Large Language Models (LLMs) have found their way into many applications. Still, they are often seen as toys to play around with, but hard to use in a productive environment, due to phenomena like hallucinations, bias and missing reproducibility. The current presentation is a short form of my lecture series ‘Einsatz von Sprachmodellen für das Studium der Festkörpermechanik’ for engineering students. The aim is to show examples for useful integration of LLMs into workflows, while avoiding hyped but not-yet-useful concepts. It is focused on providing an introduction to the field of Generative AI, Tokenzing & Embedding, Cognitive Architectures, and prompting strategies. In a seminar-style format, there will be time to try out different free chatbots and to discuss the outcomes, therefore it’s highly beneficial for the participants to bring their own laptop (or external keyboard for fast typing if a smartphone is the preferred platform).
Dr.-Ing. Adrian Ehrenhofer is a postdoc at the Institute of Solid Mechanics of Technische Universität Dreden (TUD), Germany. He completed his Diploma in 2014 and his PhD in 2018 at TUD. During his PhD, he focused on multi-field modeling and simulation of active/smart/intelligent materials. He subsequently developed an analogy description for the swelling behavior of active hydrogels, called Stimulus-Expansion-Model. During his Postdoc time he further worked on a unified description of Soft-Hard Active-Passive Embedded Structures at TUD. He is also active in the field of multi-field approaches for histopathology and was a visiting scholar at the University of Utah, Salt Lake City. Since 2021, he holds the position of Research Group Leader of the Materials Informatics Group at the Dresden Center for Intelligent Materials (DCIM), which focuses on applying machine learning methods towards material discovery.
Large Language Models (LLMs) have found their way into many applications. Still, they are often seen as toys to play around with, but hard to use in a productive environment, due to phenomena like hallucinations, bias and missing reproducibility. The current presentation is a short form of my lecture series ‘Einsatz von Sprachmodellen für das Studium der Festkörpermechanik’ for engineering students. The aim is to show examples for useful integration of LLMs into workflows, while avoiding hyped but not-yet-useful concepts. It is focused on providing an introduction to the field of Generative AI, Tokenzing & Embedding, Cognitive Architectures, and prompting strategies. In a seminar-style format, there will be time to try out different free chatbots and to discuss the outcomes, therefore it’s highly beneficial for the participants to bring their own laptop (or external keyboard for fast typing if a smartphone is the preferred platform).
Dr.-Ing. Adrian Ehrenhofer is a postdoc at the Institute of Solid Mechanics of Technische Universität Dreden (TUD), Germany. He completed his Diploma in 2014 and his PhD in 2018 at TUD. During his PhD, he focused on multi-field modeling and simulation of active/smart/intelligent materials. He subsequently developed an analogy description for the swelling behavior of active hydrogels, called Stimulus-Expansion-Model. During his Postdoc time he further worked on a unified description of Soft-Hard Active-Passive Embedded Structures at TUD. He is also active in the field of multi-field approaches for histopathology and was a visiting scholar at the University of Utah, Salt Lake City. Since 2021, he holds the position of Research Group Leader of the Materials Informatics Group at the Dresden Center for Intelligent Materials (DCIM), which focuses on applying machine learning methods towards material discovery.
