A Deep Recurrent Survival Model for Unbiased Ranking

TL;DR

This paper introduces a Deep Recurrent Survival Ranking model that jointly models user behaviors and contextual information to improve unbiased ranking in information retrieval, addressing limitations of existing methods.

Contribution

The proposed DRSR framework effectively incorporates contextual information and hidden user behavior issues, advancing unbiased ranking methods with an end-to-end neural approach.

Findings

Significant performance improvements over state-of-the-art methods.

Effective modeling of non-click queries and untrusted browsing logs.

Robustness demonstrated on large-scale industrial datasets.

Abstract

Position bias is a critical problem in information retrieval when dealing with implicit yet biased user feedback data. Unbiased ranking methods typically rely on causality models and debias the user feedback through inverse propensity weighting. While practical, these methods still suffer from two major problems. First, when inferring a user click, the impact of the contextual information, such as documents that have been examined, is often ignored. Second, only the position bias is considered but other issues resulted from user browsing behaviors are overlooked. In this paper, we propose an end-to-end Deep Recurrent Survival Ranking (DRSR), a unified framework to jointly model user's various behaviors, to (i) consider the rich contextual information in the ranking list; and (ii) address the hidden issues underlying user behaviors, i.e., to mine observe pattern in queries without any click (non-click queries), and to model tracking logs which cannot truly reflect the user browsing intents (untrusted observation). Specifically, we adopt a recurrent neural network to model the contextual information and estimates the conditional likelihood of user feedback at each position. We then incorporate survival analysis techniques with the probability chain rule to mathematically recover the unbiased joint probability of one user's various behaviors. DRSR can be easily incorporated with both point-wise and pair-wise learning objectives. The extensive experiments over two large-scale industrial datasets demonstrate the significant performance gains of our model comparing with the state-of-the-arts.