An End-to-End Multi-objective Ensemble Ranking Framework for Video Recommendation

2025-08-07

AI Summary

To address the lack of effective supervision signals and insufficient personalization in multi-objective ensemble ranking for short-video recommendation, this paper proposes EMER, an end-to-end framework. Methodologically, (i) it introduces the first learnable loss function specifically designed for multi-objective ensemble ranking—replacing hand-crafted heuristic formulas—to enable supervised learning in the absence of a single ground-truth label; (ii) it designs a Transformer-based relative relationship modeling architecture coupled with a novel sample organization strategy to explicitly capture ordinal relationships among candidate videos; and (iii) it establishes a consistent offline–online evaluation protocol. Deployed on Kuaishou’s primary recommendation scenario, EMER achieves a 1.39% increase in average app session duration and a 0.196 percentage-point improvement in 7-day user retention (LT7), significantly outperforming industrial baselines.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Computer Vision: Multi-modal Vision
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

We propose a novel End-to-end Multi-objective Ensemble Ranking framework (EMER) for the multi-objective ensemble ranking module, which is the most critical component of the short video recommendation system. EMER enhances personalization by replacing manually-designed heuristic formulas with an end-to-end modeling paradigm. EMER introduces a meticulously designed loss function to address the fundamental challenge of defining effective supervision for ensemble ranking, where no single ground-truth signal can fully capture user satisfaction. Moreover, EMER introduces novel sample organization method and transformer-based network architecture to capture the comparative relationships among candidates, which are critical for effective ranking. Additionally, we have proposed an offline-online consistent evaluation system to enhance the efficiency of offline model optimization, which is an established yet persistent challenge within the multi-objective ranking domain in industry. Abundant empirical tests are conducted on a real industrial dataset, and the results well demonstrate the effectiveness of our proposed framework. In addition, our framework has been deployed in the primary scenarios of Kuaishou, a short video recommendation platform with hundreds of millions of daily active users, achieving a 1.39% increase in overall App Stay Time and a 0.196% increase in 7-day user Lifetime(LT7), which are substantial improvements.

Problem

• Replacing heuristic formulas with end-to-end modeling for personalization
• Defining effective supervision for multi-objective ensemble ranking
• Enhancing offline model optimization with consistent evaluation

Innovation

• End-to-end modeling replaces heuristic formulas
• Transformer-based network captures candidate relationships
• Offline-online consistent evaluation enhances optimization