Back to search

INTPART-International Partnerships for Excellent Education and Research

Bio-inspired Neural Networks for Artificial Intelligence systems: A US-Norway collaboration converging Neuroscience and AI

Alternative title: Bioinspirerte nevrale nettverk for kunstig intelligenssystemer (KI): Norsk-amerikansk samarbeid for å konvergere nevrovitenskap og KI

Awarded: NOK 4.5 mill.

We will establish an activity combining extensive knowledge of the brain with computational modeling of artificial neural networks (ANN) to develop novel network architectures and perform comparable laboratory and simulation experiments in humans, rodents and machines. This will allow testing performance and comparability and generate novel predictions to drive developments in AI and bring deep insight to brain function. We build on our long-term collaboration between neuroscience and physics-based computational modeling to solve long standing questions in neuroscience and to understand and develop AI systems. We are also in lead integrating computational skills into science education. However, to reach our ambition to release the potential in the coupling between neuroscience and AI research, we need to combine our competence with complementary, world-leading competence currently lacking in Norway. This is provided by carefully selected partners who bring access to: unique hardware and neural recordings, expertise in human cognitive tasks, large-scale network simulations and expertise in statistical tools to compare brain and ANNs. Success requires long-term interdisciplinary investments, a common, ambitious research strategy, and a collective push to educate the workforce of tomorrow in these fields, which can only be achieved through a long-term joint commitment. We therefore go beyond individual international collaborations and develop long-term strategic partnerships on both research and education and will: (i) train students with the combined competences of the partners needed for breakthroughs in AI-neuroscience interactions, and (ii) strengthen our position as an internationally leading research center. Due to the Covid-19 pandemic and the lack of travelling opportunities for students and researchers, the planned activities of the project are delayed. Nevertheless, we were part of organizing an international summer school for PhD students in 2021 and organized a meeting with several of the partners in April 2022. In 2023 we have extended the partnership to also include Mila, the Quebec Artificial Intelligence Institute. Mila is recognized worldwide for its major contributions to AI. Today, the Mila community boasts the largest concentration of deep learning academic researchers globally. The collaborations were initiated by one of our researchers spending the fall of 2023 at Mila. This forms the start of series of collaborations and resprocal researcher exchanges between the University of Oslo and Mila in 2024 and 2025. In 2023 we have together with Nora AI and Simula research laboratories taken an initiative to a national working group in NeuroAI and developed a white paper describing the foundation for collaborations between Norwegian and US research institutions focusing on AI. We have planned a conference in NeuroAI in Norway with several internationally leading researchers in 2024 in a collaboration between UiO, Simula and NTNU’s groups in AI and neuroscience.

With MAIBRAIN, we will establish an activity combining extensive knowledge of the brain with computational modeling of artificial neural networks (ANN) to develop novel network architectures and perform comparable laboratory and simulation experiments in humans, rodents and machines. This will allow testing performance and comparability and generate novel predictions to drive developments in AI and bring deep insight to brain function. This may change how neuroscience research is conducted and solve current limitations within AI. With the Center for Integrative Neuroplasticity (CINPLA, UiO), we have developed an internationally leading interdisciplinary research activity. We build on established expertise in neuroscience and physics-based computational modeling to solve long standing questions in neuroscience. We are also in lead integrating computational skills into science education. However, to reach our ambition, become an international powerhouse and realize the potential in the coupling between neuroscience and AI research, we need to combine our competence with complementary, world-leading competence currently lacking in Norway. This is provided by carefully selected partners who bring access to: unique hardware and neural recordings, expertise in human cognitive tasks, large-scale network simulations and expertise in statistical tools to compare brain and ANNs. Success requires long-term interdisciplinary investments, a common, ambitious research strategy, and a collective push to educate the workforce of tomorrow in these fields, which can only be achieved through a long-term joint commitment. We must therefore go beyond individual international collaborations and develop long-term strategic partnerships on both research and education to: (i) train students with the combined competences of the partners needed for breakthroughs in AI-neuroscience interactions, and (ii) strengthen our position as an internationally leading research center.

Funding scheme:

INTPART-International Partnerships for Excellent Education and Research