“We need to advance AI together to do the right thing”

When Ioannis was a student himself in 1995 and attended the Aristotle University of Thessaloniki, Greece, he was saving up to get a fixed phone line, so that he could have a modem and connect to the university’s internet.

“I was always a geeky person, always good with mathematics,” Ioannis explains. It’s no surprise then that he started his career in electrical and computer engineering, worked in Switzerland and Greece, and came to WUR in 2015, first at the Information Technology Group, and then at the Laboratory of Geo-information Science and Remote Sensing, before his recent appointment with the Wageningen Data Competence Center. His ambition here? “To enable and facilitate data-intensive discoveries in agri-environmental systems research, and to create a dynamic between technology and scientists. That will, of course, have impact within WUR as an organisation, and it’s in our hands to extend that impact to society.”

How does AI help to achieve that?

“Artificial intelligence is the branch of computer science that aims to train machines, such as computers, to make decisions. With artificial intelligence we do not prescribe computers what to do, but rather let them discover solutions out of data or a set of rules. The goal is for machines to be able to solve problems without them being given an explicit solution. For example, in one project we’re building an AI-enabled crop yield forecasting system so that we can identify when and where crops fail at a European scale more quickly and more accurately. Consequently, this information can be used to provide policymakers and farmers with better advice. Such a forecasting system already exists, but we’re improving it with AI that uses data to learn from past situations.”

“Another example is the Digital Future Farm project, in which we’re building AI models that use data to learn how crops grow; we can employ such data to reduce nitrogen use and to make farming decisions smarter. There’s a need for this in the Netherlands, so the project is partly stimulated by that.”

Your work sounds very practical and applied.

“It is, I have an applied focus in my research. I believe it’s important to not just develop tools and algorithms, but to do it with purpose and with a good cause in mind. I want to achieve an integration-first approach, meaning we use the power of AI and data science to transform the understanding about other domains – in our case, food and living environment. Introducing AI in such domains brings several challenges with it, so we have to develop new methods and models that are tailored to these specifically, while we reuse existing knowledge in a better way.”

How much collaboration with other domain scientists does that require?

“A lot! AI advances in teams, not in individuals. My research is all about how data and data-driven processes help decision-making, which is complex, because it requires deep knowledge of data systems as well as a very deep understanding of the problems within the domains. You cannot do one without the other. That’s also why WUR is very important in this area – Wageningen has both the AI and the domain expertise, and I find it absolutely fascinating to team them up.”

“Actually, this ties in with another ambition of ours: to make WUR more visible as a university that has AI on its agenda. Right now, we’re not very visible as such just yet, but I think we’re the ideal place for advancing AI to explore the potential of nature and to improve the quality of life, as WUR has it all on one campus. That’s a unique advantage that allows for an integrative approach.”

What does the teaching component of your appointment entail?

“I teach courses on Deep Learning and Machine Learning, and I’m preparing a MOOC [Massive Open Online Course, ed.] on big data. But more so than passing on programming knowledge, I want to teach students how to think like computer scientists, so they are able to manage data and concepts that are a bit more abstract. Programming languages change so fast; if students learn how to think in abstract terms, how algorithms work, they’ll be able to cope with such challenges. That’s the essence of data science.”

So what can we expect from AI in five years’ time?

“That’s a tough question, because five years ago, AI was so different. It evolves so quickly that making a prediction for five years ahead is like a century later for AI. I don’t know what we’ll be able to do, but I’d like to have integrated an AI toolkit in Wageningen, and to have made the Wageningen agenda more central to the AI community. And in five years we should also have several examples of solutions that we’ve co-developed with AI and domain scientists alike.”

“I’ve also been looking back, reflecting on technological advancement a lot lately. What if Covid had happened before the Internet? Imagine it’s the eighties or the nineties and we’re in lockdown – how would we keep up with life? In 2020 we could facilitate that, but just imagine if... It’s hard to envision what that would have meant for all different aspects of human life, from personal communication to the economy to education and research.”

Does this digital era also present new challenges?

“Certainly. The biggest challenge is how we apply technology responsibly. In the past ten to twenty years, we’ve been flooded with data from sensors, phones, new devices, and the like. How do you manage the data they produce? And at what cost? There’s a lot of discussion about the environmental footprint of information technology; just think of ordering groceries online and having them delivered at your front door. The digital era isn’t for free and we live in very interesting times.”

“I experienced an example of what I consider to be the negative impact of technology as a tool to connect just last summer. I visited a very special spot with a palm tree forest next to a beach in Greece – a really unique place. It was full of people, but none of them saw the beauty of it: they were all Instagramming and broadcasting. I found that quite shocking, to be honest.”

“In fact, my top tip for when you have time off is to get a very old phone, to actually use it as a phone, and to make notes with old school pen and paper. Call your friends to make plans and leave social media be. The digital world can be overwhelming, as it’s used for everything: work, entertainment, staying in touch... I work with technology all day, every day; that’s why I make sure I disconnect from technology in my free time so I can connect with people and nature.”