Reproduce “Interactor: Agentic RL oriented Iterative Creation for Ad Description Generation in Sponsored Search” 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
Penghui Wei, Jiayu Wu, Chao Ye, Zhi Guo, Shuanglong Li, Lin Liu
Interactor generates search-ad descriptions by retrieving query-relevant knowledge, then revising drafts using rubric-based judgments with explanatory feedback on relevance and landing-page support. On a proprietary Chinese test set, its LLM-judged Average@10 faithfulness was 0.872 versus 0.726 for SINGLE-TURN RL, across ten candidates per input. An online A/B test reported a 0.74% revenue lift, but does not isolate iteration’s effect.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.