计算广告论文、学习资料、业界分享

在这里分享我工作中实现或者阅读过的计算广告相关论文、学习资料和业界分享。作为自己工作的整理和总结，也希望能为计算广告相关行业的技术同学带来便利。所有资料均来自于互联网，如有侵权，请联系王喆

下面将列出所有的资料目录，以及我对每篇文章的简要介绍
如有任何问题，欢迎对计算广告感兴趣的同学与我讨论，我的联系方式如下：

• email: [email protected]
• 知乎私信: 王喆的知乎
• 主页留言: 王喆的主页

目录

Allocation

广告流量的分配问题

• Ad Serving Using a Compact Allocation Plan.pdf
  雅虎的一篇比较经典的流量分配的文章，文中的HWM和DUAL算法都比较实用
• An Efficient Algorithm for Allocation of Guaranteed Display Advertising.pdf
  同样是雅虎的流量分配文章，跟上一篇文章同时发布，介绍SHALE流量分配算法

Bidding Strategy

计算广告中广告定价，RTB过程中广告出价策略的相关问题

• [Combining Powers of Two Predictors in Optimizing Real-Time Bidding Strategy under Constrained Budget.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Bidding Strategy/Combining%20Powers%20of%20Two%20Predictors%20in%20Optimizing%20Real-Time%20Bidding%20Strategy%20under%20Constrained%20Budget.pdf)
  国立台湾大学的文章，介绍一种基于流量选择的计算广告竞价方法，有别于传统的CTR CPC的方法，我在实践中尝试过该方法，非常有效
• [Real-Time Bidding Algorithms for Performance-Based Display Ad Allocation.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Bidding Strategy/Real-Time%20Bidding%20Algorithms%20for%20Performance-Based%20Display%20Ad%20Allocation.pdf)
  微软的一篇基于PID反馈控制的与效果相关的竞价算法
• [Research Frontier of Real-Time Bidding based Display Advertising.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Bidding Strategy/Research%20Frontier%20of%20Real-Time%20Bidding%20based%20Display%20Advertising.pdf)
  张伟楠博士的一篇介绍竞价算法的ppt，可以非常清晰的了解该问题的主要方法

Budget Control

广告系统中Pacing，预算控制，以及怎么把预算控制与其他模块相结合的问题

• [Budget Pacing for Targeted Online Advertisements at LinkedIn.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Budget Control/Budget%20Pacing%20for%20Targeted%20Online%20Advertisements%20at%20LinkedIn.pdf)
  linkedin的一篇非常有工程价值的解决pacing问题的文章，强烈建议计算广告系统采用此方法。
• [PID控制原理与控制算法.doc](https://github.com/wzhe06/Ad-papers/blob/master/Budget Control/PID%E6%8E%A7%E5%88%B6%E5%8E%9F%E7%90%86%E4%B8%8E%E6%8E%A7%E5%88%B6%E7%AE%97%E6%B3%95.doc)
  对于采用PID控制解决pacing问题，该文章是PID控制原理比较清晰的介绍文章。
• [PID控制经典培训教程.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Budget Control/PID%E6%8E%A7%E5%88%B6%E7%BB%8F%E5%85%B8%E5%9F%B9%E8%AE%AD%E6%95%99%E7%A8%8B.pdf)
  PID控制的经典教程
• [Real Time Bid Optimization with Smooth Budget Delivery in Online Advertising.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Budget Control/Real%20Time%20Bid%20Optimization%20with%20Smooth%20Budget%20Delivery%20in%20Online%20Advertising.pdf)
  如何将Pcaing与效果优化结合在一起，这篇文章讲的很清楚
• [Smart Pacing for Effective Online Ad Campaign Optimization.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Budget Control/Smart%20Pacing%20for%20Effective%20Online%20Ad%20Campaign%20Optimization.pdf)
  跟上篇文章一样，都是雅虎同一组人写的，解决预算控制与效果结合的问题，可以跟上篇文章一起看了

