Heterogeneous Multi-treatment Uplift Modeling for Trade-off Optimization in Short-Video Recommendation

2025-11-24

AI Summary

In short-video recommendation, multi-strategy interventions exhibit heterogeneous user responses and conflicting objectives (e.g., watch time vs. play count), posing challenges for existing models to jointly capture strategy synergy, individual causal effects, and personalized trade-offs. Method: We propose an offline-online collaborative framework for personalized objective balancing. Offline, we design a hybrid uplift modeling module that disentangles joint and independent causal effects across strategies. Online, we introduce a dynamic weight estimation mechanism that adaptively assigns personalized weights to competing metrics based on real-time user feedback. Our approach integrates causal inference, multi-task learning, and reinforcement-inspired decision-making to support high-throughput, low-latency recommendation. Contribution/Results: Extensive evaluations on public benchmarks, industrial datasets, and Kuaishou’s large-scale A/B tests demonstrate significant improvements in key business metrics. The framework has been fully deployed, serving hundreds of millions of users in production.

Technology Category

• Machine Learning: Causal Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Reasoning under Uncertainty: Causality

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
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

The rapid proliferation of short videos on social media platforms presents unique challenges and opportunities for recommendation systems. Users exhibit diverse preferences, and the responses resulting from different strategies often conflict with one another, potentially exhibiting inverse correlations between metrics such as watch time and video view counts. Existing uplift models face limitations in handling the heterogeneous multi-treatment scenarios of short-video recommendations, often failing to effectively capture both the synergistic and individual causal effects of different strategies. Furthermore, traditional fixed-weight approaches for balancing these responses lack personalization and can result in biased decision-making. To address these issues, we propose a novel Heterogeneous Multi-treatment Uplift Modeling (HMUM) framework for trade-off optimization in short-video recommendations. HMUM comprises an Offline Hybrid Uplift Modeling (HUM) module, which captures the synergistic and individual effects of multiple strategies, and an Online Dynamic Decision-Making (DDM) module, which estimates the value weights of different user responses in real-time for personalized decision-making. Evaluated on two public datasets, an industrial dataset, and through online A/B experiments on the Kuaishou platform, our model demonstrated superior offline performance and significant improvements in key metrics. It is now fully deployed on the platform, benefiting hundreds of millions of users.

Problem

• Optimizing trade-offs between conflicting metrics in short-video recommendations
• Handling heterogeneous multi-treatment scenarios with synergistic and individual effects
• Addressing limitations of fixed-weight approaches for personalized decision-making

Innovation

• Heterogeneous Multi-treatment Uplift Modeling framework
• Offline Hybrid Uplift Modeling captures synergistic effects
• Online Dynamic Decision-Making enables personalized trade-offs