Saved in:
Bibliographic Details
Main Authors: Takahashi, Tatsuki, Maru, Chihiro, Shoji, Hiroko
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.00446
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913868766248960
author Takahashi, Tatsuki
Maru, Chihiro
Shoji, Hiroko
author_facet Takahashi, Tatsuki
Maru, Chihiro
Shoji, Hiroko
contents Off-policy evaluation (OPE) in ranking settings with large ranking action spaces, which stems from an increase in both the number of unique actions and length of the ranking, is essential for assessing new recommender policies using only logged bandit data from previous versions. To address the high variance issues associated with existing estimators, we introduce two new assumptions: no direct effect on rankings and user behavior model on ranking embedding spaces. We then propose the generalized marginalized inverse propensity score (GMIPS) estimator with statistically desirable properties compared to existing ones. Finally, we demonstrate that the GMIPS achieves the lowest MSE. Notably, among GMIPS variants, the marginalized reward interaction IPS (MRIPS) incorporates a doubly marginalized importance weight based on a cascade behavior assumption on ranking embeddings. MRIPS effectively balances the trade-off between bias and variance, even as the ranking action spaces increase and the above assumptions may not hold, as evidenced by our experiments.
format Preprint
id arxiv_https___arxiv_org_abs_2506_00446
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Off-Policy Evaluation of Ranking Policies via Embedding-Space User Behavior Modeling
Takahashi, Tatsuki
Maru, Chihiro
Shoji, Hiroko
Machine Learning
Off-policy evaluation (OPE) in ranking settings with large ranking action spaces, which stems from an increase in both the number of unique actions and length of the ranking, is essential for assessing new recommender policies using only logged bandit data from previous versions. To address the high variance issues associated with existing estimators, we introduce two new assumptions: no direct effect on rankings and user behavior model on ranking embedding spaces. We then propose the generalized marginalized inverse propensity score (GMIPS) estimator with statistically desirable properties compared to existing ones. Finally, we demonstrate that the GMIPS achieves the lowest MSE. Notably, among GMIPS variants, the marginalized reward interaction IPS (MRIPS) incorporates a doubly marginalized importance weight based on a cascade behavior assumption on ranking embeddings. MRIPS effectively balances the trade-off between bias and variance, even as the ranking action spaces increase and the above assumptions may not hold, as evidenced by our experiments.
title Off-Policy Evaluation of Ranking Policies via Embedding-Space User Behavior Modeling
topic Machine Learning
url https://arxiv.org/abs/2506.00446