results, because ultimate responsibility for the design will, of course, continue to rest with humans. Consequently, utmost importance is attached to system acceptance, as the aim is for AIAx to optimally support designers in their work.
The project is based, among other things, on the numerical simulations of crash tests conducted by Mercedes-Benz over decades, which have been documented throughout the entire optimisation process of individual products. This is where Artificial Intelligence can learn and understand how experienced experts proceeded in their decision-making and optimisation.
The aim is for the methodology resulting from the research project to be applicable across all sectors to different numerical calculation processes in the field of mechanical engineering. Such a methodology has the potential to become a key technology in the digital transformation of many sectors capable of securing Germany's technological leadership as a centre of industry.
Other partners in the three-year project are DYNAmore GmbH, Stuttgart, Endress+Hauser SE & Co. KG, Maulburg, the Karlsruhe Institute of Technology, the Technical University of Berlin and USU Software AG Spitalhof, Möglingen.
AIAx is funded by the German Federal Ministry of Education and Research (BMBF). The project runs under funding code 01IS18048 A