The aim of the TAPI (Towards Autonomy in Process Industries) project is to push Norwegian land-based process industries towards more autonomous operations by exploring the intersection between machine learning (ML) and more traditional model-based control methods. The outcome has been development of techniques to learn and control systems with small amounts of data, taking uncertainty into account, and testing those methods on the real systems of the partners in the process industry. We are seeking funding for a PhD project within physics-informed ML, with applications to sensor data. In addition to TAPI, the PhD student will be connected to several other ongoing projects linked to this research area.
We aim to develop hybrid frameworks that work with interchangeable sensors, thus allowing to perform transfer learning between different sensor inputs. We call this multi-modal learning. The application areas we wish to target are all safety-critical systems that are governed by known laws of physics, but which are too complex to model from first principles, thus partially requiring data-driven methods. If those systems have different measurement options, and a trade-off between sensor quality and cost must be made, we see multi-modal hybrid machine learning as the desired method to solve those challenges. We believe the results of this project will be of interest to a broad group of customers in many sectors. Within the project period we wish to aim our communication towards the Norwegian process industry and the defence sector, utilising the expertise of the new and existing partners.