[论文解读] Rethinking Multi-objective Ranking Ensemble in Recommender System: From Score Fusion to Rank Consistency

Boyang Xia, Zhou Yu+6 more|arXiv (Cornell University)|Jan 6, 2026

Recommender Systems and Techniques被引用 0

一句话总结

HarmonRank 引入一个与排名对齐的两步多目标集成，通过可微分的 AUC 优化并通过自注意力与个性化实现目标对齐，在离线上获得显著提升并在实际直播电商中实现在线改进。

ABSTRACT

The industrial recommender systems always pursue more than one business goals. The inherent intensions between objectives pose significant challenges for ranking stage. A popular solution is to build a multi-objective ensemble (ME) model to integrate multi-objective predictions into a unified score. Although there have been some exploratory efforts, few work has yet been able to systematically delineate the core requirements of ME problem. We rethink ME problem from two perspectives. From the perspective of each individual objective, to achieve its maximum value the scores should be as consistent as possible with the ranks of its labels. From the perspective of entire set of objectives, an overall optimum can be achieved only when the scores align with the commonality shared by the majority of objectives. However, none of existing methods can meet these two requirements. To fill this gap, we propose a novel multi-objective ensemble framework HarmonRank to fulfill both requirements. For rank consistency, we formulate rank consistency (AUC) metric as a rank-sum problem and make the model optimized towards rank consistency in an end-to-end differentiable manner. For commonality modeling, we change the original relation-agnostic ensemble paradigm to a relation-aware one. Extensive offline experimental results on two industrial datasets and online experiments demonstrate that our approach significantly outperforms existing state-of-the-art methods. Besides, our method exhibits superior robustness to label skew situations which is common in industrial scenarios. The proposed method has been fully deployed in Kuaishou's live-streaming e-commerce recommendation platform with 400 million DAUs, contributing 2.6% purchase gain.

研究动机与目标

• 在直播电商推荐中平衡短期购买与长期用户–主播互动的需求。
• 解决传统多目标集成仅优化独立二分类损失且忽略目标间相关性的局限。
• 提出 HarmoRank 以对齐排序目标与目标间关系，实现个性化集成打分。
• 在工业数据集与生产部署上展示离线与在线性能提升。

提出的方法

• 将 AUC 形式化为等级和问题，并通过可微排序优化，使训练与排序评估对齐。
• 引入两步对齐与集成范式，捕捉跨目标的共性排序能力。
• 应用关系感知模块，利用自注意力对齐目标编码，并以个性化查询特征指导集成打分。
• 辅以关系无关模块，具有门控机制和线性融合路径，以保留并鲁棒融合目标得分。
• 对分数进行离散化嵌入预处理，以在进入集成前实现非线性表示。
• 以端到端可微分目标进行训练，使多目标 AUC 之和最大化。

实验结果

研究问题

• RQ1 与最先进基线相比，HarmoRank 在工业多目标推荐数据集上的性能如何？
• RQ2 模块设计与超参数对性能的影响是什么？
• RQ3 在在线生产环境（AB 测试）中，HarmoRank 相对于基线是否能带来改进？
• RQ4 HarmoRank 如何影响多目标之间的权衡并揭示目标间的对齐？
• RQ5 学到的目标间对齐关系揭示了哪些目标之间的关系？

主要发现

• HarmoRank 在两个工业数据集上对比强基线，在 3 个与 5 个目标设置下实现优越的离线 AUC-和。
• 在线 A/B 测试在核心购买与互动指标上取得提升，如购买量提升 2.635%、关注度提升 0.451%。
• 一个可微分排序的 AUC 优化在准确性和效率方面均优于成对代理损失与实例级方法。
• 消融研究显示自注意力对齐、个性化引导、门控机制与离散化预处理对性能的重要性。
• 该方法相对于多目标 BCE 实现了帕累托改进的权衡，展现出对多目标的更好同时优化。