The open source community now has access to the source code of Alibaba Group’s worldwide research program Alibaba DAMO Academy’s latest federated learning platform FederatedScope, a complete platform with easy-to-use packages. With all types of Machine Learning (ML) and Artificial Intelligence (AI) on the increase, acquiring training data to construct and advance AI models is becoming more of a focus, as the process may raise privacy concerns.
Federated learning, a method of privacy-preserving computation, has arisen to meet this difficulty. To relieve privacy concerns, intermediate training results – rather than raw user data – are supplied back to the cloud server by coordinating the training of micro-tasks across several end devices. Nonetheless, data analytics and machine learning tasks can be performed across end devices. Discovery, Adventure, Momentum, and Outlook are abbreviated as DAMO.
“By sharing our self-developed federated learning technologies with the open-source community, we hope to promote the research and industrial deployment of privacy-preserving computation in different sectors, such as healthcare and smart mobility that usually involves sensitive user data and requires strict privacy protection practices,” says Bolin Ding, research scientist at Alibaba DAMO Academy.
FederatedScope also offers flexible support and comprehensive tools, such as a large collection of benchmark datasets, well-known model architectures, advanced federated learning algorithms, easy-to-use automatic tuning functionalities, and friendly interfaces, thanks to a newly implemented event-driven framework.
Researchers and developers can use these to swiftly create and customise task-specific federated learning applications in domains such as computer vision, natural language processing, audio recognition, graph learning, and recommendation. The platform also includes cutting-edge technologies for privacy protection, such as differential privacy and multi-party computation, to suit various privacy requirements.
“We believe privacy-preserving computation is an important and essential trend,” added Ding. “Training AI models without compromising privacy is critical and that’s why we have devoted a lot of resources to drive the research of federated learning. We hope that by sharing our source codes and technology platform, we can support global developers in the community and encourage more innovation in this emerging field.”
By 2025, 60% of major organisations are projected to implement one or more privacy-enhancing compute approaches, according to Gartner.