the
∞
.ai
Search
/
Concepts
trail ·
0 nodes
Every concept
Recommenders
8 concepts, 8 reviewed and merged.
Evaluation
Offline-Online Gap
A change that improves every offline metric can lose an A/B test, and this happens often enough that mature teams…
verified
Foundations
Collaborative Filtering
Recommend to you what people like you liked.
verified
Implicit Feedback
Nobody rates things, but everybody clicks.
verified
Methods
Matrix Factorization
Give every user and every item a short vector, and predict an interaction by their dot product.
verified
Regimes
Cold Start
A new user has no history and a new item has no audience, so the method that works by finding similar behaviour has…
verified
Feedback Loop
A recommender trains on what people clicked, and people click on what it recommended.
verified
Systems
Candidate Generation
You cannot score ten million items for every request.
verified
Two-Tower Retrieval
Encode the user with one network and the item with another, and let the score be a dot product.
verified