AI4B.io is a collaboration between TU Delft and DSM that focuses on improving biotechnology by applying novel artificial intelligence technologies. The collaboration aims to enhance the effectiveness and efficiency of research and business processes in the biotech industry on different scales. Novel machine learning techniques are developed and applied in real-world biomanufacturing settings.
The researcher projects covered by AI4b.io are diverse. Researchers look at the coordination between bioreactors and other processing units using machine learning algorithms for their scheduling to shorten their lead time. Another venture is to look at the spatial mass flows and processes in a bioreactor in interaction with the microbes that perform the actual bioconversions using physics-informed learning to predict and set continuity of what comes out of the reactor under which conditions. On a different scale the researchers are investigating the biochemical pathways within microbes using a similar scheduling approach to predict how the pathways change when targeted modifications in the genome of the microbes are made. Moreover, the team works on developing a self-learning system where a robot can make its own interventions to get the desired end-product in an iterative autonomous process. Here, the researchers combine self-learning algorithms with experimental data and AI-guided experimental designs. In a fifth project they are working on using AI for metagenomics checking out gut bacterial populations to make predictions on the influence of food towards health benefits.

With this initiative TU Delft and DSM want to boost the use of AI within the biosciences setting to solve some of the big challenges the world is facing. We intend to become THE center in the world adopting such an approach. This can and should not be done alone. Therefore, we will extend and expand our collaboration at a larger scale including other bioscience companies and universities to become a world-wide hub.