Reproduce “JevRec: Linguistic Encoding with Jev-like Decision Modeling for LLM-based Generative Recommendation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Qiyong Zhong, Jiajie Su, Ming Yang, Li Zhang
JevRec assigns semantic ID lengths by training interaction frequency, giving popular items shorter codes and rare items longer ones. It predicts the target length separately from the length-conditioned item-code content. On Amazon Reviews 2023’s Video Games category, it reaches NDCG@10 of 0.0514, versus 0.0490 for ActionPiece, under leave-one-out full-catalog evaluation; results cover three tested categories.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.