1 code implementation • 11 Dec 2023 • Jinseok Seol, Minseok Gang, Sang-goo Lee, Jaehui Park
Additionally, that the proxy embeddings are shared across all items allows the infrequent items to borrow training signals of frequent items in a unified model structure and end-to-end manner.
Ranked #1 on Recommendation Systems on Amazon Beauty (using extra training data)