David Elsweiler is a private lecturer at the Chair of Information Science at the University of Regensburg. Since the summer semester 2019, he is besides this the academic councilor of the chair. Before he came to Regensburg, he was holder of the “Alexander von Humboldt”-scholarship and lecturer at the University of Erlangen.
In his researches he uses many different techniques to learn how people deal with information. He uses the knowledge gained from this to design tools that support people and work in such a way that they match their thinking and behavior.
In some of his work, including his dissertation, he examines the management and retrieval of personal information. He is also interested in how people look for fun. He also investigates information systems that help people live healthier lives.
Information Science
Chair for Information Science
Empirical Methods, Information Behaviour, Information Retrieval
All of my classes have at least some link to the named topics albeit that they are tangential. For example, an understanding of basic statistical concepts (as is taught in experimental design) is necessary to understand AI and machine learning approaches. Some of the techniques taught in that class relate directly to machine learning, such as regression analysis. In information behaviour and information retrieval classes many of the papers we read and discuss make use of machine learning approaches and some exploit large datasets which could be considered Big Data. In these classes, one goal is always to help students recognize the links between what they are working with and what they may have seen in other classes.
I am fortunate to have regular contact with students as they progress through our degree programme. I am very aware of what they learn at different time points and can try to emphasise links.
I think three strong points that our students have are empirical thinking (they experience different kinds of data collection methods and taught to understand the strengths and limitation of these), solid statistical knowledge, as well as a human focus.
Artificial Intelligence or at least the methods that are associated with this term are valuable in just about every walk of life. An understanding of how they work, what is possible (as well as what is not) will stand students in good stead in their future work and private lives.
My research always revolves around information behaviour in some way. At the moment I focus on two topics: 1) food recommendation, where we attempt to assist people make food choices that they will like to eat, but which also match their goals (healthy, environmentally-friendly etc.) and 2) information credibility, where we try to understand how people judge which information they can trust or not and how we can help them do this effectively.
Artificial Intelligence, while not central to my research, is never very far away. Very often to understand human behaviour we make statistical models and use these models to make predictions about future behaviours (e.g. will a person find this dish tasty? Will this document be judged as credible?). These models make use of AI technologies. My PhD. students have been using deep learning approaches in their projects. For example, Qing Zhang has been using visual processing methods to understand food culture differences and Alex Frummet has been using Bert to predict information needs from spoken utterances.
I have a computer science background, but avoided AI during my undergraduate degree. I chose to focus on software engineering and information retrieval modules. It was only in my post-doc time in Erlangen when I really came into contact with AI. Firstly, my colleagues there helped me analyse data in new and exciting ways. Secondly, I suddenly had to teach these techniques to students, which deepened my understanding and opened my eyes to new possibilities for my research.
I am not really sure if this answers your question, but I am less interested in the technology and more interested in the use cases (i.e. what can we learn or which problems can we solve using this technology?).
I am interested in human behaviour and AI methods provide us with tools that can act as a lens to understanding how people behave, when and why.
I am currently involved in a project with partners in the United Kingdom, which aims to understand how we can build conversational assistants to help older adults improve their nutrition by providing personalized food recommendation. This is a multi-disciplinary group with social-scientists, psychologists, nutritional scientists, as well as computer and information scientists. This is challenging not only because we are developing complex systems, but because our research consortium is melting pot of different ideas and backgrounds. This is what makes it exciting though.
I think the project is an important one. There are so many good people at the University of Regensburg actively applying data science methods and AI technology, but before this project existed I had next to no knowledge of them or their work. This project changes that and I am already aware of much more networking and collaboration as a result of this.
My own contribution so far has been modest. I feature on the website, attend events, including the data science meetup, and am answering requests for interviews, such as this one.
Mr. Elsweiler, thank you for taking the time to answer these questions. We wish you a nice day!