Reproduce “Beyond Independent Predictions: Learning the Browsing Journey for Sequence-Level Re-ranking” 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
Jiayi Zhang, Jingxin Hai, Zhenjia Ma, Hongbin Pei, Dagui Chen, Han Zhu, Jian Xu
The authors introduce ICE, an evaluator for recommendation sequences that estimates browsing depth and item impressions before predicting clicks. It attenuates the influence of less likely seen items and models click allocation across a sequence, while keeping the generator fixed. In a 14-day Taobao feed A/B test, ICE increased clicks and commercial revenue; these results apply to that production setting.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.