Nathan van de Wouw, TU/e
Great opportunities lie in the synergy between high-tech systems and AI
As announced by Katja Pahnke and Maarten Steinbuch on 15 December 2020 the TU/e High Tech Systems Center (HTSC) is being integrated into the recently established TU/e Eindhoven Artificial Intelligence Systems Institute (EAISI). New opportunities are emerging for HTSC’s AI focus – Robotics, Digital Engineering, AI for Engineering, Internet of Things and AgriFoodTech – while the center continues its activities under the current name. Nathan van de Wouw, full professor in the TU/e Dynamics and Control group, discusses the benefits of this merger and how the divide between ‘AI’ and ‘non-AI’ may be smaller than we think.
The next step in decision-making
“Up to now,” Nathan begins, “AI techniques have largely been used for advisory systems such as the algorithms of Netflix and Spotify or for speech and image recognition. In these cases, you use the data to do an analysis and then hand over the information to a human. But in automation – my field – we want to know how we can use developments in AI and data-driven science to also make a step towards more autonomous decision-making.”
Nathan’s home at TU/e is the Department of Mechanical Engineering, which he represents as a member of EAISI’s Scientific Board. Alongside a team of scientists, fellows and program managers, he assists the institute in setting up a vision on research (and associated education), scouting for talent and securing a good balance in program management. With the current merger, Nathan now feels a responsibility for supporting the smooth integration of HTSC within EAISI.
“Mechanical Engineering has had strong involvement in HTSC and high-tech systems is one of the key application domains within EAISI,” he explains. “As this is also important to the Brainport area, it’s essential that the connection between developments in AI and high-tech systems be shaped as successfully as possible. Keeping them apart would mean missing out on an opportunity. It makes sense to integrate the two and ensure that we can achieve the most at this interface.”
AI versus non-AI
In a nutshell, EAISI’s research philosophy is guided by three pillars: engineering systems, data & algorithms and humans & ethics. In order to achieve AI for the real world – in this case, the application domains of high-tech systems, health and mobility – there must be a recognition that no one element exists in a vacuum.
“There’s a lot of things here that people would call ‘non-AI’,” says Nathan, “but even the things that might not directly be AI are important to this overall picture. The ultimate innovations will be achieved by bringing together things from these three pillars in a multidisciplinary manner. What is AI and what is not AI becomes less important. The engineering side of HTSC – the design of high-tech systems, the systems engineering – will remain important under the banner of EAISI.”
"It’s essential that the connection between developments in AI and high-tech systems be shaped as successfully as possible."
By bringing together HTSC’s high-tech focus and all of TU/e’s AI activities, the institute will build on a large number of successes already achieved at the university. These range from high-tech motion stages with more accurate performance to tremendous breakthroughs in AI for healthcare diagnostics. EAISI also inherits HTSC’s longstanding approach of bringing together disciplines to maximize results, such as in the setting up of an exploratory, multidisciplinary AI research program.
Nathan: “If we want to really cash in on the promise of our three domains, we have to cross boundaries much more. I see interest in the development of AI from the whole breadth of Mechanical Engineering and TU/e, and I think that the biggest potential – but also challenge – for EAISI is further stimulating multi-disciplinary collaboration. That’s where the biggest gain is, not just for the scientific side but also for the industrial valorization side. This is at the core of HTSC’s vision, which can bring a lot to EAISI.”
Plenty of selling points
As stated in last year’s unveiling, EAISI aims to take on fifty new professors in the next four years alone. However, Nathan again stresses that there’s more to this than just AI expertise. “These should be people who are also very excited about the engineering domain – the next generation of semiconductor machines and so on – and the human aspect. The fact that we’re so strong in high-tech systems in this region uniquely enables us to attract people who are interested in this. There are so many opportunities to connect academia to industry in Eindhoven, so EAISI is more than just saying to talented people that they can come work on AI. It’s also letting them do that in the exciting Brainport region, which is definitely a selling point.”
"EAISI also inherits HTSC’s longstanding approach of bringing together disciplines to maximize results."
For students as well, opportunities are arising through the integration of high-tech systems and AI, such as TU/e’s upcoming Master’s Program in Artificial Intelligence Engineering Systems. This seeks to readdress an imbalance in industry: the fact that many well-trained engineers are inexperienced in AI while many Data Science students are less knowledgeable about engineering systems.
“Graduates that understand both are needed to push this development forward,” notes Nathan. “There will be many opportunities for students to set up a program related to high-tech systems, health or mobility. Students who are interested in AI should find it very exciting.”
Shooting for the moon
As for the future, EAISI has already identified a number of ‘moonshots’ through which AI could be a gamechanger for societal and industrial challenges. Examples include personalized health support outside of hospitals and zero waste from factories, but also a serious look at both the strengths and weaknesses of human-machine collaboration. “Think of an autonomous car,” says Nathan. “One aspect related to ethics is that if you make processes more autonomous, there’s a transfer of responsibility. How do we guarantee safety? As humans, we’re flexible with rules and make judgements on when it’s ethical to break them. But this question is difficult for AI.”
“Some people say that we’ll automate everything. Honestly, I don’t believe in this. We have to carefully think about where it helps humanity as a whole and where we should leave the decision-making to people. AI is not a goal in itself; we do it so we can have less accidents, greener mobility, more sustainable industry, improved healthcare and so on. In my opinion, it should serve a higher purpose.”