Open utterances-bot opened 3 years ago
A whirlwind tour of bandits, embedding+MLP, sequences, graph, and user embeddings.
https://eugeneyan.com/writing/patterns-for-personalization/
Thanks for the posts. I'm looking forward if you can share the pros/cons of these approach in term of inference latency, how frequent we need to retrain the embedding etc.
awesome overview!
Patterns for Personalization in Recommendations and Search
A whirlwind tour of bandits, embedding+MLP, sequences, graph, and user embeddings.
https://eugeneyan.com/writing/patterns-for-personalization/