Click-through rate (CTR) prediction is a crucial task in web search, recommender systems, and online advertisement displaying. In practical application, CTR models often serve with high-speed user-generated data streams, whose underlying distribution rapidly changing over time. The concept drift problem inevitably exists in those streaming data, which can lead to performance degradation due to the timeliness issue. To ensure model freshness, incremental learning has been widely adopted in real-world production systems. However, it is hard for the incremental update to achieve the balance of the CTR models between the adaptability to capture the fast-changing trends and generalization ability to retain common knowledge. In this paper, we propose adaptive mixture of experts (AdaMoE), a new framework to alleviate the concept drift problem by statistical weighting policy in the data stream of CTR prediction. The extensive offline experiments on both benchmark and a real-world industrial dataset, as well as an online A/B testing show that our AdaMoE significantly outperforms all incremental learning frameworks considered.
翻译:点击通速( CTR) 预测是网络搜索、 推荐系统和在线广告显示的一项关键任务。 在实际应用中, CTR 模型通常用于高速用户生成的数据流,其基本分布随着时间的推移而发生迅速变化。 概念漂移问题不可避免地存在于这些流流数据中,由于及时性问题,这可能导致性能退化。为了确保模型的更新,在现实世界的生产系统中广泛采用了渐进式学习。然而,对于渐进式更新来说,实现CTR 模型在适应性以捕捉快速变化的趋势和普及性能力以保留共同知识之间的平衡是困难的。 在本文中,我们提出了适应性专家混合(AdaMoE),这是一个通过CTR预测数据流中的统计加权政策来缓解概念漂移问题的新框架。关于基准和实际世界工业数据集的广泛离线实验以及在线A/B测试表明,我们的AdaMoE 明显超越了所考虑的所有递增学习框架。