CytoNet, a foundation model for neural microarchitecture analysis based on EBRAINS atlases has been presented during the International Supercomputing Conference in Hamburg this week as part of the Nvidia Executive Briefing.
Ian Buck, Vice President and General Manager of Hyperscale and High Performance Computing at NVIDIA, showed Cytonet as one of the first research projects carried out on the JUPITER Exascale computer in Germany.
CytoNet uses imaging data from different cellular scales available through the EBRAINS atlas services to model relationships between individual cell structures and higher-level patterns of brain organization and function. The foundation model was trained on JUPITER in less than five days, using 6.5 petabytes of data from 21 postmortem brains on 4,096 NVIDIA Grace Hopper Superchips.
The Cytonet project is led by EBRAINS Joint-CEO Katrin Amunts and Helmholtz AI research group leader Christian Schiffer, both from Forschungszentrum Jülich in Germany.
NVIDIA also highlights the project in a dedicated blog post and describes the team's next steps toward AI assistants that can help scientists interrogate brain data directly. “For the first time, we’re not just using AI to analyze the brain — we’re building an agent that can think through the experiment itself”, Katrin Amunts is quoted.
Further Information
Create an account
EBRAINS is open and free. Sign up now for complete access to our tools and services.