Reproduce “Addressing Purchase-Impression Gap through a Sequential Re-ranker” 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
Shubhangi Tandon, Saratchandra Indrakanti, Amit Jaiswal, Svetlana Strunjas, Manojkumar Rangasamy Kannadasan
The authors present a sequential reranker for e-commerce search that selects each next item using its production relevance score and the aspect mix already selected. Historical query-specific purchase shares provide a target-mix proxy. Offline, at α = 0.5, it reduced the measured gap by 8% and shifted MRR 3.7%, but p = 0.5 and no live multi-aspect test is reported.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.