A house for the Centre's machine learning and AI tools.
Machine learning models, analysis pipelines, AI agents and the scripts that go with them, built by members and kept in one place. Once a deposit has been reviewed, a member who is granted access either takes the tool away and runs it on their own machine, or attaches it to their own assistant and uses it where it stands. CISBM houses the tools; it does not sell models, inference or credits.
How the house works
What it holds
- Machine learning models, analysis pipelines and agents that members have built, reviewed before deposit and catalogued by method.
- Two ways to use one: take it away and run it on your own machine, or attach it to your own Claude or GPT and use it where it stands.
- Access given per member, per tool, on the terms the depositor set.
- An enclosure: a depositor may let their tool be used while keeping the source unread.
Where the boundaries fall
- Model access, inference and credits stay with the member, under their own account with their own provider.
- Compute costs sit wherever the tool is run, and each depositor states what a tool needs.
- An attached tool answers to the assistant you already use, and you connect it yourself.
- A deposit is software with a licence; findings and manuscripts are handled by the Journal, and progression levels by the councils.
Catalogue
What is in the house, and how each tool may be used. Signed-in members may ask for access; depositing a tool of your own is done in your portal.
Loading…