计算广告 权威资料总结

一、阿里巴巴
1.阿里自主创新的下一代匹配&推荐技术：任意深度学习+树状全库检索 https://zhuanlan.zhihu.com/p/35030348
解读博客 http://d0evi1.com/tdm/
Tree based deep model for rec
2. Behavior Sequence Transformer for E-commerce Recommendation in Alibaba
3. Billion-scale Commodity Embedding for E-commerce Recommendation in Alibaba

4.阿里技术 https://mp.weixin.qq.com/s/MBVac45WpZy4c208zF3Ang

二.Brandit算法与计算广告
http://xudongyang.coding.me/linucb/
https://blog.csdn.net/heyc861221/article/details/80129310

三、腾讯
1、Real-time Attention based Look-alike Model for Recommender System
参考：https://zhuanlan.zhihu.com/p/71951411
paper：https://arxiv.org/pdf/1906.05022.pdf
人群扩展相关：
（1）yahoo：https://blog.csdn.net/AckClinkz/article/details/82665578，http://proceedings.mlr.press/v53/ma16.pdf
（2）LSH相关：https://www.cnblogs.com/sddai/p/6110704.html， http://infolab.stanford.edu/~ullman/mmds/ch3.pdf
四、DataFun
1、新浪：DL 特征 交叉组合layer：FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Prediction：https://mp.weixin.qq.com/s/VGtfGPxu_unbge3JU0OL-A