Off-policy evaluation for slate recommendation

Abstract

This paper studies the evaluation of policies that recommend an ordered set of items (e.g., a ranking) based on some context---a common scenario in web search, ads, and recommendation. We build on techniques from combinatorial bandits to introduce a new practical estimator that uses logged data to estimate a policy's performance. A thorough empirical evaluation on real-world data reveals that our estimator is accurate in a variety of settings, including as a subroutine in a learning-to-rank task, where it achieves competitive performance. We derive conditions under which our estimator is unbiased---these conditions are weaker than prior heuristics for slate evaluation---and experimentally demonstrate a smaller bias than parametric approaches, even when these conditions are violated. Finally, our theory and experiments also show exponential savings in the amount of required data compared with general unbiased estimators.

31 pages (9 main paper, 20 supplementary), 12 figures (2 main paper, 10 supplementary)

References in corpus (2)

• Doubly Robust Policy Evaluation and Learning
• Introducing LETOR 4.0 Datasets

Cited by in corpus (9)

• Counterfactual Evaluation of Slate Recommendations with Sequential Reward Interactions
• Doubly Robust Off-Policy Evaluation for Ranking Policies under the Cascade Behavior Model
• Off-Policy Evaluation of Ranking Policies under Diverse User Behavior
• Off-Policy Evaluation of Slate Bandit Policies via Optimizing Abstraction
• Exploiting Neighborhood Interference with Low Order Interactions under Unit Randomized Design
• Contextual Bandit Applications in Customer Support Bot
• Offline Evaluation of Ranking Policies with Click Models
• Learning Action Embeddings for Off-Policy Evaluation
• Non-Stationary Off-Policy Optimization