Click-through rate (CTR) prediction is one of the fundamental tasks for online advertising and recommendation. Although applying a vanilla MLP network alone is inefficient in learning multiplicative feature interactions, in this work, we show that a well-tuned two-stream MLP model that simply combines two MLPs can achieve surprisingly good performance. Based on this observation, we further propose feature selection and interaction aggregation layers that can be easily plugged in to build an enhanced two-stream MLP model, FinalMLP. We envision that the simple yet effective FinalMLP model could serve as a new strong baseline for future developments of two-stream CTR models.
Kelong Mao, Jieming Zhu, Liangcai Su, Guohao Cai, Yuru Li, Zhenhua Dong. FinalMLP: An Enhanced Two-Stream MLP Model for CTR Prediction, in AAAI 2023.