Reproduce “LLM-based User-Product Relevance Evaluation for Dynamic Product Advertising” 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
Wen Zhang, Xiao Bai, Hsiang-wei Chao, Mengfan Zhu, Tong Zhao, Mingxuan Ju, Xiaodong Liu, Congfei Zhang and 4 more
The authors propose an offline method to assess retrieval for advertising by comparing products with interests inferred from engagement histories. Semantic IDs and event types group histories into weighted intents, which a language model judge compares with candidates. The weighted method uses 3.4% of pairwise calls and retains 0.989 rank correlation, with evidence bounded to sources and experiments.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.