Our research
SURE-AI balances fundamental research with industry-driven projects
The centre fosters interdisciplinary collaboration to prepare Norway with essential AI competencies.
By improving the fundamental technology behind AI, our aim is to advance current and emerging AI technologies, aligned with national priorities for creating economic and societal value.
Through dedicated innovation clusters, we ensure that the research outcomes are not only publishable, but deployable, transforming mathematical theory into applied solutions with our industry partners from energy, finance and public services.
Together, we co-develop trustworthy, efficient, robust and impactful AI systems that meet real world needs. To expand our reach and curiosity, we welcome new industry partners to contact us.
Sustainable
Sustainable AI means building and using AI in a way that’s good for people and the planet over the long run. It keeps energy, water, and hardware use low and is affordable to run and maintain. In practice, that means building efficient models with a low environmental impact. This is a central topic for the SURE AI research center.
Risk-Averse
Risk-averse AI means building systems that avoid unnecessary risk and harm, and that are robust to handle challenges and scale without breaking. They know when they’re unsure, choose the safer option, and ask a human when needed. We test them thoroughly, set clear limits, and add backups so they fail safely and keep mistakes small.
Ethical
Ethical AI means building human values right into the algorithms, so they’re ethical from the inside out. Ethical AI means building systems that treat people fairly and with respect. They protect privacy, use data responsibly, and avoid bias, with clear rules about what the AI can and cannot do. We keep choices transparent, explain decisions in plain language, and make sure people—not machines—stay accountable and in control.
Toward trustworthy AI: recovering reliable decisions from corrupted data
Risk-averse ensemble control: why "average" performance is no longer good enough
The Geometry of Data - Why Machine Learning Needs Signatures
SURE-AI kicks off: “of vital importance to our future”
Beyond the Monolith: how brain-inspired AI is building a more reliable and sustainable future
Signals with shape: why topology matters for modern data?
Why horizon matters more than you think in decision making
The economics of overlapping generations: a stochastic lens
Unifying Risk and Belief: A Foundation for Coherent AI Decision-Making
Research outputs
Toward trustworthy AI: recovering reliable decisions from corrupted data
Risk-averse ensemble control: why "average" performance is no longer good enough
The Geometry of Data - Why Machine Learning Needs Signatures
Beyond the Monolith: how brain-inspired AI is building a more reliable and sustainable future
Signals with shape: why topology matters for modern data?
Why horizon matters more than you think in decision making
The economics of overlapping generations: a stochastic lens