Reproduce “Adaptive Doubly Robust Off-Policy Evaluation for Ranking Policies under Diverse User Behavior” 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
Kosuke Iguchi, Ren Kishimoto
ADR estimates candidate ranking-policy value from existing logs by combining a reward prediction with an adaptively weighted correction for its error. In synthetic simulations, at ranking length 14 with 8,000 logged samples, it reduced MSE by 37.2% versus AIPS, from 10.581 to 6.645. The finding is limited to synthetic data; real-log validation remains needed.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.