Abstract: It has never been easier to train and use machine/deep-learning models. At the same time, building real, intelligent software products and systems is harder than expected. It is intriguing that 78% of all such projects fail, with 96% facing challenges with data quality and quantity. A core problem is that data science is model-centric, while many problems in reality (including ethical AI, safety, and security) cannot be solved by focusing on models. In the talk, I will discuss important competences and the field of AI engineering (including MLOps and DataOps, a.k.a. AIOps) needed to build intelligent products and systems that provide actual value. I will provide examples from our application domains in autonomous driving systems, robotics, or other cyber-physical systems. While my background is rather technical in computer science, the talk will provide an interdisciplinary perspective, strongly believing that these are the areas in which we could collaborate across our disciplines.
About the speaker:
Thorsten Berger is a Professor in Computer Science at Ruhr University Bochum in Germany. His research focuses on automating software engineering for the next generation of intelligent, autonomous, and variant-rich software systems—exploring new ways of software creation, analysis, and evolution. Before coming to Bochum, he was an Associate Professor at the Chalmers University of Technology and the University of Gothenburg in Sweden, after stations as a postdoctoral researcher in Canada (University of Waterloo) and Denmark (ITU Copenhagen), and as a PhD student until 2013 in Germany (University of Leipzig), supported by a scholarship from the German National Academic Foundation (Studienstiftung des deutschen Volkes). He received grants from the Swedish Research Council (competitive early-career grant), the Wallenberg AI, Autonomous Systems and Software Program (WASP), Vinnova Sweden (EU ITEA project), the European Union (H2020 project), as well as other EU and German national (e.g., state-funding for a startup in AI engineering) and local funding sources (e.g., from the German Research Council’s cluster of excellence CASA). He received a fellowship from the Royal Swedish Academy of Sciences and the Wallenberg Foundation, one of the highest recognitions for researchers in Sweden. He received best-paper awards at the 2015 ACM SIGPLAN conference on MODULARITY and the 2013 European Conference on Software Maintenance and Reengineering (CSMR, now IEEE SANER), as well as most influential paper awards at the VaMoS’20 and the VaMoS’23 conferences. His service was recognized with distinguished reviewer awards at the A* conferences ASE 2018 and ICSE 2020, and the SPLC 2022 conference.