DOC | Quick Start | 中文
PaddleFL is an open source federated learning framework based on PaddlePaddle. Researchers can easily replicate and compare different federated learning algorithms with PaddleFL. Developers can also benefit from PaddleFL in that it is easy to deploy a federated learning system in large scale distributed clusters. In PaddleFL, several federated learning strategies will be provided with application in computer vision, natural language processing, recommendation and so on. Application of traditional machine learning training strategies such as Multi-task learning, Transfer Learning in Federated Learning settings will be provided. Based on PaddlePaddle's large scale distributed training and elastic scheduling of training job on Kubernetes, PaddleFL can be easily deployed based on full-stack open sourced software.
Data is becoming more and more expensive nowadays, and sharing of raw data is very hard across organizations. Federated Learning aims to solve the problem of data isolation and secure sharing of data knowledge among organizations. The concept of federated learning is proposed by researchers in Google [1, 2, 3]. PaddleFL implements federated learning based on the PaddlePaddle framework. Application demonstrations in natural language processing, computer vision and recommendation will be provided in PaddleFL. PaddleFL supports the current two main federated learning strategies[4]: vertical federated learning and horizontal federated learning. Multi-tasking learning [7] and transfer learning [8] in federated learning will be developed and supported in PaddleFL in the future.
- Horizontal Federated Learning: Federated Averaging [2], Differential Privacy [6], Secure Aggregation[11]
- Vertical Federated Learning: Two-party training with PrivC[5], Three-party training with ABY3 [10]
There are mainly two components in PaddleFL: Data Parallel and Federated Learning with MPC (PFM).
-
With Data Parallel, distributed data holders can finish their Federated Learning tasks based on common horizontal federated strategies, such as FedAvg, DPSGD, etc.
-
PFM is implemented based on secure multi-party computation (MPC) to enable secure training and prediction. As a key product of PaddleFL, PFM intrinsically supports federated learning well, including horizontal, vertical and transfer learning scenarios. Users with little cryptography expertise can also train models or conduct prediction on encrypted data.
In Data Parallel, the whole process of model training is divided into two stages: Compile Time and Run Time. Components for defining a federated learning task and training a federated learning job are as follows:
-
FL-Strategy: a user can define federated learning strategies with FL-Strategy such as Fed-Avg[2]
-
User-Defined-Program: PaddlePaddle's program that defines the machine learning model structure and training strategies such as multi-task learning.
-
Distributed-Config: In federated learning, a system should be deployed in distributed settings. Distributed Training Config defines distributed training node information.
-
FL-Job-Generator: Given FL-Strategy, User-Defined Program and Distributed Training Config, FL-Job for federated server and worker will be generated through FL Job Generator. FL-Jobs will be sent to organizations and federated parameter server for run-time execution.
-
FL-Server: federated parameter server that usually runs in cloud or third-party clusters.
-
FL-Worker: Each organization participates in federated learning will have one or more federated workers that will communicate with the federated parameter server.
-
FL-scheduler: Decide which set of trainers can join the training before each updating cycle.
For more instructions, please refer to the examples
PaddleFL MPC implements secure training and inference tasks based on the underlying MPC protocol like ABY3[10] and PrivC[5], which are high efficient multi-party computing model. In PaddeFL, two-party federated learning based on PrivC mainly supports linear/logistic regression and DNN model. Three-party federated learning based on ABY3 supports linear/logistic regression, DNN model, CNN model and FM.
In PaddleFL MPC, participants can be classified into roles of Input Party (IP), Computing Party (CP) and Result Party (RP). Input Parties (e.g., the training data/model owners) encrypt and distribute data or models to Computing Parties (There exist three computing parties in ABY3 protocol while two computing parties in PrivC protocol). Computing Parties (e.g., the VM on the cloud) conduct training or inference tasks based on specific MPC protocols, being restricted to see only the encrypted data or models, and thus guarantee the data privacy. When the computation is completed, one or more Result Parties (e.g., data owners or specified third-party) receive the encrypted results from Computing Parties, and reconstruct the plaintext results. Roles can be overlapped, e.g., a data owner can also act as a computing party.
A full training or inference process in PFM consists of mainly three phases: data preparation, training/inference, and result reconstruction.
-
Private data alignment: PFM enables data owners (IPs) to find out records with identical keys (like UUID) without revealing private data to each other. This is especially useful in the vertical learning cases where segmented features with same keys need to be identified and aligned from all owners in a private manner before training.
