AI Applications

Enabling Diverse AI Applications

FedML Ecosystem facilitates federated learning research and productization in diverse application domains. With the foundational support from FedML Core Framework, it supports FedNLP (Natural Language Processing), FedCV (Computer Vision), FedGraphNN (Graph Neural Networks), and FedIoT (Internet of Things).


Due to privacy regulations, the small amount of datasets in each hospital are not allowed to be moved to a central cloud for data sharing/large-model training for disease diagnosis, while the data on each node is far from enough to train a deep learning model.
FedML provides a solution which not only includes a platform to preserve the data privacy to be in compliance with the regulations without complicated deployment but also a paradigm to boost the accuracy.


Financial data are not sharable among banks, Regulation Institutions, and other financial institutions . By leveraging FedML, customers can enhance Risk Control/Anti-fraud in the scenarios:
1. risk control before/after loan
2. credit report generation/evaluation
3. fraud prevention/recognition
4. blacklist query
5. regulation within a financial institution/over risk control institutions.


FedML can help Web3/blockchain based companies to train recommendation systems by leveraging the decentralized data at the edge, while protecting privacy and security for digital assets.

Autonomous Driving

Autonomous driving requires low latency, real-time response to operate correctly, therefore it has a great need for distributed data processing at the edge.
By using FedML, training for obstacle/lane detection etc. can happen at the edge to avoid huge data offloading and storage on the cloud, reduce the time to train and update the model back to vehicles, and preserve users’ private data locally to compy with the regulations.

Cloud Computing

Multi-cloud/Sky Computing is the next generation of cloud computing, so each cloud service provider can make the most of its strengths and leverage other providers’ strengths to make up for the shortcomings. Cloud service providers can leverage FedML to empower their platforms the capability of collaborative federated learning over multiple clouds from anywhere at any scale, without concerning moving the data, and simultaneously maintain the advantage of Multi-cloud/Sky Computing.