A premature baby is at risk for complications and could end up in the NICU, an intensive care unit for babies. These NICUs are becoming more and more innovative, as we can see, for example, with the new unit at Radboudumc. A major risk that premature babies run in the NICU is blood poisoning and that can have serious consequences.
Sepsis can lead to tissue or organ damage and even death. Specialists from UMC Utrecht therefore want to predict as accurately as possible how high the risk of a baby actually is. By identifying the very high-risk babies, doctors can take measures as early as possible to prevent health risks or even death. This risk assessment process can be further improved with better AI, which was recently worked on by six students in the Fruit Punch Challenge.
Six students challenged
A machine learning model is already being used at UMC Utrecht to map the risk babies as much as possible. With the fruit punch challenge, six students were challenged to further refine this AI application. In the end, two groups each took a slightly different path with their ideas. The creative thinking process yielded concrete new insights that can be used to improve the algorithm of the existing machine learning prediction model.
AI & fruit punch challenges
Concrete results were a refinement of the existing logistic regression model, a new XGBoost model and an elaboration with SHAP for XAI (Explainable Artificial Intelligence). According to those involved, this FruitPunch AI Challenge certainly leaves you wanting more. The FruitPunch AI develops challenges for students to teach them about AI, based on their own experience and while developing their own skills. Together with Utrecht University, Eindhoven University of Technology and Wageningen University, UMC Utrecht stimulates the use of AI by supporting and guiding these challenges.
Cross-pollination in AI
In any case, the six students found it valuable to get started with clinical data and in turn inspired the professionals involved. Utrecht UMC is calling on students but also data scientists to take part in such a challenge, because cross-pollination is particularly useful in a field such as AI. Artificial Intelligence (AI) offers the potential to come up with solutions for the world’s biggest (healthcare) problems, but it is currently underused because there are simply too few people how to apply AI.
.