Computational Advertising Architect

广告系统的架构问题

• [Display Advertising with Real-Time Bidding (RTB) and Behavioural Targeting.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Computational Advertising Architect/Display%20Advertising%20with%20Real-Time%20Bidding%20%28RTB%29%20and%20Behavioural%20Targeting.pdf)
  张伟楠博士的RTB过程所有相关算法的书，全而精，非常棒
• [大数据下的广告排序技术及实践.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Computational Advertising Architect/%E5%A4%A7%E6%95%B0%E6%8D%AE%E4%B8%8B%E7%9A%84%E5%B9%BF%E5%91%8A%E6%8E%92%E5%BA%8F%E6%8A%80%E6%9C%AF%E5%8F%8A%E5%AE%9E%E8%B7%B5.pdf)
  阿里妈妈的一篇广告排序问题的ppt，模型、训练、评估都有涉及，很有工程价值
• [美团机器学习 吃喝玩乐中的算法问题.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Computational Advertising Architect/%E7%BE%8E%E5%9B%A2%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%20%E5%90%83%E5%96%9D%E7%8E%A9%E4%B9%90%E4%B8%AD%E7%9A%84%E7%AE%97%E6%B3%95%E9%97%AE%E9%A2%98.pdf)
  美团王栋博士的一篇关于美团机器学习相关问题的介绍，介绍的比较全但比较粗浅，可以借此了解美团的一些机器学习问题

CTR Prediction

CTR预估模型相关问题

• [Ad Click Prediction a View from the Trenches.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR Prediction/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。
• [Adaptive Targeting for Online Advertisement.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR Prediction/Adaptive%20Targeting%20for%20Online%20Advertisement.pdf)
  一篇比较简单但是全面的CTR预估的文章，有一定实用性
• [Logistic Regression in Rare Events Data.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR Prediction/Logistic%20Regression%20in%20Rare%20Events%20Data.pdf)
  样本稀少情况下的LR模型训练，讲的比较细
• [Practical Lessons from Predicting Clicks on Ads at Facebook.pdf](https://github.com/wzhe06/Ad-papers/blob/master/CTR Prediction/Practical%20Lessons%20from%20Predicting%20Clicks%20on%20Ads%20at%20Facebook.pdf)
  Facebook的一篇非常出名的文章，GBDT＋LR/FM解决CTR预估问题，工程性很强

Explore and Exploit

探索和利用问题，计算广告中非常经典的问题， 也是容易被大家忽视的问题，其实所有的广告系统都面临如何解决新广告主冷启动的问题，以及在效果不好的情况下如何探索新的优质流量的问题，希望该目录下的几篇文章能搞帮助到你。

• [A Contextual-Bandit Approach to Personalized News Article Recommendation(LinUCB).pdf](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/A%20Contextual-Bandit%20Approach%20to%20Personalized%20News%20Article%20Recommendation%28LinUCB%29.pdf)
  
• [An Empirical Evaluation of Thompson Sampling.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/An%20Empirical%20Evaluation%20of%20Thompson%20Sampling.pdf)
  
• [Analysis of Thompson Sampling for the Multi-armed Bandit Problem.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/Analysis%20of%20Thompson%20Sampling%20for%20the%20Multi-armed%20Bandit%20Problem.pdf)
  
• [Finite-time Analysis of the Multiarmed Bandit Problem.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/Finite-time%20Analysis%20of%20the%20Multiarmed%20Bandit%20Problem.pdf)
  
• [Thompson Sampling PPT.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/Thompson%20Sampling%20PPT.pdf)
  
• [UCT算法.doc](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/UCT%E7%AE%97%E6%B3%95.doc)
  
• [基于UCT的围棋引擎的研究与实现.doc](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/%E5%9F%BA%E4%BA%8EUCT%E7%9A%84%E5%9B%B4%E6%A3%8B%E5%BC%95%E6%93%8E%E7%9A%84%E7%A0%94%E7%A9%B6%E4%B8%8E%E5%AE%9E%E7%8E%B0.doc)
  
• [对抗搜索、多臂老虎机问题、UCB算法.ppt](https://github.com/wzhe06/Ad-papers/blob/master/Explore and Exploit/%E5%AF%B9%E6%8A%97%E6%90%9C%E7%B4%A2%E3%80%81%E5%A4%9A%E8%87%82%E8%80%81%E8%99%8E%E6%9C%BA%E9%97%AE%E9%A2%98%E3%80%81UCB%E7%AE%97%E6%B3%95.ppt)
  

