README.md

View on GitHub

(中文文档|简体中文|English)

What is recommendation system ?

Recommendation system helps users quickly find useful and interesting information from massive data.
Recommendation system is also a silver bullet to attract users, retain users, increase users' stickness or conversionn.

Who can better use the recommendation system, who can gain more advantage in the fierce competition.

At the same time, there are many problems in the process of using the recommendation system, such as: huge data, complex model, inefficient distributed training, and so on.

What is PaddleRec ?

A quick start tool of search & recommendation algorithm based on PaddlePaddle
A complete solution of recommendation system for beginners, developers and researchers.
Recommendation algorithm library including content-understanding, match, recall, rank, multi-task, re-rank etc.

Getting Started

Online running

AI Studio Online Running

Environmental requirements

Python 2.7/ 3.5 / 3.6 / 3.7 , Python 3.7 is recommended ,Python in example represents Python 3.7 by default
PaddlePaddle >=2.0
operating system: Windows/Mac/Linux

Linux is recommended for distributed training

Installation

Install by pip in GPU environment

python -m pip install paddlepaddle-gpu==2.0.0

Install by pip in CPU environment

python -m pip install paddlepaddle # gcc8

For download more versions, please refer to the installation tutorial Installation Manuals

Download PaddleRec

git clone https://github.com/PaddlePaddle/PaddleRec/
cd PaddleRec

Quick Start

We take the dnn algorithm as an example to get start of PaddleRec, and we take 100 pieces of training data from Criteo Dataset:

python -u tools/trainer.py -m models/rank/dnn/config.yaml # Training with dygraph model
python -u tools/static_trainer.py -m models/rank/dnn/config.yaml #  Training with static model

