Machine learning: Help for engineering and medicine
Roland Aydin is a new professor at the TU Hamburg
25.10.2023
Artificial intelligence (AI) has arrived in our everyday lives. For example, you feed a text-based program such as ChatGPT with a few keywords and it spits out a text on the desired topic in a fraction of a second. The program uses machine learning technology for its answers. Among other things, Roland Aydin is working on applying such "large language models" as part of machine learning to the field of engineering. "My aim is to adapt the latest innovations from AI research for our fields of research, such as so-called transformer architectures, which also form the basis of ChatGPT, for example. It is possible that the definition of language or grammar is much more general than we thought and, with a little creativity, we could use current language models in innovative ways," explains Junior Professor Aydin, offering an insight into his research area "Machine Learning in Virtual Materials Design".
Combining computer simulation and machine learning
By efficiently coupling computer simulations and machine learning, advantageous models can also be developed for other applications of materials science, such as in medicine, for example by working with so-called digital twins. They provide a digital image of reality. The computer scientist explains: "An important goal of mine is also to make machine learning accessible to research groups in materials science who do not have expert knowledge in the field of data science. By making such technologies more widely usable, we can find better and more efficient solutions to real-world challenges that arise in many fields of science and engineering."
Interdisciplinary training for AI experts
Roland Aydin studied computer science at the Technical University of Munich and the Georgia Institute of Technology in the USA. He then went on to study human medicine at the LMU in Munich. His first professional position took him back to the TU Munich as a research assistant at the Chair of Computational Mechanics, specializing in AI and biomechanics. In 2018, he moved to the Helmholtz Center Hereon with the task of developing or adapting AI algorithms for materials science. The joint appointment allows the TU professor to continue his work at Hereon parallely, where he is building up the Department of Machine Learning and Data. "Training the next generation of interdisciplinary AI experts is a key aspect of this," says Roland Aydin.
Aydin considers it a privilege to be rewarded for learning and researching new things as part of the professorship, especially in an area as transformative for humanity as AI. "My children give me new perspectives on the world every day," says the proud father of four sons. In his free time, Aydin plays chess and is passionate about reading books in "the classic genres of a computer scientist": science fiction and fantasy. His research interests also shape other hobbies in his private life, such as generating texts with ChatGPT: "The beautiful architecture of Hamburg with its canals and cultural richness offer my family and I an environment worth living in."
TUHH - Public Relations Office
Elke Schulze
E-Mail: elke.schulze@tuhh.de
