跳转至

Awesome Ads 🎯

经典论文

经典论文 (Classic Papers)¶

一句话概述¶

广告技术领域的经典论文，涵盖 CTR 预估、出价优化、推荐系统、隐私计算等方向，是深入理解广告算法的必读材料。

CTR/CVR 预估¶

年份	论文	机构	核心贡献
2014	Practical Lessons from Predicting Clicks on Ads at Facebook	Facebook	GBDT+LR 组合模型
2016	Wide & Deep Learning for Recommender Systems	Google	Wide&Deep 架构
2017	DeepFM: A Factorization-Machine based Neural Network	Huawei	FM + Deep 端到端
2017	Deep & Cross Network for Ad Click Predictions	Google	显式交叉网络 DCN
2018	DIN: Deep Interest Network for Click-Through Rate Prediction	Alibaba	Attention 建模用户兴趣
2019	DIEN: Deep Interest Evolution Network	Alibaba	兴趣演化建模
2018	ESMM: Entire Space Multi-Task Model	Alibaba	CVR 样本选择偏差
2018	MMOE: Modeling Task Relationships in Multi-task Learning	Google	多任务学习
2020	PLE: Progressive Layered Extraction	Tencent	改进多任务学习
2019	DSIN: Deep Session Interest Network	Alibaba	Session 级兴趣建模
2019	FiBiNET: Feature Importance and Bilinear feature Interaction	Sina	特征重要性 + 双线性交互
2021	DCN V2: Improved Deep & Cross Network	Google	DCN 改进版
2020	AutoInt: Automatic Feature Interaction Learning	Peking Univ	自动特征交互

出价与预算优化¶

年份	论文	机构	核心贡献
2017	Bid Optimizing and Inventory Scoring in Targeted Online Advertising	Yahoo	出价优化理论
2018	Budget Constrained Bidding by Model-free Reinforcement Learning	Alibaba	RL 出价
2018	Real-Time Bidding with Multi-Agent Reinforcement Learning	UCL	多智能体 RTB
2019	Bid Optimization by Multivariable Control in Display Advertising	Alibaba	PID 控制出价
2020	OCPC: A Bidding Strategy for Sponsored Search	Baidu	oCPC 两阶段
2021	Auto-bidding in Real-Time Bidding	Alibaba	自动出价框架

推荐与广告融合¶

年份	论文	机构	核心贡献
2020	Ads Allocation in Feed via Constrained Optimization	Meta	信息流广告分配
2019	Deep Reinforcement Learning for Ad Allocation in Feed	Alibaba	RL 广告插入
2021	Jointly Optimizing Revenue and User Experience	ByteDance	收入与体验平衡

搜索广告¶

年份	论文	机构	核心贡献
2013	DSSM: Learning Deep Structured Semantic Models	Microsoft	语义匹配模型
2007	Predicting Clicks: Estimating the CTR of New Ads	Microsoft	搜索广告 CTR 预估
2015	A Sub-linear, Massive-scale Look-alike Audience Extension	Yahoo	Lookalike 扩量

竞价机制¶

年份	论文	机构	核心贡献
2007	Internet Advertising and the Generalized Second-Price Auction	Google/Yahoo	GSP 竞价理论
2019	Reserve Price Optimization at Scale	Facebook	底价优化
2021	From GSP to First-Price Auction	—	一价竞价迁移

隐私与联邦学习¶

年份	论文	机构	核心贡献
2017	Federated Learning: Strategies for Improving Communication Efficiency	Google	联邦学习基础
2019	Federated Learning for Mobile Keyboard Prediction	Google	联邦学习应用
2006	Calibrating Noise to Sensitivity in Private Data Analysis	—	差分隐私基础
2014	RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response	Google	本地差分隐私

反作弊¶

年份	论文	机构	核心贡献
2012	Click Fraud Detection	Microsoft	点击欺诈检测
2020	A Survey on Click Fraud in Online Advertising	—	点击欺诈综述

在线学习¶

年份	论文	机构	核心贡献
2013	Ad Click Prediction: a View from the Trenches	Google	FTRL 在线学习
2010	Web-Scale Bayesian Click-Through Rate Prediction	Microsoft	贝叶斯 CTR 预估

Embedding 与表示学习¶

年份	论文	机构	核心贡献
2018	Billion-scale Commodity Embedding for E-commerce Recommendation	Alibaba	大规模 Embedding
2019	Real-time Attention Based Look-alike Model	Tencent	实时 Lookalike
2016	Deep Neural Networks for YouTube Recommendations	Google	YouTube 推荐系统

阅读建议¶

入门路线:
  1. LR → GBDT+LR (Facebook 2014)
  2. FM → DeepFM
  3. Wide & Deep (Google 2016)
  4. DIN → DIEN (Alibaba 2018-2019)
  5. ESMM (多任务学习)
  6. MMOE → PLE

进阶路线:
  7. 出价优化 (RL Bidding)
  8. 在线学习 (FTRL)
  9. 联邦学习
  10. 竞价机制理论