SWaT: Statistical Modeling of Video Watch Time through User Behavior Analysis

2024-08-14

arXiv.org

AI Summary

To address inaccurate video watch-time prediction and poor model interpretability in short-video recommendation, this paper proposes a user-centric, white-box statistical modeling framework. The method formalizes domain-knowledge-driven behavioral hypotheses—such as non-stationary viewing probability progression—into interpretable probabilistic models, and introduces a progress-bar binning mechanism to jointly model continuous watch-time regression and binary viewing events. The framework ensures statistical rigor while maintaining industrial deployability, supporting both offline training and online serving. Evaluated on two public benchmarks, large-scale industrial offline data, and an A/B test on a platform with hundreds of millions of daily active users, the approach consistently outperforms strong baselines. It significantly improves watch-time prediction accuracy and key business metrics—including average watch time and session engagement—demonstrating both technical soundness and practical impact.

Technology Category

• Computer Vision: Video Understanding & Activity Analysis
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The significance of estimating video watch time has been highlighted by the rising importance of (short) video recommendation, which has become a core product of mainstream social media platforms. Modeling video watch time, however, has been challenged by the complexity of user-video interaction, such as different user behavior modes in watching the recommended videos and varying watching probability over the video progress bar. Despite the importance and challenges, existing literature on modeling video watch time mostly focuses on relatively black-box mechanical enhancement of the classical regression/classification losses, without factoring in user behavior in a principled manner. In this paper, we for the first time take on a user-centric perspective to model video watch time, from which we propose a white-box statistical framework that directly translates various user behavior assumptions in watching (short) videos into statistical watch time models. These behavior assumptions are portrayed by our domain knowledge on users' behavior modes in video watching. We further employ bucketization to cope with user's non-stationary watching probability over the video progress bar, which additionally helps to respect the constraint of video length and facilitate the practical compatibility between the continuous regression event of watch time and other binary classification events. We test our models extensively on two public datasets, a large-scale offline industrial dataset, and an online A/B test on a short video platform with hundreds of millions of daily-active users. On all experiments, our models perform competitively against strong relevant baselines, demonstrating the efficacy of our user-centric perspective and proposed framework.

Problem

• Modeling video watch time complexity
• Incorporating user behavior assumptions
• Handling non-stationary watching probability

Innovation

• User-centric statistical framework
• Bucketization for non-stationary probability
• Behavior assumptions into watch models