Deconfounding Duration Bias in Watch-time Prediction for Video Recommendation

TL;DR

This paper identifies and addresses duration bias in watch-time prediction for video recommendation, proposing a causal inference-based framework that improves prediction accuracy and user engagement in industry settings.

Contribution

It introduces the first causal graph analysis of duration bias and proposes a scalable deconfounding method for watch-time prediction in video recommendation systems.

Findings

Significantly outperforms state-of-the-art baselines in offline evaluations.

Improves real-time video consumption on Kuaishou App.

Effectively removes duration bias while preserving true user interests.

Abstract

Watch-time prediction remains to be a key factor in reinforcing user engagement via video recommendations. It has become increasingly important given the ever-growing popularity of online videos. However, prediction of watch time not only depends on the match between the user and the video but is often mislead by the duration of the video itself. With the goal of improving watch time, recommendation is always biased towards videos with long duration. Models trained on this imbalanced data face the risk of bias amplification, which misguides platforms to over-recommend videos with long duration but overlook the underlying user interests. This paper presents the first work to study duration bias in watch-time prediction for video recommendation. We employ a causal graph illuminating that duration is a confounding factor that concurrently affects video exposure and watch-time prediction -- the first effect on video causes the bias issue and should be eliminated, while the second effect on watch time originates from video intrinsic characteristics and should be preserved. To remove the undesired bias but leverage the natural effect, we propose a Duration Deconfounded Quantile-based (D2Q) watch-time prediction framework, which allows for scalability to perform on industry production systems. Through extensive offline evaluation and live experiments, we showcase the effectiveness of this duration-deconfounding framework by significantly outperforming the state-of-the-art baselines. We have fully launched our approach on Kuaishou App, which has substantially improved real-time video consumption due to more accurate watch-time predictions.