Artificial intelligence flourishing in Leiden, Delft and Rotterdam
Leiden University, Delft University of Technology and Erasmus University Rotterdam are important players in the world of self-learning machines that can work together with people. Moreover, the three have close ties in the field of AI and hope to further strengthen this collaboration in the future. An overview of artificial intelligence in Zuid-Holland.
This article was previously published in LDE Magazine, a magazine about collaboration between the universities in Leiden, Delft and Rotterdam.
From image recognition in MRI scans to information searches in law firms, artificial intelligence is everywhere. The same is true for universities, which are conducting research across the entire spectrum of this acclaimed yet sometimes maligned field. At Leiden University, much but by no means all of this takes place at the Leiden Institute of Advanced Computer Science (LIACS). AI is everywhere, from archaeology to biology and from linguistics to healthcare.
Privacy by design
The Faculty of Law, for example, has Bart Custers, who is Professor of Law and Data Science and head of eLaw, the Center for Law and Digital Technologies. One of the centre’s focal areas is to ensure that people don’t get crushed by the cogs of technology. ‘We don’t want to end up in a Kafkaesque society,’ Custers says. He tries to ensure that technicians, ethicists and legal experts work together. ‘Then you can achieve privacy by design: building in all kinds of safeguards beforehand so you don’t encounter problems afterwards.’
Today’s ‘hot’ intelligent systems are no longer merely programmed to perform searches or fast calculations; they are self-learning on the basis of big data. Self-learning systems are making quick advances in the medical world. An AI system can examine a million X-ray images to learn to identify tumours, in theory with a greater degree of accuracy than a radiologist can. The same applies to other forms of imaging technology, such as MRI and CT scans. Leiden is at the forefront of this technology.
Professor in Leiden and Delft
Boudewijn Lelieveldt is Professor of Biomedical Imaging and head of the Division for Image Processing (LKEB) at LUMC, where they are developing intelligent systems that can recognise hidden patterns in huge amounts of imaging data. However, Lelieveldt stresses: ‘No matter how intelligent systems may be, the doctor must always be the one to make the decisions.’
Lelieveldt is a professor in Delft as well as Leiden, at the Faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS), where Catholijn Jonker also works. Like Lelieveldt she holds a post at both universities.
Jonker is a proponent of optimal hybrid intelligence, an interactive system in which humans and systems complement each other. ‘You can compare it to a sheepdog that improves the shepherd’s ability to herd sheep.’ Achieving this requires a dialogue between humans and machines, and at the two universities Jonker can collaborate with the colleagues needed to accomplish this.
In addition to EEMCS, scientists at many other parts of Delft University of Technology are working on artificial intelligence or autonomous systems, be it on pure software, such as decision support programs, on embedded programming in robots, cars or drones, or on research into how to design AI-based solutions in an ethically responsible and people-centred way and how to introduce these to our complex society.
Not just zeros and ones
Jonker’s work and that of eLaw researcher Custers shows that AI research is not just about zeros and ones, but is also about interaction with and consequences for people. This is also true for the AI research at Erasmus University Rotterdam. The Erasmus Centre for Data Analytics (ECDA) coordinates and facilitates AI and data-related research at all the University’s faculties. Twelve professors collaborate in interdisciplinary Expert Practices, such as the audit, accounting and control analytics lab, which develops fraud detection algorithms for accountancy.
One of ECDA’s Academic Directors is Antoinette de Bont, Professor of Sociology of Innovations in Healthcare at the Erasmus School of Health Policy & Management. Her work includes looking from a societal perspective at digitisation and the use of big data: ‘How can you ensure data is stored and re-used safely and securely? How do you determine whether an algorithm is reliable? This raises all kinds of legal, ethical and moral questions, particularly if algorithms fall into the hands of commercial companies.’
Technological solutions for sustainable care...
The three universities and the Erasmus MC and LUMC medical centres also collaborate within the Medical Delta, which also includes universities of applied sciences, healthcare organisations, businesses and local authorities in Zuid-Holland. Since the launch of a major academic programme in 2019, over 280 researchers in 13 consortia have been working on technological solutions for sustainable care. One interesting example that uses AI is wireless neuromonitoring of people with migraine or autism. Here ultrasound is used to measure their brain activity over a longer period of time while in their own surroundings, for an optimum diagnosis. The ultimate aim is to develop software that can also correct their brain activity.
...and for large companies
Besides making improvements in healthcare, the wide AI expertise in the three universities in Zuid-Holland may also benefit major players in the region for whom AI is of crucial importance to competing in the world market. For example, the Rotterdam Port Authority, the petrochemicals industry or other multinationals. The perspective from all academic disciplines makes the Zuid-Holland approach to artificial intelligence unique and relevant to society and academia alike.