TriMLP: A Foundational MLP-Like Architecture for Sequential Recommendation.

Autor: Jiang, Yiheng, Xu, Yuanbo, Yang, Yongjian, Yang, Funing, Wang, Pengyang, Li, Chaozhuo, Zhuang, Fuzhen, Xiong, Hui
Zdroj: ACM Transactions on Information Systems; Nov2024, Vol. 42 Issue 6, p1-34, 34p
Abstrakt: The article focuses on the triangular multi-layer perceptron (TriMLP)-like artificial intelligence (AI) architecture model for sequential recommendation, balancing computational efficiency with improved performance. Topics include TriMLP's addressing of limitations of existing MLP models in sequential contexts, empirical studies highlighting the incompatibility of standard MLP structures with sequential data and TriMLP's competitive performance across various datasets.
Databáze: Complementary Index