Documentation

Background

Introductory tutorial

Advanced tutorial

FAQ

Common Problem FAQ

Acknowledgements

Contributions from external developer

Support model list

Type	Algorithm	Online Environment	Parameter-Server	Multi-GPU	version	Paper
Content-Understanding	TextCnn (doc)	Python CPU/GPU	✓	x	>=2.1.0	[EMNLP 2014]Convolutional neural networks for sentence classication
Content-Understanding	TagSpace (doc)	Python CPU/GPU	✓	x	>=2.1.0	[EMNLP 2014]TagSpace: Semantic Embeddings from Hashtags
Match	DSSM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[CIKM 2013]Learning Deep Structured Semantic Models for Web Search using Clickthrough Data
Match	MultiView-Simnet (doc)	Python CPU/GPU	✓	x	>=2.1.0	[WWW 2015]A Multi-View Deep Learning Approach for Cross Domain User Modeling in Recommendation Systems
Match	Match-Pyramid (doc)	Python CPU/GPU	✓	x	>=2.1.0	[2016]Text Matching as Image Recognition
Recall	TDM	-	✓	>=1.8.0	1.8.5	[KDD 2018]Learning Tree-based Deep Model for Recommender Systems
Recall	FastText	-	x	x	1.8.5	[EACL 2017]Bag of Tricks for Efficient Text Classification
Recall	MIND (doc)	Python CPU/GPU	x	x	>=2.1.0	[2019]Multi-Interest Network with Dynamic Routing for Recommendation at Tmall
Recall	Word2Vec (doc)	Python CPU/GPU	✓	x	>=2.1.0	[NIPS 2013]Distributed Representations of Words and Phrases and their Compositionality
Recall	DeepWalk (doc)	Python CPU/GPU	x	x	>=2.1.0	[SIGKDD 2014]DeepWalk: Online Learning of Social Representations
Recall	SSR	-	✓	✓	1.8.5	[SIGIR 2016]Multi-Rate Deep Learning for Temporal Recommendation
Recall	Gru4Rec (doc)	-	✓	✓	1.8.5	[2015]Session-based Recommendations with Recurrent Neural Networks
Recall	Youtube_dnn	-	✓	✓	1.8.5	[RecSys 2016]Deep Neural Networks for YouTube Recommendations
Recall	NCF (doc)	Python CPU/GPU	✓	✓	>=2.1.0	[WWW 2017]Neural Collaborative Filtering
Recall	GNN	-	✓	✓	1.8.5	[AAAI 2019]Session-based Recommendation with Graph Neural Networks
Recall	RALM	-	✓	✓	1.8.5	[KDD 2019]Real-time Attention Based Look-alike Model for Recommender System
Rank	Logistic Regression (doc)	Python CPU/GPU	✓	x	>=2.1.0	/
Rank	Dnn (doc)	Python CPU/GPU	✓	✓	>=2.1.0	/
Rank	FM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[IEEE Data Mining 2010]Factorization machines
Rank	BERT4REC	-	✓	x	>=2.1.0	[CIKM 2019]BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer
Rank	FAT_DeepFFM	-	✓	x	>=2.1.0	[2019]FAT-DeepFFM: Field Attentive Deep Field-aware Factorization Machine
Rank	FFM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[RECSYS 2016]Field-aware Factorization Machines for CTR Prediction
Rank	FNN	-	✓	x	1.8.5	[ECIR 2016]Deep Learning over Multi-field Categorical Data
Rank	Deep Crossing	-	✓	x	1.8.5	[ACM 2016]Deep Crossing: Web-Scale Modeling without Manually Crafted Combinatorial Features
Rank	Pnn	-	✓	x	1.8.5	[ICDM 2016]Product-based Neural Networks for User Response Prediction
Rank	DCN (doc)	Python CPU/GPU	✓	x	>=2.1.0	[KDD 2017]Deep & Cross Network for Ad Click Predictions
Rank	NFM	-	✓	x	1.8.5	[SIGIR 2017]Neural Factorization Machines for Sparse Predictive Analytics
Rank	AFM	-	✓	x	1.8.5	[IJCAI 2017]Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks
Rank	DMR (doc)	Python CPU/GPU	x	x	>=2.1.0	[AAAI 2020]Deep Match to Rank Model for Personalized Click-Through Rate Prediction
Rank	DeepFM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[IJCAI 2017]DeepFM: A Factorization-Machine based Neural Network for CTR Prediction
Rank	xDeepFM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[KDD 2018]xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems
Rank	DIN (doc)	Python CPU/GPU	✓	x	>=2.1.0	[KDD 2018]Deep Interest Network for Click-Through Rate Prediction
Rank	DIEN (doc)	Python CPU/GPU	✓	x	>=2.1.0	[AAAI 2019]Deep Interest Evolution Network for Click-Through Rate Prediction
Rank	GateNet (doc)	Python CPU/GPU	✓	x	>=2.1.0	[SIGIR 2020]GateNet: Gating-Enhanced Deep Network for Click-Through Rate Prediction
Rank	DLRM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[CoRR 2019]Deep Learning Recommendation Model for Personalization and Recommendation Systems
Rank	NAML (doc)	Python CPU/GPU	✓	x	>=2.1.0	[IJCAI 2019]Neural News Recommendation with Attentive Multi-View Learning
Rank	DIFM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[IJCAI 2020]A Dual Input-aware Factorization Machine for CTR Prediction
Rank	DeepFEFM (doc)	Python CPU/GPU	✓	x	>=2.1.0	[arXiv 2020]Field-Embedded Factorization Machines for Click-through rate prediction
Rank	BST (doc)	Python CPU/GPU	✓	x	>=2.1.0	[DLP-KDD 2019]Behavior Sequence Transformer for E-commerce Recommendation in Alibaba
Rank	AutoInt	-	✓	x	1.8.5	[CIKM 2019]AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks
Rank	Wide&Deep (doc)	Python CPU/GPU	✓	x	>=2.1.0	[DLRS 2016]Wide & Deep Learning for Recommender Systems
Rank	FGCNN	-	✓	✓	1.8.5	[WWW 2019]Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction
Rank	Fibinet	-	✓	✓	1.8.5	[RecSys19]FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Prediction
Rank	Flen	-	✓	✓	1.8.5	[2019]FLEN: Leveraging Field for Scalable CTR Prediction
Multi-Task	PLE (doc)	Python CPU/GPU	✓	✓	>=2.1.0	[RecSys 2020]Progressive Layered Extraction (PLE): A Novel Multi-Task Learning (MTL) Model for Personalized Recommendations
Multi-Task	ESMM (doc)	Python CPU/GPU	✓	✓	>=2.1.0	[SIGIR 2018]Entire Space Multi-Task Model: An Effective Approach for Estimating Post-Click Conversion Rate
Multi-Task	MMOE (doc)	Python CPU/GPU	✓	✓	>=2.1.0	[KDD 2018]Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts
Multi-Task	ShareBottom (doc)	Python CPU/GPU	✓	✓	>=2.1.0	[1998]Multitask learning
Multi-Task	Maml (doc)	Python CPU/GPU	x	x	>=2.1.0	[PMLR 2017]Model-agnostic meta-learning for fast adaptation of deep networks
Re-Rank	Listwise	-	✓	x	1.8.5	[2019]Sequential Evaluation and Generation Framework for Combinatorial Recommender System