Joakim Löfgren from Aalto University is applying AI machine learning capabilities to data produced in lab experiments to find the optimal process parameters and drastically speed up research and development times.
Coupling lab experiments with machine learning capabilities is a novel area of application for AI that can provide answers to some of the common challenges in the lab such as sparse amounts of data, long testing times and unstable process conditions.
Joakim utilizes a machine learning technique called Bayesian optimization that works very well with limited data amounts. This research can have a great impact with scientists in the lab as the AI model is capable of active learning, as not only able to provide solutions based on given data but also to suggest how to run new experiments to generate useful data.
The team has so far been working successfully with lignin process optimization to directly map the processing conditions for the maximum amount of lignin with specific properties.
