Learning to Rank with Selection Bias in Personal Search

By Xuanhui Wang et al

Read the original document by opening this link in a new tab.

Table of Contents

1. ABSTRACT
2. INTRODUCTION
3. RELATED WORK
4. PROBLEM FORMULATION
5. PROPOSED METHODS

Summary

This paper discusses the problem of selection bias in learning-to-rank models for personal search scenarios. It addresses the challenges of leveraging biased click-through data and proposes methods to estimate inverse propensity weights to overcome selection bias. The application scenario involves a search engine for a commercial email service. The study highlights the importance of accurate bias estimation for improving the effectiveness of learning-to-rank models.