SURE-AI
The Centre for Sustainable, Risk-averse and Ethical AI (SURE-AI) believes that AI technologies can be made more sustainable, ethical, and risk-averse by advancing the mathematical foundations underlying the models, training, and implementation.
Lower row: Lena Korsnes (Simula), Giulia Di Nunno (UiO), Baltasar Beferull-Lozano (Simula), Asbjørn Torvanger (CICERO), Hans Munthe-Kaas (UiT), Sven-Arne Reinemo (Simula)
SURE-AI is one of six new national centres for research in AI, composed of more than 60 percent mathematicians.
When conceiving the centre, we asked ourselves "what can our particular team of scientists here in Norway do to address the issues facing AI today and in the future?"
Thomas Surowiec, Deputy Centre Director
This resulted in three focus areas:
- AI has excessive resource consumption: The excessive use of resources is due to the massive number of parameters in the models that need to be tuned for the most successful AI technologies. We seek to find smaller, more agile models and better techniques to do the training.
- AI has a transparency problem: This is often due to a lack of understanding of the fundamental principles. This often requires deep mathematical analysis to provide things like performance guarantees. Or simply making it clear to people that there are mathematical limitations to what a given technology can actually tell us.
- AI poses various types of risk due to unpredictable behavior: Many models tune according to the mean, ignoring the overall risks involved in deviations from the mean and the inherent randomness in the data and training process itself. We seek to bridge tools of risk management into the training process to reduce risk and align the technologies better with human values.
SURE-AI leadership
Centre Director
Deputy Centre Director
Director of Research
Director of Administration and Operations
Director of Innovation
If you have any questions or would like to get involved, please contact us at: contact@sure-ai.no