An ontology modeling classes and relationships describing deep learning networks, their component layers and activation functions, machine learning methods, as well as AI/ML potential biases.
we previously curated 40 mappings between AIO and ITO in Biomappings (summary at https://biopragmatics.github.io/biomappings), and there are for sure more that are possible. The Biomappings SSSOM file can be consumed by the ODK to incorporate these mappings directly into your ontology artifact
I just skimmed https://arxiv.org/abs/2404.03044 and have two suggestions: