View all blog posts

Behind the Scenes at GENIUS Lab

Earlier this year, Kickstart AI, together with TU Delft, Maastricht University, DSM-Firmenich we launched a new ICAI lab: GENIUS, which stands for Generative Enhanced Next-Generation Intelligent Understanding Systems lab. In this blog I am thrilled to share what efforts went into this collaborative effort in shaping the future of AI:
May 6, 2024

At Kickstart AI, we believe the future is positively shaped by AI, and so our enthusiasm for the GENIUS lab stems from its dedication to advancing generative AI methods. This new research lab aims for GenAI to become not only more useful for human experts but also more understandable and trustworthy, adding to impactful AI solutions for a better future.

Valley of Death

Everyone who has worked in R&D knows the infamous Valley of Death, the place where our awesome ideas are left to die due to lack of funding, tangible results and sometimes an incomplete business case. At Kickstart AI it is our aim to keep our research and ideas out of the Valley of Death. We do this by collaborating with businesses and society to develop practical AI solutions that address their most pressing challenges. That’s why we are so happy that the idea behind GENIUS saw the light of day and that we are partnering with some of the best knowledge institutions and organisations to make it happen.

To give you a bit of a behind the scenes I will share what I was able to learn when speaking to some  of the bright minds behind a new AI lab at the official GENIUS launch back in February. First off, I spoke to Geert-Jan Houben, Scientific Director from TU Delft, who played an important role in the creation of the GENIUS lab. Houben, together with Professor Michel Dumontier, is lab director of the Genius lab. He leads the activities of the TU Delft in the field of AI.

"The context and cases these students are going to work on as part of the GENIUS lab are not only useful but also very interesting. What excites me is that we can look at what is needed from different perspectives and work on it together. We have gathered scientists from various knowledge institutions here. Together with Kickstart AI and dsm-firmenich we want to explore in what kind of settings we can implement findings and what is needed to do it well— to truly apply it. Because it's more than just technology; knowledge is needed, talent is needed, research is needed. This is how you can make innovation happen."

One of those PhD students is Anne Arzberger, a PhD candidate for AI value alignment at TU Delft. I spoke to her about her motivation for joining the lab and her experiences collaborating with the GENIUS lab team up until today.

“I have a background in Design, so human-centeredness is something I have been trained on and know by heart. I always want to find ways in which technology can serve human needs, and I think AI has a huge potential to do that if we go about it the right way. We have people from different backgrounds, with different practices and with different knowledge involved, and because of the transdisciplinary nature of this project, we have the chance of seeing things from different angles. The very openness I have experienced so far towards different ways of knowing and gathering knowledge is a very important characteristic to have for such a project as the lab.”

I acknowledge that the lab is in its early stages when I ask Anne about hopes for her research at GENIUS. It is clear that the lab is working towards something with tangible impact as we talk about her goals for this project.  

“Ultimately, my hope for this project is that I can contribute something to ethical AI, and we, as colleagues with different backgrounds, can learn from each other’s knowledge and way of working.”

Next up, I had the privilege of speaking to Dr. Jie Yang Assistant Professor, Lab manager for GENIUS lab, with whom I talked about the transdisciplinary collaboration in our conversations as he tells me AI and the relationship between humans and AI have always been the core focus of his research.

"When Chat-GPT launched, AI became well known, and everybody started talking about it. But generative AI or large-scale models are actually not that new. For example, trustworthiness, an important principle of the EU’s approach to AI, is a large part of the current conversation around AI. Many, including me, have been working on the reliability and robustness of AI long before this conversation. But from my perspective, this attention that AI has gained adds to opportunities for us to work on societally impactful use cases as the use of AI becomes more popular.”

Our conversation moves towards recent years development and popularisation of large language models and he tells me about this switch in the AI space and how this has been an exciting time to work in artificial intelligence.

“It is exciting to be at this point where AI is making a real impact. This sounds obvious, but for technologies, it can take a huge amount of effort and many years for everything to come together like things are coming together right now around AI. Now that AI is at a level where it is widely used, it becomes more important to understand the potential implications of the usage of these technologies; it can mean very different things across different contexts. That is also why I believe that we cannot bring forward technology by working alone; we need to bring in domain experts from different contexts and explore how we can really impact their daily work.”

Jie Yang is the lab manager, responsible for the lab on the TU Delft side, but also responsible for administrative management, project coordination and planning. Besides that Yang is work package leader and PhD advisor in WP3: Integration of structured knowledge in generative AI models. In this work package, the researchers try to improve the reliability of neurosymbolic AI by adding symbolic knowledge. I ask him what he, as the lab manager, envisions for the future of GENIUS lab.

“I would like to look back on this lab and see that we actually find that there are people who can benefit from the things we’ve created in the lab. This is why I am pleased that we have people from different fields and organizations for this project; that, to me, is essential for creating something of real value."

Reflecting on my conversations with the team at GENIUS lab, it is clear that the lab stands as a testament to a shared commitment to advancing AI responsibly and creating something of tangible impact. Simultaneously, it promotes transdisciplinary collaboration that goes beyond technological boundaries. But what excites me most about the talks I had with the team was everyone’s enthusiasm for the lab and working with their collaborators. I look forward to the future of GENIUS lab and will take you, the reader, along for future updates on our blog and LinkedIn.

By Lesley-Ann de Wit, Quality Assurance Manager and occasional writer at Kickstart AI.

Receive Our Newsletter
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form. Please try again.