Federated AI Ecosystem

Collaborative Learning and Knowledge Transfer with Data Protection


  • Security & Compliance

    Meet security and
    compliance requirements;
    Ensure data privacy and
    model security

  • Joint & Expansion

    Connect with business partners
    of various industries;
    Exploit and extend
    the value of data

  • Promotion & Empowerment

    Sustainable and intelligent
    incentive mechanisms;
    Stable and win-win
    business ecosystem


  • Data Isolation

    No data leakage to the outside

  • Lossless

    Federated learning preserves model quality

  • Equivalence

    Equivalence of participants status

  • Mutual Benefits

    Mutual benefits for participants


Federated learning is a machine learning framework that helps multiple organizations effectively and collaboratively use data and modeling in accordance with user privacy protection, data security, data confidentiality and government regulations

White Paper

Depending on data distribution, we separate Federated Learning into three categories

  • The overlap of features (X1, X2, …) is large, whereas the overlap of users (U1, U2, …) is small

  • The overlap of users (U1, U2, …) is large, whereas the overlap of features (X1, X2, …) is small

  • The overlap of users (U1, U2, …) and the overlap of features (X1, X2, …) are both small

A Demo of Federated Learning System:


Federated learning can be used in many industries, including financial services, logistics, supply chain, operators, health care, etc. We aim to solve some of the most challenging business problems through joint modeling and platform services.

  • Small Business Loans

    Internet finance platforms, micro loan companies and banks which are qualified for small business loans, may collaboratively explore a federated learning framework for risk management and share the final prediction model with participating organizations.

    • Ensure Data Safety and No Data Leakage

    • Improve Model Performance and Mutual Benefit

  • Anti Money Laundering

    Financial institutions may not have enough money laundering cases to build an efficient machine learning model. Through federated learning framework, they can collaboratively build a detection model without sharing their data, improve accuracy and efficiency of the anti money laundering system.

    • Meeting Requirements on Financial Data Security and Compliance

    • Joint Modeling to Solve Few Labels Problem

  • Intelligent Anomaly Detection

    The federated modeling can guarantee the data security between different equipment operators, overcome the key issues of less labels and low prediction accuracy of single equipment, and improve anomaly detection accuracy and operational efficiency globally.

    • Improve The Anomaly Detection Accuracy

    • Enhance Operational Efficiency and Reduce Costs

  • Federated Image Recognition

    Medical institutions do not share their data, and users are reluctant to disclose personal disease data. The application of federated learning on medical images can solve the problem of medical data isolation and improve AI applications in the computer-aided diagnosis field.

    • Enhances The Ability of Medical Institutions to Applying Data

    • Federated Multi-party Data and Improve Detection Key Metrics


Federated AI Technology Enabler (FATE)

FATE is an open-source project initiated by Webank's AI Department to provide a secure computing framework to support the federated AI ecosystem. It implements secure computation protocols based on homomorphic encryption and multi-party computation (MPC). It supports federated learning architectures and secure computation of various machine learning algorithms, including logistic regression, tree-based algorithms, deep learning and transfer learning.

  • It helps academic researcher develop algorithm prototype rapidly
  • It provides a simple and effective solution to develop applications quickly, and supports development and application in various scenarios
  • With flexible architecture, users can easily deploy computing work to multiple platforms (CPU, GPU) and devices (desktop, server, mobile, etc.)


  • Analysts advise widespread AI use


  • How to solve the big data plight of AI? federal transfer learning




The Federated AI Ecosystem project is initiated by WeBank's AI Department. WeBank is a technology driven internet bank based in Shenzhen, China.  The team's aim is to develop and promote advanced AI technology that preserves data security, data confidentiality and user privacy.


"Build a Federated AI Ecosystem based on federated learning and transfer learning. So that our partners can fully exploit their data values and promote vertical applications. It can carry on AI empowerment to related entities, and enhance their modeling technologies and abilities"


  • Open Source Technology

    Accelerate the open-sourcing of federated transfer learning, and develop more applications

  • Standards and Guidelines

    Formulate architectural framework and application guideline of federated learning, and facilitate industry cooperation

  • Multi-Party Consensus Mechanisms

    Encourage more institutions to participate and establish a multi-party consensus mechanism based on technologies like blockchain

  • Vertical Fields Applications

    Focused on vertical field data and scenarios, built a new business model and ecology

  • email
  • backtop