One by one the participants arrive, from Dutch Universities, including Groningen, Eindhoven, and Amsterdam, but also as far away as Paris. They are pursuing MSc or PhD degrees in various fields, including Mathematics, Physics, and Earth-Sciences. What unites them is their interest in novel data-driven techniques for inverse problems. At the 2-day masterclass, organized as part of the CWI Semester Program on Data-driven methods for Inverse Problems, they came to the right place.
Inverse problems occur in many applications, including medicine, materials science, and geosciences. The basic task in all these applications is to fit a mathematical model to data, in order to learn something about the system of investigation. Where traditionally these mathematical models where hand-crafted based on the underlying physics, the latest development is the use of abundant data to inform model-building directly using machine learning. In the 2-day Masterclass, 5 speakers presented the state-of-the-art in a tutorial fashion, including hands-on examples where participants could work with small data-sets. The topics included Bayesian statistics, machine learning for computational finance, fluid dynamics, and image processing.
We look back at a very successful event, with insightful lectures, interesting discussions and interactive demos. As one participant put it:
“The variety in inverse problem applications with machine learning was great, it gave it a more broad perspective. The speakers were all great and because we were with a small group it was easy to approach the speaker after to ask further questions, or talk to them during the drinks. The day was organized very well and all information we needed was sent ahead of time so there were no miscommunications or unclear things.”
We are looking forward to hosting more masterclasses of this kind. To stay updated on events like these, you can subscribe to the mailing list.