-
Encryption and distribution: PFM provides both online and offline data encryption and distribution solutions. If users choose the offline data sharing scheme, data and models from IPs will be encrypted using Secret-Sharing[9], and then be sent to CPs, via directly transmission or distributed storage like HDFS. If users adopt the online solution, IPs encrypt and distribute data and models online at the beginning of the training phase. Each CP can only obtain one share of each piece of data, and thus is unable to recover the original value in the Semi-honest model.
A PFM program is exactly a PaddlePaddle program, and will be executed as normal PaddlePaddle programs. Before training/inference, user needs to choose a MPC protocol, define a machine learning model and their training strategies. Typical machine learning operators are provided in paddle_fl.mpc
over encrypted data, of which the instances are created and run in order by Executor during run-time.
For more information of Training/inference phase, please refer to the following doc.
Upon completion of the secure training (or inference) job, the models (or prediction results) will be output by CPs in encrypted form. Result Parties can collect the encrypted results, decrypt them using the tools in PFM, and deliver the plaintext results to users (Currently, data sharing and reconstruction can be supported in both offline and online modes).
For more instructions, please refer to mpc examples
We highly recommend to run PaddleFL in Docker
#Pull and run the docker
docker pull paddlepaddle/paddlefl:1.1.2
docker run --name <docker_name> --net=host -it -v $PWD:/paddle <image id> /bin/bash
#Install paddle_fl
pip3 install paddle_fl
If you want to compile and install from source code, please click here to get instructions.
If you are using the gloo communication model, you will need Redis. We also provide a stable Redis installation package for download.
wget --no-check-certificate https://paddlefl.bj.bcebos.com/redis-stable.tar
tar -xf redis-stable.tar
cd redis-stable && make
kubectl apply -f ./python/paddle_fl/paddle_fl/examples/k8s_deployment/master.yaml
Please refer K8S deployment example for details
You can also refer K8S cluster application and kubectl installation to deploy your K8S cluster
FL-mobile is a framework integrated algorithm simulation , training and deployment. The simulator is part of FL-mobile.
The design purpose of the simulator is to simulate the actual cooperated training among multiple mobile terminal devices online. FL-mobile simulates multiple mobile terminal devices online on the server to verify the effect of algorithms rapidly. The advantages of the simulator are as follows:
- Support training for single machine and distributed networks
- Support training for common open source datasets
- Support private and shared parameters in the models, the private parameters do not participant in the global update
-
Vertical Federated Learning will support more algorithms.
-
Add K8S deployment scheme for PFM.
-
FL mobile simulator will be open sourced in following versions.
[1]. Jakub Konečný, H. Brendan McMahan, Daniel Ramage, Peter Richtárik. Federated Optimization: Distributed Machine Learning for On-Device Intelligence. arXiv preprint 2016
[2]. H. Brendan McMahan, Eider Moore, Daniel Ramage, Blaise Agüera y Arcas. Federated Learning of Deep Networks using Model Averaging. arXiv preprint 2016
[3]. Jakub Konečný, H. Brendan McMahan, Felix X. Yu, Peter Richtárik, Ananda Theertha Suresh, Dave Bacon. Federated Learning: Strategies for Improving Communication Efficiency. arXiv preprint 2016
[4]. Qiang Yang, Yang Liu, Tianjian Chen, Yongxin Tong. Federated Machine Learning: Concept and Applications. ACM Transactions on Intelligent Systems and Technology 2019
[5]. Kai He, Liu Yang, Jue Hong, Jinghua Jiang, Jieming Wu, Xu Dong et al. PrivC - A framework for efficient Secure Two-Party Computation. In Proc. of SecureComm 2019
[6]. Martín Abadi, Andy Chu, Ian Goodfellow, H. Brendan McMahan, Ilya Mironov, Kunal Talwar, Li Zhang. Deep Learning with Differential Privacy. In Proc. of CCS 2016
[7]. Virginia Smith, Chao-Kai Chiang, Maziar Sanjabi, Ameet Talwalkar. Federated Multi-Task Learning. In Proc. of NIPS 2017
[8]. Yang Liu, Tianjian Chen, Qiang Yang. Secure Federated Transfer Learning. IEEE Intelligent Systems 2018
[9]. https://en.wikipedia.org/wiki/Secret_sharing
[10]. Payman Mohassel and Peter Rindal. ABY3: A Mixed Protocol Framework for Machine Learning. In Proc. of CCS 2018
[11]. Aaron Segal Antonio Marcedone Benjamin Kreuter Daniel Ramage H. Brendan McMahan Karn Seth K. A. Bonawitz Sarvar Patel Vladimir Ivanov. Practical Secure Aggregation for Privacy-Preserving Machine Learning. In Proc. of CCS 2017