Reproduce “Generative Query Suggestion via Intent Coverage and Query-Level Credit Assignment” 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
Xinpeng Liu, Lu Ma, Jiayi Qiao, Mengyu Zhou, Linglong Li, Xiaofeng Bian, Haonan Chen, Xiaoxi Jiang and 1 more
The authors train a query-suggestion system to cover distinct intents rather than vary wording. It combines intent-aware supervised data with reinforcement learning, assigning quality signals to individual queries while sharing intent coverage across each slate. In a one-week A/B test, it improved relative CTR by 3.25% versus GRPO and human-evaluated GSB by 0.07, with evidence limited to this experiment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.