A conversation with Hedderik van Rijn

Giving back to society

Inspired by washing-up sessions with his mother, Hedderik van Rijn, experimental psychologist at the University of Groningen, plunged into developing a method to improve how we learn facts. And his efforts were successful. His method resulted in a limited company, but the route to this achievement was anything but smooth.

‘Capital of France?’
‘Paris.’
‘Capital of Iceland?’
Silence. ‘Reykjavik?’ he said, with an audible question mark at the end.

When Van Rijn was in high school, his mother regularly tested his knowledge of topography while they were washing-up together. If he said the name of a place convincingly, then his mother repeated that question less often than when he hesitated. His mother intuitively felt that the place names he hesitated about were the ones he had mastered less well, so she continued to ask about these.

Van Rijn was lucky to have a mother who was a teacher, just like several of his aunts. He therefore always wanted to become a teacher and still enjoys teaching. However, he also liked psychology and wanted to unravel how humans think. And another one of his passions was computers. All of this led to him to studying cognitive sciences at Radboud University, where you learn to capture human behaviour in mathematical models. ‘My work is building things that are similar to how people do it’, is what he says about that. After a PhD at the University of Amsterdam, he did a postdoc at Carnegie Mellon University in Pittsburg (US), where he saw how you can apply knowledge from psychology and artificial intelligence in education. His fellow researchers there were working on implementing models of the human memory in learning systems. They had written an algorithm that could train students to remember facts by simply asking about the facts and then, in a smart way, more often repeating the questions that were answered incorrectly. One experiment had students learning Japanese symbols with the help of a computer program. The computer showed a symbol and the student had to state the meaning. When the students subsequently had to take the actual test, they scored higher than students who had practised with a standard algorithm.

Hedderik van Rijn is Professor of Cognitive Sciences and Neurosciences at the University of Groningen.

He studied cognitive sciences at Radboud University and gained his doctorate in computer models of scientific discovery and cognitive development processes from the University of Amsterdam. As a postdoc at Carnegie Mellon University (US), he studied how information is stored in the brain, and the role played by time in cognitive processes.

'I’d like to spread this way of learning across the world.'

Finally, a breakthrough

Van Rijn was not so impressed with the outcomes in Pittsburg since the students were allowed a good six hours to learn and had to learn things that were not relevant for them. ‘That did not concur in the slightest with a high school pupil who has to learn French for a test that counts towards their final mark and who only wants to spend 10 minutes cramming.’ Once he had taken up his appointment at the University of Groningen, Van Rijn thought: I’ll show those researchers in Pittsburgh that the results they obtained from the lab cannot be replicated in the everyday reality of a class of pupils at a high school in Groningen. However, he did add a smart trick. Instead of using an algorithm that only learnt from incorrect answers, his algorithm obtained information from good answers. Inspired by the washing-up sessions with his mother, he also made the time factor part of his algorithm; the time it took somebody to come up with an answer. As a result, the algorithm not only repeats the questions that received a wrong answer, but also those that were answered correctly but hesitantly. The result? With a score of 6.5, ‘the Pittsburg algorithm’ scarcely scored higher than the control group, but the Groningen algorithm achieved a score of 7.5. Nevertheless, Van Rijn was still not particularly impressed. ‘I thought: nice result, but let’s try it again next year.’ Time and time again, the same result emerged, however, even at different schools.

Slowly it began to dawn on Van Rijn that they were onto something. Consequently, the next step would be to incorporate the improved algorithm in teaching methods. ‘I’d like to spread this way of learning across the world because not everybody has a mother who can quiz them.’ And that is when the hard work began. ‘The step from research to practice is rather huge. How should you go about it? Which people do you need to approach? I had no idea. And especially because in the social sciences, there is so little experience in this area.’ It just so happened that the educational publisher Noordhoff was looking to provide better ways to learn vocabulary and one of Van Rijn’s mentor students was doing an internship there. Thus, a splendid, broad application was found.

 

In 2005, he was appointed assistant professor at the Department of Artificial Intelligence of the University of Groningen. Four years later, he made the switch to the Department of Psychology, where he was the chair of the basic unit experimental psychology for several years and helped to establish the Research Master Behavioral and Cognitive Neurosciences.

In 2017, he received a prestigious Vici grant for research into how people estimate short time intervals. In addition to his academic work, he has been the director of SlimStampen BV since 2021.

