Amsterdam University of Applied Sciences combines AI and practice

The Netherlands’ first university of applied sciences, Amsterdam University of Applied Sciences (HvA), has launched a centre of expertise in applied artificial intelligence (AI), where students from all faculties – from nursing and ICT to teacher training – will learn how to apply AI in their field of study.  

Lecturers and researchers in various application areas at the university had already been working on AI for some time. Nanda Piersma, scientific director at the HvA’s centre for applied AI, drew up a plan with colleagues Pascal Wiggers and Martijn de Waal to bring all these initiatives together in a structured way.

“Applied AI is significantly different from academic AI,” said Piersma. “In the scientific world, the development of innovative AI algorithms is central, whereas the goal of applied AI is primarily successful implementation. This is not about the most beautiful or best algorithm, but rather about understanding the application of AI.” 

The strength of the expertise centre lies in the fact that not only AI experts from the HvA are involved, but they work closely with experts from the field of AI application. “Co-creation is the big secret to actually applying AI in the world,” said Piersma. “You can’t do AI in your own little office in India or Poland. You have to do it together with people from the field in which the AI is applied.”

The seven laboratories the HvA has set up at the expertise centre are used by people with domain knowledge and experts who can program the algorithms. “This gives us the unique opportunity to generalise findings from one lab so that it can be applied in other domains,” said Piersma. “Findings become transferable to a different context.”

In the Netherlands, there is an urgent need for more knowledge and talent around AI, especially from small and medium-sized enterprises (SMEs) and the public sector. The HvA is responding to this need with the centre of expertise as part of the AI Technology for People partnership, under which Amsterdam is making significant efforts in the field of AI technology.

Participants in the AI Technology for People partnership include Amsterdam’s knowledge institutes, the City of Amsterdam, Amsterdam Data Science and the Amsterdam Economic Board.

“We want all students at the HvA to have the skills and knowledge to apply AI in their own field,” said Piersma. On the one hand, the HvA is contributing to more knowledge about AI in the region, and on the other hand, it prepares students better for the future, as many professions will change due to new applications of AI, she added.

“For example, we have developed an AI system for architects that automatically searches for reference locations,” said Piersma. “But for the application of AI in a domain to be successful, it is important that people learn to work with it and become familiar with it.”

This is why the centre of expertise works closely with project participants during the development and design of an algorithm, so they are trained during the project and can start working with it immediately. “This way, architects can expand the system themselves with, for example, additional filters,” she added. “When they understand how it works and can adjust it themselves, this ensures a higher adoption rate.”

Adoption is an important goal for the success of the centre of expertise, said Piersma. “The first success is the centre itself, but in addition, we want to make AI understandable so that people in the field can easily adopt it,” she said. “We do this by developing solutions with domain experts and the training programme. A project can start small, but then it can be scaled up and people can get on with it themselves, like architects, for example.”

In the short term, the centre of expertise wants to develop successful AI products with the market. In the medium term, after the initial startup, the intention is for these successes to be picked up and continued in the field. “Ultimately, we hope people will have more confidence in how algorithms can work for them,” said Piersma. “That it comes closer to people, and they embrace it.”