推荐系统与深度学习中的模型综述-优快云博客

这篇博客汇总了推荐系统中常见的模型，包括协同过滤、GBDT、矩阵分解、因子分解机、深度学习模型如DeepFM、NFM、AFM等。还介绍了网络嵌入、NLP相关技术和深度学习的最新进展，如Attention机制和Graph Embedding。

推荐相关

CF（Collaborative Filtering）

Item-Based Collaborative Filtering Recommendation Algorithms , paper链接 , paper-biji
Collaborative Metric Learning , paper链接 , paper-biji
Item2Vec: Neural Item Embedding for Collaborative Filtering , paper链接 , paper-biji
少数人的智慧(The Wisdom of the Few) , paper链接, zhihu
Neural Collaborative Filter , paper链接 , code , blog , paper-biji
Outer Product-based Neural Collaborative Filtering , paper链接 , code , blog , paper-biji
ALS , blog

GBDT及各种变形

greedy function approximation: a gradient boosting machine , paper链接 , code , blog , paper-biji
A Regression Framework for Learning Ranking Functions Using Relative Relevance Judgments , paper链接 , code , blog , paper-biji
From RankNet to LambdaRank to LambdaMART: An overview , paper链接 , code , blog , paper-biji
GBDT+LR , paper链接 , xgboost , lightgbm , paper-biji

MF（Matrix Factorization）

FM = LR + MF

Factorization Machines ,paper链接 , libfm , ppyFM , FM

FFM

Field-aware Factorization Machines for CTR Prediction , paper链接 ,FFM ,xlearn , libffm ,
深入FFM原理与实践_美团 , 3 Idiots’ Approach for Display Advertising Challenge

MLR（多元线性回归模型）= Embedding + MF + LR

Learning Piece-wise Linear Models from Large Scale Data for Ad Click Prediction , paper链接, blog

FNN = FM + MLP

Deep Learning over Multi-Field Categorical Data: A Case Study on User Response Prediction, paper链接 , code , blog , paper-biji

PNN = FNN + product layer

Product-based Neural Networks for User Response Prediction , paper链接 ,
code1-torch ,
code2-tf ,
code3-tf ,
code4-tf , blog , paper-biji ()

DeepFM = FM + Embedding + MLP

DeepFM: A Factorization-Machine based Neural Network for CTR Prediction, paper链接 ,
code1-tf ,
code2-tf ,
code3-tf ,
code4-torch ,
code5-tf ,
code6-tf ,
code7-tf , paper-biji

NFM（Neural FM） = FM + Embedding + MLP + Bi-Interaction Pooling

Neural Factorization Machines for Sparse Predictive Analytics, paper链接 , code, code , blog , paper-biji

AFM（Attentional FM）= LR + Embedding + MLP + Attention

Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks, paper链接 ,
code1-tf ,
code2-torch ,
code3 ,
code4-tf , blog , paper-biji

WDL（Wide & Deep）= LR + Embedding + MLP

【github】 wide&deep, ichuang, tensorflow
【github】wide&deep, jrzaurin, keras
【github】wide&deep, jorahn, keras
【github】wide&deep, zhougr1993, tensorflow
【github】wide&deep, lambdaji, tensorflow
【API】wide&deep, tensorflow, tf-API
【Paper】Wide & Deep Learning for Recommender Systems

, paper链接 , code , blog , paper-biji

YouTube Recommendation DNN

Deep Neural Networks for YouTube Recommendations paper链接 , code , blog , paper-biji

DIN（Deep Interest Network）= Embedding + MLP + Attention

Deep Interest Network for Click-Through Rate Prediction (阿里提出)
paper链接 , code , blog , paper-biji

MIND

Multi-Interest Network with Dynamic Routing for Recommendation at Tmall(阿里提出)
paper链接
 paper-biji

DeepCTR = CNN + Embedding + MLP

【Paper】Deep CTR Prediction in Display Advertising

DCN （Deep & Cross Network） = Embedding + ResNet + LR

Deep & Cross Network for Ad Click Predictions paper链接 , code , blog , paper-biji
Deep Crossing: Web-Scale Modeling without Manually Crafted Combinatorial Features paper链接 , code , blog , paper-biji

