arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2106.14979cs.IRcs.LGstat.ML
收藏

两阶段推荐系统中组件交互的研究

On component interactions in two-stage recommender systems

University of Cambridge(剑桥大学);UC Berkeley(加州大学伯克利分校);Technical University of Munich(慕尼黑工业大学);Helmholtz AI, Munich(亥姆霍兹人工智能中心,慕尼黑)

Jiri Hron, Karl Krauth, Michael I. Jordan, Niki Kilbertus

摘要:

由于其可扩展性,两阶段搜索器被当今许多最大的在线平台使用,包括YouTube,LinkedIn和Pinterest。这些系统通过两个步骤产生推荐:(i)多个提名者,针对低预测延迟进行调整,从整个项目池中预选候选者的一个小子集;(ii)更慢但更准确的排名器进一步缩小提名项目,并为用户服务。尽管它们很受欢迎,但关于两阶段算法的文献相对较少,并且算法通常被视为仅仅是其部分的总和。这种处理的前提是,两阶段的性能是由单独的组成部分的行为解释。事实并非如此:使用合成和真实世界的数据,我们证明,排名和提名者之间的相互作用,大大影响整体性能。受这些发现的启发,我们推导出一个泛化下限,表明独立提名人训练可以导致与均匀随机推荐相当的性能。我们发现,仔细设计的项目池,每个分配给不同的提名人,这些问题。由于手动搜索一个好的池分配是困难的,我们建议学习一个,而不是使用混合的专家为基础的方法。这显著提高了K的查准率和查全率。

英文摘要:

Thanks to their scalability, two-stage recommenders are used by many of today's largest online platforms, including YouTube, LinkedIn, and Pinterest. These systems produce recommendations in two steps: (i) multiple nominators, tuned for low prediction latency, preselect a small subset of candidates from the whole item pool; (ii) a slower but more accurate ranker further narrows down the nominated items, and serves to the user. Despite their popularity, the literature on two-stage recommenders is relatively scarce, and the algorithms are often treated as mere sums of their parts. Such treatment presupposes that the two-stage performance is explained by the behavior of the individual components in isolation. This is not the case: using synthetic and real-world data, we demonstrate that interactions between the ranker and the nominators substantially affect the overall performance. Motivated by these findings, we derive a generalization lower bound which shows that independent nominator training can lead to performance on par with uniformly random recommendations. We find that careful design of item pools, each assigned to a different nominator, alleviates these issues. As manual search for a good pool allocation is difficult, we propose to learn one instead using a Mixture-of-Experts based approach. This significantly improves both precision and recall at K.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