This is really cool. You might be interested in this talk which shows you can incrementally add some preference vectors as a way to improve the recommendation.

This is really cool. You might be interested in this talk which shows you can incrementally add some preference vectors as a way to improve the recommendation. https://haystackconf.com/us2023/talk-20/ https://haystackconf.com/us2023/talk-20/