一文尽览推荐系统模型演变史(文末可下载)

前言

1. 本次整理方式为时间线的形式，未来可能从其他视角和维度进行整理。

2. 由于本人知识有限，对于未出现在下文的推荐模型并不代表不经典，欢迎大家补充。

3. 文中提及的模型都标有对应参考文献，具体细节可阅读论文原文。

4. 整理此文的目的是给大家一个清晰的脉络，可当作一篇小小综述。从信息过载概念的提出到推荐系统的起源，从前深度学习时代的推荐系统到劲头正热的深度推荐系统，再到最后对于深度学习技术带来的推荐系统性能提升的质疑，每个阶段都是必不可少的。

5. 希望推荐系统领域未来仍然可以在曲折中前进。

6. 提前祝大家端午节快乐。

结尾

1. 回看推荐发展的30年，尽管前进的路上有过批评、有过怀疑，但一直在进步。

2. 至于推荐系统未来的发展何去何从，需要我们每个当前参与的人一起努力。

3. 也希望后人再总结的时候发现推荐系统在大方向上一直在茁壮成长。

参考文献

[1] Gross et al. The Managing Organizations: The Administrative Struggle. 1964.

[2] Karlgren et al. An Algebra for Recommendations.1990.

[3] Goldberg et al. Using collaborative filtering to weave an information tapestry.1992. 

[4] User-based CF of Netnews GroupLens: An Open Architecture for Collaborative Filtering of Netnews. 1994.

[5] Balabanovic et al. Fab: Content-based, collaborative recommendation.1997.

[6] Daniel et al. Learning Collaborative Information Filters. ICML,1998.

[7] Sarwar et al. Item-based collaborative filtering recommendation algorithms. WWW, 2001.

[8] Linden et al. Amazon.com recommendations: Item-to-item collaborative filtering. IEEE INTERNET COMPUT, 2003.

[9] Lemire et al. Slope One Predictors for Online Rating-Based Collaborative Filtering, SDM, 2005 .

[10] Bell et al. The BellKor solution to the Netflix Prize.2007.

[11] Simon Funk. Netflix Update: Try This at Home:https://sifter.org/~simon/journal/20061211.html, 2006.

[12] Ruslan et al. Restricted Boltzmann Machines for Collaborative Filtering,ICML, 2007.

[13] Mnih et al. Probabilistic matrix factorization. NIPS, 2008.

[14] Ma et al. Sorec: social recommendation using probabilistic matrix factorization. CIKM, 2008.

[15] Rendle et al. BPR: Bayesian personalized ranking from implicit feedback. UAI, 2009.

[16] Jamali et al. Trustwalker: a random walk model for combining trust-based and item-based recommendation. KDD, 2009.

[17] Koren et al. Matrix factorization techniques for recommender systems. Computer, 2009.

[18] Rendle. Factorization machines. ICDM, 2010.

[19] Ning et al. SLIM: Sparse linear methods for top-n recommender systems. ICDM, 2011.

[20] Ma et al. Recommender systems with social regularization. WSDM, 2011.

[21] Huang et al. Learning deep structured semantic models for web search using clickthrough data. 2013.

[22] Zhao et al. Leveraging social connections to improve personalized ranking for collaborative filtering. CIKM, 2014.

[23] He et al. Practical lessons from predicting clicks on ads at facebook. 2014.

[24] Sedhain et al. Autorec: Autoencoders meet collaborative filtering. WWW, 2015.

[25] Guo et al. TrustSVD: Collaborative Filtering with Both the Explicit and Implicit Influence of User Trust and of Item Ratings. AAAI, 2015.

[26] Barkan et al. Item2vec: neural item embedding for collaborative filtering. 2016.

[27] Covington et al. Deep Neural Networks for YouTube Recommendations. 2016.

[28] Cheng et al. Wide & Deep Learning for Recommender Systems. 2016.

[29] Shan et al. Deep crossing: Web-scale modeling without manually crafted combinatorial features. KDD, 2016.

[30] Yuchin et al. Field-aware Factorization Machines for CTR Prediction. RecSys, 2016.

[31] Zhang et al. Deep Learning over Multi-field Categorical Data. UCL, 2016.

[32] Yanru et al. Product-based Neural Networks for User Response Prediction. ICDM, 2016.

[33] Kim et al. Convolutional Matrix Factorization for Document Context-Aware Recommendation. RecSys, 2016.

[34] Hidasi et al. Session-based Recommendations with Recurrent Neural Networks. 2016.

[35] He et al. Neural collaborative filtering. WWW, 2017.

[36] Guo et al. Deepfm: A factorization-machine based neural network for ctr prediction. IJCAI, 2017.

[37] Hsieh et al. Collaborative metric learning. WWW, 2017.

[38] Gai et al. Learning Piece-wise Linear Models from Large Scale Data for Ad ClickPrediction, 2017.

[39]He et al. Neural Factorization Machines for Sparse Predictive Analytics. 2017.

[40] Xiao et al. Attentional factorization machines: Learning the weight of feature interactions via attention networks. IJCAI, 2017.

[41] Hsieh et al. Collaborative metric learning. WWW, 2017.

[42] Lei et al. Joint Deep Modeling of Users and Items Using Reviews for Recommendation. WSDM, 2017.

[43] Ruoxi et al. Deep & Cross Network for Ad Click Predictions. ADKDD, 2017.

[44] Ying et al. Graph Convolutional Neural Networks for Web-Scale Recommender Systems. KDD, 2018.

[45] Lian et al. xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems. KDD, 2018.

[46] Chen et al. Sequential Recommendation with User Memory Networks. WSDM 2018.

[47] Hongwei et al. RippleNet: Propagating User Preferences on the Knowledge Graph for Recommender Systems. CIKM, 2018.

[48] Zhou et al. Deep interest network for click-through rate prediction. KDD, 2018.

[49] Zhou et al. Deep interest evolution network for click-through rate prediction. AAAI, 2019.

[50] Huafeng et al. Deep Generative Ranking for Personalized Recommendation. Recsys, 2019.

[51] Xiang et al. Neural Graph Collaborative Filtering. SIGIR, 2019.

[52] Zhi-Hong et al. DeepCF: A Unified Framework of Representation Learning and Matching Function Learning in Recommender System. AAAI, 2019.

[53] Wu et al. Session-based Recommendation with Graph Neural Networks. AAAI, 2019.

[54] Maurizio et al. Are We Really Making Much Progress? A Worrying Analysis of Recent Neural Recommendation Approaches. RecSys, 2019.

[55] Rendle et al. Neural Collaborative Filtering vs. Matrix Factorization Revisited. arXiv, 2020.

[56] Noveen et al. How Useful are Reviews for Recommendation? A Critical Review and Potential Improvements. SIGIR, 2020

以上文献均可在 https://github.com/hongleizhang/RSPapers 中找到。

获取以上高清文件，公众号后台回复“ timeline “即可。

推荐阅读

[0].MLP or IP：推荐模型到底用哪个更好？

[1]. 评论文本信息对推荐真的有用吗？

[2]. 推荐系统之FM与MF傻傻分不清楚

[3].利用对抗技术来权衡推荐精度与用户隐私

---

喜欢的话点个在看吧👇