‘The step from research to practice is rather huge.'

Frustrating process

But when foreign clients also started to show interest, it quickly became apparent that the existing contracts for intellectual property, such as those concluded with Noordhoff, fell short of the mark. This was mainly because these new clients wanted certainties that a research institution could not provide. Therefore, a spin-off was founded, partly at the initiative of the University of Groningen. However, the next problem immediately presented itself: who should run the spin-off? The intellectual property for the work was owned by the university, whereas the copyright belonged to the researcher. And at that moment, the University of Groningen employees could not own more than 4.99% of the shares in a company, which meant that a researcher could not be extensively involved in a business. Yet that kind of involvement is needed to make a company attractive for investors. Another problem was that algorithms like this cannot be patented in Europe, which makes it hard to transfer knowledge. ‘Investors do not want a company in which the founder scarcely has any shares because that is too great a risk. It would mean that if I left and took my knowledge with me, they would be left behind empty-handed.’

Therefore Van Rijn needed to play an important role in the company himself. ‘But I was not the type of person who had always dreamed of establishing a company. Plus, as an academic, you already have to deal with rather a heavy workload. So where would I find the time? I am more interested in the content; to reduce the inequality of opportunity between children. That gives me a greater sense of satisfaction than the idea that I could be running my own company.’ Everything would have been much easier if Van Rijn had simply chosen to leave academia and run his company, but he very much wanted to pursue his academic career. Meanwhile, it was already clear from a number of other spin-offs from the University of Groningen that the 4.99 % ruling did not work. Consequently, the university’s regulation concerning external activities was modified, partly on the basis of Van Rijn’s case. Now the agreement is that if he works at the University of Groningen for a maximum of 80% of his time, he may own 20% of the shares in his company.

This outcome was not easily arrived at. It took Van Rijn two years of negotiations. ‘That is time I could not devote to my research and company'. However, the problem was solved in the end, and because he did that via all official procedures, the University of Groningen now has a structure in place that others can copy. ‘I am pleased that my colleagues can now benefit from this new situation because we are at the bottom of the pile when it comes to knowledge utilisation, and a more stimulating regulation could help to improve that.’ Previously, there existed an unclear situation in which a new solution had to be improvised for each new company, which often led to undesirable situations that had to be corrected in retrospect. Now the “procedural hygiene” is safeguarded, says Van Rijn.

Although the process was frustrating, he also understands that universities have a difficult role in this regard. ‘They find themselves in a tight corner. On the one hand, they have to focus on fundamental research, but they also want to serve society. And they want to do more outreach, but society also demands that universities prevent conflicts of interest. So the path they have to follow is very narrow and many universities are still trying to find their feet in this regard.’ Van Rijn effectively had to reinvent the wheel with his employer because he was the first of his kind in many areas. The university’s intentions were always good, he says, because generating spin-offs, certainly in the social sciences, is an important aim of the university.

Full circle

In 2021, the company could be established at last and it was given the name SlimStampen (Smart Cramming). Noordhoff has incorporated the method in its teaching modules, as a result of which 800,000 high school pupils in the Netherlands now use it. Publishers in Belgium and Switzerland also use Van Rijn’s method, as well as students at universities in Ghent, Groningen, Utrecht and Seattle. The next step is that adults outside of universities will also be able to use this method, such as people who have to follow safety courses, for example, to work at a drilling platform or as a pilot. For them, it is also important that they not only master the knowledge at the moment of testing but also retain it subsequently. Besides helping with the testing, SlimStampen could therefore also be used to derive the knowledge level from the learning process. In addition, Van Rijn is also deploys SlimStampen to examine whether speech signals, such as intonation, can be predictive for the learning result. An example is the audible question mark when you hesitate a moment before giving an answer.

‘Society gives many possibilities to conduct completely curiosity-driven research and I will always be a firm supporter of that. But it is also good to occasionally pause and consider whether you can give something back to society that it genuinely benefits from. I’d like to be an example in that respect and show that my research leads to applications and that those applications, in turn, lead to new research. Then you have come full circle.’

Three points for further reflection and discussion:

  • How can we make it easier for social scientists to set up a company?
  • How do you prevent conflicts of interest?
  • How do you ensure that top science and the generation of spin-offs go hand-in-hand and strengthen the reputation of a university?

Read the stories of other entrepreneurial scientists