Analytics and Artificial Intelligence - Research Group

Our field of research

AI encompasses a wide array of specific methods. We specialise in AI for physical systems and adopt a multidisciplinary approach to find optimal AI solutions for the industry and the public sector. Our three main research areas are:

• Knowledge-informed AI: We develop models that combine domain knowledge with data-driven methods. This domain knowledge may stem from the expertise and insights of professionals and operators, or it may be mathematical models based on the underlying physics of the system.
• Machine learning and optimization: We leverage machine learning to continuously improve optimization algorithms from data, particularly when the system to be optimized change over time. This includes but is not limited to system control with reinforcement learning.
• Time series learning: Mainly on physical systems with sensor data, with applications in predictive maintenance, condition monitoring and anomaly detection.

Our expertise

The Analytics and AI group is composed of researchers with diverse backgrounds in physics, mathematics, statistics, cybernetics and informatics. This multi-disciplinary background enables us to combine aspects of AI and machine learning with domain knowledge for physical processes and industrial systems. The applications span various domains including process industry, logistics, construction, energy and many others.

How we work

Together with our partners, we take time to understand the problem we want to solve, and the relevance of various AI. We believe that the best AI solutions for complicated physical systems can only be found through a fundamental understanding of what the data represents. We stay updated on the latest advancements in AI, conduct fundamental algorithmic research, and develop machine learning models tailored to the specific challenges of our partners.

What sets us apart

Our researchers blend a profound understanding of AI and data science principles with practical application experience and a passion for solving real-world problems. When necessary, we collaborate with domain experts across SINTEF to leverage existing knowledge and domain expertise.

How can we help you?

• Struggling to model your physical system or detect anomalous behaviour? We offer deep expertise on modelling of complex physical systems by combining machine learning and domain knowledge whether for detecting anomalies, predicting future behaviour, control, or simply increasing understanding of the system.
• Want to optimize a dynamic system but lack the connection between data and optimizer? We develop data-driven models tailored for optimization and decision processes.
• Are your operators sceptic of AI models? We co-develop with end users to ensure relevance and trustworthiness.
• Do you struggle to get started with AI? We have long experience with developing successful AI cases directly with clients or in collaboration with the digital innovation hub Nemonoor.

International orientation

Our group collaborate with leading academic institutions both nationally (NTNU, UiO) and internationally. Specifically, we have strategic partnerships with Prof. Karniadakis at Brown University and Prof. Kutz at University of Washington.

Strategic Directions

While focusing on the three topics listed above, we follow the rapid development of AI technologies and integrate them in our daily workflows. This includes exploring language models as code-pilots or interfaces to physics models or new directions such as quantum machine learning and bio-inspired neuromorphic computing. 

Join Us

We are always open to new collaboration opportunities. Contact us to learn more about our research and how we can work together to solve complex scientific and technological challenges. You can also engage with our expertise through tailored courses, strategic consulting, or by spending time with us as a visiting researcher or collaborator.