Category-Specific CNN for Visual-aware CTR Prediction at JD.com

1.运用场景

????CTR模型/加入图像特征。

2.创新点

????propose CSCNN,a novel visual embedding module specially for CTR prediction.The key idea is to conduct category-specific channel and spatial self-attention to emphasize features that are both important and category related.
????build efficient infrastructure to apply CNN in the real online e-commerce advertising system.

3.算法原理

3.1 网络框架

3.2 CSCNN

????CSCNN论文

4.算法理解

????引入visual embedding，且训练图像的特征时与物品类别对应，使特征能更好的用于CTR模型。

Category-Specific CNN for Visual-aware CTR Prediction at JD.com

原文：https://www.cnblogs.com/LuckPsyduck/p/13197338.html