Deep Embedding Forest = Embedding + Forest

Deep Embedding Forest: Forest-based Serving with Deep Embedding Features , paper链接
Deep Forest: Towards An Alternative to Deep Neural Networks , paper链接 code , blog , paper-biji

xDeepFM

【Paper】https://arxiv.org/abs/1803.05170
, paper链接 , code , blog , paper-biji

其他

【Paper】https://www.ijcai.org/proceedings/2018/0277.pdf

强化学习在推荐系统中的应用

DRN

【Paper】DRN: A Deep Reinforcement Learning Framework for News Recommendation

【Paper】Deep Reinforcement Learning for Page-wise Recommendations

Multi-Task Learning
An Overview of Multi-Task Learning in Deep Neural Networks∗

SEMAX: Multi-Task Learning for Improving Recommendations

Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics

Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces

Improving Entity Recommendation with Search Log and Multi-Task Learning

Perceive Your Users in Depth: Learning Universal User Representations from Multiple E-commerce Tasks

Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts

多样性
Tagging like Humans: Diverse and Distinct Image Annotation （行列式点过程）

多目标

孪生网络
【ICML-2015 workshop】Siamese Neural Networks for One-shot Image Recognition

【NIPS-2005】Learning a Similarity Metric Discriminatively, with Application to Face Verificatio

【CVPR-2015】FaceNet: A Unified Embedding for Face Recognition and Clustering

Graph Embedding
【2017】graph2vec: Learning Distributed Representations of Graphs

【Paper】Graph Convolutional Neural Networks for Web-Scale Recommender Systems

【Paper】Graph Convolutional Matrix Completion

《Graph Neural Networks: A Review of Methods and Applications》

《A Comprehensive Survey on Graph Neural Networks》

深度学习

Attention is all your need , paper链接 , paper-biji
Variation Autoencoder Based Network Representation Learning for Classification , paper链接 , paper-biji

Network Embedding/Graph Embedding

DeepWalk: Online Learning of Social Representations , paper链接 , code , zhihu , paper-biji
node2vec: Scalable Feature Learning for Networks , paper链接 , code , blog , paper-biji
Line: Large-scale information network embedding , paper链接 , code , blog , paper-biji
Graph embedding techniques, applications, and performance: A survey , paper链接 , code , blog , paper-biji
SDNE模型（Structure Deep Network Embedding） , paper链接 , paper-biji
Semi-Supervised Classification with Graph Convolutional Networks , paper链接 , code , blog , paper-biji
Fast Network Embedding Enhancement via High Order Proximity Approximation , paper链接 , paper-biji
CANE: Context-Aware Network Embedding for Relation Modeling , paper链接 , paper-biji
Deep Neural Networks for Learning Graph Representations , paper链接 , paper-biji
Geometric deep learning: going beyond euclidean data , paper链接 , paper-biji

NLP 相关文本生成系列

Generative Adversarial Nets 普通对抗网络 Ian Goodfellow∗
GANs for Sequence of Discrete Elements with the Gumbel-softmax Distribution
Generating Text via Adversarial Training-NIPS2016
Adversarial Feature Matching for Text Generation
SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient
MaskGAN: Better Text Generation via Filling in the-ICLR 2018
Long Text Generation via Adversarial Training with Leaked Information-AAAI 2018

NLP 相关 vector 系列

[github] word2vec

Distributed Representations of Words and Phrases and their Compositionality , paper链接 , paper-biji
word2vec Parameter Learning Explained , paper链接 , paper-biji
A Neural Probabilistic Language Model , paper链接 , paper-biji
Efficient Estimation of Word Representations in Vector Space , paper链接 , paper-biji
秒懂词向量Word2vec的本质 , zhihu
word2vec 笔记 , csdn
Bag of Tricks for Efficient Text Classification（fasttext）
Distributed Representations of Sentences and Documents , paper链接 , paper-biji
Topic2Vec: Learing Distributed Representations of Topics , paper链接 , paper-biji
A Hierarchical Neural Autoencoder for Paragraphs and Documents , paper链接 , paper-biji
Convolutional Neural Networks for Sentence Classification , paper链接 , paper-biji
E-commerce in Your Inbox: Product Recommendations at Scale , paper链接 , paper-biji
Supervised Learning of Universal Sentence Representations from Natural Language Inference Data , paper链接 , paper-biji
Deep contextualized word representations（ELMo） , paper链接 , blog
Improving Language Understanding by Generative Pre-Training（GPT) , [https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf) , paper-biji
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding（BERT） , paper链接 , code paper-biji
Language Models are Unsupervised Multitask Learners（GPT-2） , paper链接 , code , blog , paper-biji