HAISEM-Lab – Lab on Hardware-optimized AI Applications using modern Software Engineering Methods

The diverse usage of Artificial Intelligence (AI) has extensive impact on the society, but also on technical systems and related processes. Nowadays, systems and processes are driven by various types of (sensor) data, novel options to control and automatically adapt systems as well as the increasing interconnectedness of systems. Example application areas are intelligent production (Industrie 4.0), intelligent mobility or energy supply. These changes significantly affect the work of engineers and technicians and creates new challenges for the development, evolution, usage and maintenance of intelligent AI-based systems.

However, the current shortage of required skills in many areas hinders the effective usage of AI-methods in the practice. Current studies show that in particular missing know-how and a lack of skilled personnel are a serious obstacle for the implementation of novel AI-applications. This skill shortage slows down the realization of planned AI-projects.

The goal of the project HAISEM-Lab, which is funded by the German Ministry of Education and Research (BMBF), is to create an AI-lab based on a sustainable and flexible concepts for education and training. HAISEM-Lab shall empower developers to realize effective and efficient AI-applications for modern hardware architectures through Software Engineering Methods. To realize this goal, the working group SSE (University of Hildesheim) collaborates with the Universität Hannover (Forschungszentrum L3S, Institut für Mikroelektronische Systeme) to build up a (regionally distributed) AI-Lab. HAISEM-Lab will be based on the existing education and training offers of the  Applied Machine Learning Academy (AMA) and extend these offers for the specific topics of system development, software engineering, and specialized hardware architectures such as graphical processing units (GPU), tensor processing units (TPU) or field-programmable gate arrays (FPGAs).

The research at the Universität of Hildesheim will concentrate on software engineering methods for hardware-supported AI applications, in particular on software architectures, performance engineering, model-based software development or search-based software engineering.

Runtime: 2 years

Contact: Dr. Holger Eichelberger

The HAISEM-Lab project is being funded through the programme "Künstliche Intelligenz" (BMBF) in the funding area "KI4 - KI-Labore".