Balancing Semantic Relevance and Engagement in Related Video Recommendations

TL;DR

This paper presents a multi-objective retrieval framework for related video recommendations that balances semantic relevance and user engagement by combining multi-task learning, multimodal content features, and off-policy correction, leading to improved relevance and reduced popularity bias.

Contribution

It introduces a novel multi-objective retrieval approach that explicitly balances semantic relevance and engagement, incorporating multimodal features and bias mitigation techniques, with demonstrated industrial-scale effectiveness.

Findings

Semantic relevance improved from 51% to 63%

Popularity bias reduced by 13.8%

User engagement metric increased by 0.04%

Abstract

Related video recommendations commonly use collaborative filtering (CF) driven by co-engagement signals, often resulting in recommendations lacking semantic coherence and exhibiting strong popularity bias. This paper introduces a novel multi-objective retrieval framework, enhancing standard two-tower models to explicitly balance semantic relevance and user engagement. Our approach uniquely combines: (a) multi-task learning (MTL) to jointly optimize co-engagement and semantic relevance, explicitly prioritizing topical coherence; (b) fusion of multimodal content features (textual and visual embeddings) for richer semantic understanding; and (c) off-policy correction (OPC) via inverse propensity weighting to effectively mitigate popularity bias. Evaluation on industrial-scale data and a two-week live A/B test reveals our framework's efficacy. We observed significant improvements in semantic relevance (from 51% to 63% topic match rate), a reduction in popular item distribution (-13.8% popular video recommendations), and a +0.04% improvement in our topline user engagement metric. Our method successfully achieves better semantic coherence, balanced engagement, and practical scalability for real-world deployment.