Factorization Machines

FM因子分解机模型的相关paper，在计算广告领域非常实用的模型

• [Factorization Machines Rendle2010.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Factorization Machines/Factorization%20Machines%20Rendle2010.pdf)
  

• [Fast Context-aware Recommendations with Factorization Machines.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Factorization Machines/Fast%20Context-aware%20Recommendations%20with%20Factorization%20Machines.pdf)
  

• [fastFM- A Library for Factorization Machines.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Factorization Machines/fastFM-%20A%20Library%20for%20Factorization%20Machines.pdf)
  

• [FM PPT by CMU.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Factorization Machines/FM%20PPT%20by%20CMU.pdf)
  

• [libfm-1.42.manual.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Factorization Machines/libfm-1.42.manual.pdf)
  

• [Scaling Factorization Machines to Relational Data.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Factorization Machines/Scaling%20Factorization%20Machines%20to%20Relational%20Data.pdf)
  

Google Three Papers

Google三大篇，HDFS，MapReduce，BigTable，奠定大数据基础架构的三篇文章，应该读一读

• [Bigtable A Distributed Storage System for Structured Data.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Google Three Papers/Bigtable%20A%20Distributed%20Storage%20System%20for%20Structured%20Data.pdf)
  
• [MapReduce Simplified Data Processing on Large Clusters.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Google Three Papers/MapReduce%20Simplified%20Data%20Processing%20on%20Large%20Clusters.pdf)
  
• [The Google File System.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Google Three Papers/The%20Google%20File%20System.pdf)
  

Guaranteed Contracts Ads

• [Pricing Guaranteed Contracts in Online Display Advertising.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Guaranteed Contracts Ads/Pricing%20Guaranteed%20Contracts%20in%20Online%20Display%20Advertising.pdf)
  
• [Pricing Guidance in Ad Sale Negotiations The PrintAds Example.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Guaranteed Contracts Ads/Pricing%20Guidance%20in%20Ad%20Sale%20Negotiations%20The%20PrintAds%20Example.pdf)
  
• [Risk-Aware Revenue Maximization in Display Advertising.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Guaranteed Contracts Ads/Risk-Aware%20Revenue%20Maximization%20in%20Display%20Advertising.pdf)
  

Machine Learning Tutorial

机器学习方面一些非常实用的学习资料

• [Deep Learning Tutorial.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Machine Learning Tutorial/Deep%20Learning%20Tutorial.pdf)
  
• [Rules of Machine Learning- Best Practices for ML Engineering.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Machine Learning Tutorial/Rules%20of%20Machine%20Learning-%20Best%20Practices%20for%20ML%20Engineering.pdf)
  
• [关联规则基本算法及其应用.doc](https://github.com/wzhe06/Ad-papers/blob/master/Machine Learning Tutorial/%E5%85%B3%E8%81%94%E8%A7%84%E5%88%99%E5%9F%BA%E6%9C%AC%E7%AE%97%E6%B3%95%E5%8F%8A%E5%85%B6%E5%BA%94%E7%94%A8.doc)
  
• [各种回归的概念学习.doc](https://github.com/wzhe06/Ad-papers/blob/master/Machine Learning Tutorial/%E5%90%84%E7%A7%8D%E5%9B%9E%E5%BD%92%E7%9A%84%E6%A6%82%E5%BF%B5%E5%AD%A6%E4%B9%A0.doc)
  
• [广义线性模型.ppt](https://github.com/wzhe06/Ad-papers/blob/master/Machine Learning Tutorial/%E5%B9%BF%E4%B9%89%E7%BA%BF%E6%80%A7%E6%A8%A1%E5%9E%8B.ppt)
  
• [机器学习总图.jpg](https://github.com/wzhe06/Ad-papers/blob/master/Machine Learning Tutorial/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E6%80%BB%E5%9B%BE.jpg)
  
• [贝叶斯统计学(PPT).pdf](https://github.com/wzhe06/Ad-papers/blob/master/Machine Learning Tutorial/%E8%B4%9D%E5%8F%B6%E6%96%AF%E7%BB%9F%E8%AE%A1%E5%AD%A6%28PPT%29.pdf)
  

