SINTEF's Vision and Opportunities
An overview of SINTEF and the exciting opportunities they offer for summer internships and master's thesis.
Physics informed machine learning
Physics-informed machine learning consists, as the name implies, of combining machine learning methods with physics-based modelling, in order to better utilize the structures that lie within data. Pseudo-Hamiltonian Neural Networks (PHNN) do this through learning conservative and dissipative functions of partial differential equations (PDEs) from data, with the key benefit of not having to inform the system of the explicit equation.
This approach allows for more robust and trustworthy predictions for complex physical systems.
Data Informed Assessment of Rock Quality for the Circular Economy
In the industry project "Circular Management of Soil, Rocks, and Aggregates," we analysed geological data related to rock quality. One of the goals was to assist our customer, Feiring Pukkverk, in making more informed and efficient decisions when accepting stone from external sources.
In this presentation, we will provide insight into how basic statistical methods, data visualization techniques, and regression models were utilized to deliver value to the customer.
As usual, pizza and refreshments will be served in advance.