Or the standard given a bunch of tags (from the instance segmentation) and find your ranked list of preferences. https://librecommender.readthedocs.io/en/latest/ has a list of 5-10 standard recommender algorithms.
Or use autogluon https://auto.gluon.ai/stable/tutorials/tabular/tabular-multi... where they take Meta's large repository of annotated-images (DINOv2) and use them to classify then do the recommendation system on tabular (database tables).
This stuff has been standard for years now.
duxup•3mo ago
How would this work? It identifies just objects / places that imply interests?
Sounds like a blackbox that assumes I want to ski ... because some pics look "like" skiing or something. Or tries to say connect someone with their ex or people who look like their ex ...
Granted I suspect this is just AI for their sake / identifying faces and collecting mass data ...
iFire•3mo ago
Kinda. But collaborative filtering by the textbook definition tries to find the people who have those embedding and then match those nodes with other similar nodes.
So because you are a person who likes skiing, the system tries to find other people's existing preferences that likes skiing and match those to the items to show you.
https://en.wikipedia.org/wiki/Collaborative_filtering