Optimization Method

Online Optimization，Parallel SGD，FTRL等优化方法，很实用的一些文章

• [A Survey on Algorithms of the Regularized Convex Optimization Problem.pptx](https://github.com/wzhe06/Ad-papers/blob/master/Optimization Method/A%20Survey%20on%20Algorithms%20of%20the%20Regularized%20Convex%20Optimization%20Problem.pptx)
  
• [Follow-the-Regularized-Leader and Mirror Descent- Equivalence Theorems and L1 Regularization.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Optimization Method/Follow-the-Regularized-Leader%20and%20Mirror%20Descent-%20Equivalence%20Theorems%20and%20L1%20Regularization.pdf)
  
• [Hogwild A Lock-Free Approach to Parallelizing Stochastic Gradient Descent.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Optimization Method/Hogwild%20A%20Lock-Free%20Approach%20to%20Parallelizing%20Stochastic%20Gradient%20Descent.pdf)
  
• [Parallelized Stochastic Gradient Descent.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Optimization Method/Parallelized%20Stochastic%20Gradient%20Descent.pdf)
  
• [在线最优化求解(Online Optimization)-冯扬.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Optimization Method/%E5%9C%A8%E7%BA%BF%E6%9C%80%E4%BC%98%E5%8C%96%E6%B1%82%E8%A7%A3%28Online%20Optimization%29-%E5%86%AF%E6%89%AC.pdf)
  
• [非线性规划.doc](https://github.com/wzhe06/Ad-papers/blob/master/Optimization Method/%E9%9D%9E%E7%BA%BF%E6%80%A7%E8%A7%84%E5%88%92.doc)
  

Recommendation

推荐系统相关文章，研究不多，欢迎补充

• 基于BPR-MF算法的推荐系统设计.docx
  
• 微博推荐策略平台Eros.pdf
  

Topic Model

话题模型相关文章，PLSA，LDA，进行广告Context特征提取，创意优化肯定会用到Topic Model

• [Dirichlet Distribution, Dirichlet Process and Dirichlet Process Mixture(PPT).pdf](https://github.com/wzhe06/Ad-papers/blob/master/Topic Model/Dirichlet%20Distribution%2C%20Dirichlet%20Process%20and%20Dirichlet%20Process%20Mixture%28PPT%29.pdf)
  
• [LDA数学八卦.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Topic Model/LDA%E6%95%B0%E5%AD%A6%E5%85%AB%E5%8D%A6.pdf)
  
• [Parameter estimation for text analysis.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Topic Model/Parameter%20estimation%20for%20text%20analysis.pdf)
  
• [概率语言模型及其变形系列.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Topic Model/%E6%A6%82%E7%8E%87%E8%AF%AD%E8%A8%80%E6%A8%A1%E5%9E%8B%E5%8F%8A%E5%85%B6%E5%8F%98%E5%BD%A2%E7%B3%BB%E5%88%97.pdf)
  
• [理解共轭先验.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Topic Model/%E7%90%86%E8%A7%A3%E5%85%B1%E8%BD%AD%E5%85%88%E9%AA%8C.pdf)
  

Transfer Learning

迁移学习相关文章，计算广告中经常遇到新广告冷启动的问题，利用迁移学习能较好解决该问题

• [A Survey on Transfer Learning.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Transfer Learning/A%20Survey%20on%20Transfer%20Learning.pdf)
  
• [Scalable Hands-Free Transfer Learning for Online Advertising.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Transfer Learning/Scalable%20Hands-Free%20Transfer%20Learning%20for%20Online%20Advertising.pdf)
  

Tree Model

树模型和基于树模型的boosting模型，树模型的效果在大部分问题上非常好，在CTR，CVR模型以及特征工程方面的应用非常广，值得深入研究

• [Classification and Regression Trees.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Tree Model/Classification%20and%20Regression%20Trees.pdf)
  
• [Classification and Regression Trees.ppt](https://github.com/wzhe06/Ad-papers/blob/master/Tree Model/Classification%20and%20Regression%20Trees.ppt)
  
• [Greedy Function Approximation A Gradient Boosting Machine.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Tree Model/Greedy%20Function%20Approximation%20A%20Gradient%20Boosting%20Machine.pdf)
  
• [Introduction to Boosted Trees.pdf](https://github.com/wzhe06/Ad-papers/blob/master/Tree Model/Introduction%20to%20Boosted%20Trees.pdf)