29 June 2021, Iris Proff
© Arco Mul
Sonja Smets had a challenging task: For five years, she managed and supported researchers from two faculties and diverse fields. She brought together logicians, linguists, cognitive scientists, musicologists, mathematicians, computer scientists and philosophers and helped shaping a common direction for the ILLC. She did so with great competence, curiosity and idealism and was always eager to help the members of the institute perform at their best.
The art of interdisciplinarity
Sonja, are you sad to leave the director position behind? Or are you happy to focus more on your own research in the future?
Sonja Smets: Both. I already have many plans for new research. But giving things out of hand is not easy. Of course, I am still going to care a lot – that will not stop.
What were your highlights of your years as a director? What are you proud of?
We continued strengthening our research but also took into account new developments, for instance in AI. When I started my directorship in 2016, I wanted to make the ILLC also more visible at the Bachelor level as we were already doing excellent at the Master level with our Master of Logic. We designed a new minor in logic and computation and that’s going really well. Another highlight was that we installed two new joint PhD programs with the University of Sint Andrews in Scotland and with Tsinghua University in China.
Is it difficult to manage an interdisciplinary institute?
That was a challenge, indeed. When I became a director, I could relate to a large chunk of the research at the institute from my experience as a logician. But there were some areas like machine translation or the work on predictability of music that were entirely new to me. I learned a lot from everybody, and I learned very fast, because I had to.
Once, we were asked by a journalist to give our opinion on machine translation and natural language processing. I called Khalil Sima’an who is an expert in that area to have the interview. But I wasn’t sure he could come to the appointment, so I took two days to study the whole topic. It was a real crash course. My whole desk was full of post-it notes, to make sure I covered every possible question. In the end, Khalil came to my office and gave excellent answers. I was so happy!
Sonja Smets: between epistemology and quantum logic
- In epistemic logic and epistemology, Sonja Smets investigates how individuals and groups form beliefs, convert their beliefs into knowledge and how these beliefs can shift over time.
- In quantum logic and quantum information, she designs logical formalisms to study the information states of quantum systems.
- Before joining the ILLC in 2012, Smets researched at the Free University of Brussels and at the University of Groningen and visited numerous other research institutes all over the world.
- She holds a visiting professor position at the University of Bergen and is the vice president of the Association for Logic, Language and Information.
Logic and AI remain entangled
Logic and AI are central topics at the ILLC and they are often portrayed as competing fields. Is it an opportunity that AI is booming at the moment? Or is a threat to research in logic?
This is clearly an opportunity. AI was an interdisciplinary field from its beginning. Cognitive science and logic had a strong influence on its development. Mainly the data driven approach is booming today. But there are lots of challenges concerning how to make data driven applications responsible and interpretable. Here, input is needed from other areas. Logicians and cognitive scientists can help shed some light on these big open problems. Responsible AI is one of the areas where a lot of research at the ILLC comes together.
Is this also what your work in the university’s Humane AI research priority area is about?
Currently there is a lack of trust in AI systems. They entered society very fast and gave a lot of power to the people who are hosting them. Governments and society now ask the designers of AI systems to take more responsibility and to make sure that their systems are explainable and make fair decisions. That is what Humane AI focuses on. We combine designs by computer scientists with input from philosophy, social science and law to build applications that accommodate society’s needs and to ensure a better integration of AI in society.
We want to make sure that something like the Cambridge Analytica scandal cannot happen again. We should not cut down our AI applications, but improve on them and use them for what they are meant for: not to replace us, but to help us.
In its beginnings, the ILLC was mainly famous for its fundamental research in formal logic, which was not directly linked to the real world. Is the institute’s research moving closer to reality?
I don’t think our work isn’t directly linked to the real world. We still address fundamental questions and they are always connected to the real world, but there might be many steps in between.
But you do see a growing interest of society in fundamental questions, such as: What is intelligence or rationality? We as researchers have the duty to explain what we are doing more than we had before. That is a good development. But it can be difficult for mathematicians and logicians, as it is not always easy to explain where our research will have an effect in the future. We can give examples from the past, but there is a big unknown about future impact.
Heading towards a quantum era?
Next to the booming AI research, what changed at the ILLC over the last 5 years?
There is a growing interest in quantum theory. The Dutch government has recently announced to invest 615 million euros in quantum research and quantum applications over the coming years. This means this field is now taking a big jump forward. Similar to AI, quantum research will no longer be only interesting to computer science and physics. A lot of societal questions will pop up, requiring interdisciplinary research in which computer scientists and physicists work together with researchers in the humanities and social sciences.
Part of your own research is in that area. You investigate the logic underlying quantum information states. How is this linked to the real world?
We work with the basic principles from quantum physics such as entanglement and superposition of states and try to put these into an axiomatic system. It can be used to specify what happens in a quantum protocol and to verify its correctness. If you build a quantum computer, such an axiomatic system is ultimately what you need. You have the physical level with the actual implementation, just like the circuits described by logical gates, zeros and ones, in a classical computer. All that information needs to be translated into a programming language which you can use to operate the quantum computer. Quantum logic can help make that translation. Ultimately, everybody should be able to program with it!
Gender equality in science
Logic research has historically been male dominated. You are the first female director of the ILLC and one of few high-ranking female researchers in your field. Are things moving forward in terms of gender equality?
When I did my PhD, there were a few female role models in the quantum logic community. They were extremely powerful and full professors already. Looking back, I’m really happy these women were there. But all the rest was male – I always wondered how this was possible.
I think it is changing. Gender inequality is something that a lot of universities are fighting against. Also at the ILLC it became a major focus. At some point we had no female researcher in a permanent position within one of the sections of our institute. What I did was controversial: I went to the faculty and argued for a need to open a position only for women. This way we hired a woman and made a first step forward. It was urgent.
Why was it controversial?
It’s not a trivial decision, because you are discriminating against men. Our law says that you cannot discriminate in a vacancy, unless you can show that you have done everything else to fight inequality and you didn’t succeed. We had a good case and never entered into a lawsuit. Other universities did – some won, some did not win.
In many technical fields, the rates of males and females are still quite balanced at the PhD level. But for postdocs and professors, this often changes. How to keep more female PhD students in science?
Times are changing, a lot of women joined the Faculty of Science recently on permanent positions. This creates a more welcoming atmosphere for other women. Sometimes conferences offer workshops specifically for women, where young female scholars can present their work. These events also help. We now have more and more female PhD students that graduate and want to continue in research. They see what their career could look like and that it is compatible with having a family life.
But it’s not an easy question. An academic career comes with many insecurities and temporary contracts. Plus, especially now, if you are very good in computing and AI, then there will be many companies interested in you which pay much more. The downside is that you have to work on what the company dictates; you often do not have the same academic freedom. My hope is that young researchers try it for a while and eventually come back to academia.