diff --git a/CTR Prediction/Deep Learning over Multi-field Categorical Data.pdf b/CTR Prediction/Deep Learning over Multi-field Categorical Data.pdf new file mode 100644 index 0000000..9964660 Binary files /dev/null and b/CTR Prediction/Deep Learning over Multi-field Categorical Data.pdf differ diff --git a/CTR Prediction/Product-based Neural Networks for User Response Prediction.pdf b/CTR Prediction/Product-based Neural Networks for User Response Prediction.pdf new file mode 100644 index 0000000..c82d743 Binary files /dev/null and b/CTR Prediction/Product-based Neural Networks for User Response Prediction.pdf differ diff --git a/CTR Prediction/Wide & Deep Learning for Recommender Systems.pdf b/CTR Prediction/Wide & Deep Learning for Recommender Systems.pdf new file mode 100644 index 0000000..34b4b58 Binary files /dev/null and b/CTR Prediction/Wide & Deep Learning for Recommender Systems.pdf differ diff --git a/Computational Advertising Architect/Overlapping Experiment Infrastructure More, Better, Faster Experimentation.pdf b/Computational Advertising Architect/Overlapping Experiment Infrastructure More, Better, Faster Experimentation.pdf new file mode 100644 index 0000000..027118e Binary files /dev/null and b/Computational Advertising Architect/Overlapping Experiment Infrastructure More, Better, Faster Experimentation.pdf differ diff --git a/README.md b/README.md index 80e1227..ba63f4b 100644 --- a/README.md +++ b/README.md @@ -45,6 +45,12 @@ CTR预估模型相关问题 * [Learning Piece-wise Linear Models from Large Scale Data for Ad Click Prediction.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/Learning%20Piece-wise%20Linear%20Models%20from%20Large%20Scale%20Data%20for%20Ad%20Click%20Prediction.pdf)
* [Entire Space Multi-Task Model_ An Effective Approach for Estimating Post-Click Conversion Rate.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/Entire%20Space%20Multi-Task%20Model_%20An%20Effective%20Approach%20for%20Estimating%20Post-Click%20Conversion%20Rate.pdf)
* [Deep Interest Network for Click-Through Rate Prediction.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/Deep%20Interest%20Network%20for%20Click-Through%20Rate%20Prediction.pdf)
+* [Wide & Deep Learning for Recommender Systems](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/Wide%20%26%20Deep%20Learning%20for%20Recommender%20Systems.pdf)
+Google 的 Wide & Deep 模型，论文将模型用于推荐系统中，但也可用于 CTR 预估中 +* [Deep Learning over Multi-field Categorical Data](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/Deep%20Learning%20over%20Multi-field%20Categorical%20Data.pdf)
+张伟楠博士的论文，提出了 FNN 模型，类似 Wide & Deep 的 Deep 部分，亮点在于用 FM 预训练的隐向量初始化 embedding 层 +* [Product-based Neural Networks for User Response Prediction](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/Product-based%20Neural%20Networks%20for%20User%20Response%20Prediction.pdf)
+张伟楠博士的另外一篇论文，提出了 PNN 模型，在 FNN 基础上对特征的隐向量进行了 inner product 作为新特征 * [Ad Click Prediction a View from the Trenches.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/Ad%20Click%20Prediction%20a%20View%20from%20the%20Trenches.pdf)
Google大名鼎鼎的用FTRL解决CTR在线预估的工程文章，非常经典。 * [DeepFM- A Factorization-Machine based Neural Network for CTR Prediction.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR%20Prediction/DeepFM-%20A%20Factorization-Machine%20based%20Neural%20Network%20for%20CTR%20Prediction.pdf)
@@ -173,6 +179,8 @@ PID控制的经典教程 张伟楠博士的RTB过程所有相关算法的书，全而精，非常棒 * [Efficient Query Evaluation using a Two-Level Retrieval Process.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Computational%20Advertising%20Architect/Efficient%20Query%20Evaluation%20using%20a%20Two-Level%20Retrieval%20Process.pdf)
搜索广告中经典的搜索算法 Wand(Weak AND) +* [Overlapping Experiment Infrastructure:More, Better, Faster Experimentation.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Computational%20Advertising%20Architect/Overlapping%20Experiment%20Infrastructure%20More%2C%20Better%2C%20Faster%20Experimentation.pdf)
+Google 一篇关于 A/B 测试框架的论文，涉及到如何切分流量以同时进行多个 A/B 测试，工程性很强 * [Scaling Distributed Machine Learning with the Parameter Server.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Computational%20Advertising%20Architect/Scaling%20Distributed%20Machine%20Learning%20with%20the%20Parameter%20Server.pdf)