New professor is creating AI that imitates animal cognition
Claes Strannegård, recently promoted to professor of cognitive science at the University of Gothenburg, wants his research to lead to more informed decisions that combine economic benefit with ecological sustainability.
What does being promoted to professor of cognitive science mean to you?
It is fun and inspiring, especially to be promoted in cognitive science, which is an interdisciplinary field that I have worked with for a long time and want to continue to develop. At the same time, I can sometimes feel a certain concern about being perceived as an expert who can comment on all sorts of things, for example about the future with AI. I look forward to continuing to explore how ideas from different cognitive science fields, such as AI and psychology, can cross-reference each other. In the past, I have studied how to use psychological knowledge about limitations in human cognition to make more effective AI programs, who among other things can solve IQ tests. I have also studied how to use AI technology to build computer models of human cognition, for example in logical problem solving.
A personal goal is to create an AI program that can measure up to a fruit fly when it comes to adapting and surviving in previously unseen environments.
How would you describe your research?
My research is very much about trying to create artificial intelligence that imitates certain aspects of the cognition of humans and other animals. A personal goal is to create an AI program that can measure up to a fruit fly when it comes to adapting and surviving in previously unseen environments.
I am working on four different research projects now. One of them aims to build digital twins of real ecosystems. Part of that project is to build AI-based computer models that simulate the cognition of different animals. Another project is a collaboration with Karolinska Institutet, which is about looking for patterns in cancer data using machine learning. A third project is about developing computer games that offer cognitive training for people with intellectual disabilities. I am also involved in researching neural networks in a collaboration with Mathematical Sciences at Chalmers. Among other things, we have studied networks with dynamic architectures and different methods to make the networks generalize better and become more useful. As a next step, I would like to study neural networks that can manage with relatively little energy, just like the human brain.
In what way does your research benefit society?
I think the method we use in the ecosystem project has great potential in terms of utilization. The idea is to construct a digital twin of a real ecosystem, which you then subject to various experiments and observe the consequences for the animal and plant populations. For example, you can see what happens in the computer models if you cut down a forest, build a residential area, or ban fishing in a certain sea area. In this way, more informed decisions can be made, which combine financial benefit with ecological sustainability.
The cancer project can potentially contribute to understanding the origin of certain forms of cancer. We have found seasonal variations in certain types of cancer and wonder if there might be a connection with seasonal viral infections.
In the past, I have worked with research utilization both within the academy, among other things as vice prefect for utilization at the Department of Computer Science and Engineering at Chalmers, and outside the academy, with various IT companies that I started up and managed.
My way of solving issues is to spend time with them and twist and turn them for a long time. In order to do that, you have to think the issue is really interesting.
What motivates you as a researcher?
It is probably mainly a desire to understand certain things. My way of solving issues is to spend time with them and twist and turn them for a long time. In order to do that, you have to think the issue is really interesting. I have always been fascinated by AI and interested in how humans and animals think. I am drawn to issues that are both theoretically interesting and have a clear application potential. Luckily, there are plenty of such issues in cognitive science.
Do you have anything special going on right now that you want to mention?
I am involved in organizing this year's conference in Artificial General Intelligence (AGI-23), which takes place in Stockholm in June. It's fun because general AI is a pretty hot topic right now. We have received a record number of conference contributions and a good line-up of invited speakers with everything from brain researchers to AI experts and robot specialists. Hopefully, ethical issues connected to AI and not least the debate about large language models will get a lot of space at the conference. Also, we have our ongoing seminar series in cognitive science at the department, which is open to everyone.