Build out the classes for embeddings

berkeleybop / artificial-intelligence-ontology

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.

https://berkeleybop.github.io/artificial-intelligence-ontology/

16 stars 3 forks source link

Build out the classes for embeddings #88

Open caufieldjh opened 3 months ago

caufieldjh commented 3 months ago

re: discussion on Jun 10 2024 with @realmarcin @turbomam @cmungall

Classes like AIO:word2vec are somewhat isolated as children of AIO:MachineLearning, and incompletely modeled.

Would like to see:

More terms for embeddings and vectorization approaches
Modeling of these in a way that includes inputs and outputs rather than hierarchy alone