SUNNYVALE, Calif.--(BUSINESS WIRE)--FedML today announced $6 million in funding to spearhead a “collaborative AI” movement that empowers companies and developers to work together on machine learning tasks by sharing data, models and compute resources – fueling waves of AI innovation beyond the largest technology companies. To meet that need, FedML has built an open-source community, enterprise platform, and software tools that make it easier to train, deploy and customize machine learning models at scale across edge and cloud nodes.
FedML was founded to create an ecosystem that helps enterprises customize and deploy AI models, including generative AI and other large language models. Many businesses are eager to train or fine-tune AI models on company-specific and/or industry data, so they can deploy AI-powered applications that improve customer service, business automation, content creation, product design, etc. Because that company and industry data is often sensitive, regulated and/or siloed, traditional cloud-based AI training solutions are not suitable for the task.
FedML addresses those challenges using federated learning technology, which enables training of AI models using private or siloed data at the edge without the need to share or move such data (i.e., “learning without sharing”). For example, federated learning would allow a retail, e-commerce or social media company to build models for personalized content without pulling customers’ private data, or enable a healthcare company to develop models for rare disease detection by using scarce datasets spread across many hospitals. In fact, FedML recently announced partnerships with Theta Network and Konica Minolta for both of these applications.
“The future of AI depends on large-scale collaboration,” said Salman Avestimehr, co-founder and CEO of FedML. “We want to create a community that trains, serves and mines the best AI models. For example, we enable data owners to contribute their data to a machine learning task, and they can work with AI developers or training specialists to build a customized machine learning model, and everyone gets rewarded for their contributions.”
Since launching in March 2022 after three years of development, FedML has quickly become a leader in community-driven AI, hosting the top-ranked open source library for federated machine learning, which surpassed TensorFlow Federated from Google in November 2022. FedML also provides an MLOps ecosystem for training, serving, and monitoring machine learning models anywhere at the edge or the cloud, with 1900+ users globally who have deployed FedML over 3500+ edge devices, and performed 6500+ training jobs.
FedML has now raised $6 million in seed and pre-seed funding, led by Camford Capital, along with additional investors Plug and Play Ventures, AimTop Ventures, Acequia Capital, LDV Partners and other undisclosed investors. FedML has also signed 10 enterprise contracts spanning healthcare, financial services, logistics, retail, smart city, generative AI and web3 applications.
“FedML has a compelling vision and unique technology to enable open, collaborative AI at scale,” said Ali Farahanchi, partner at Camford Capital. “Their leadership team combines humility, hard work and perseverance with deep technical capabilities, and they’ve already made strong progress. In a world where every company needs to harness AI, we believe FedML will power both company and community innovation that democratizes AI adoption.”
“FedML’s collaborative AI unlocks unprecedented opportunities in the entertainment industry, when coupled with Theta’s distributed global network of edge nodes,” said Mitch Liu, co-founder and CEO of Theta Network, the leading Web3 blockchain infrastructure for video, media and entertainment. “We’re seeing significant consumer demand for generative text-to-image and text-to-video AI to create new content and new experiences. With FedML technology, it is possible to transform media businesses by offering personalized AI-based experiences, while rewarding users for contributing data, compute and storage resources.”
FedML was co-founded by Avestimehr, a Dean’s professor at USC and the inaugural director of the USC-Amazon Center on Trustworthy AI, and his former PhD student Dr. Chaoyang He, who published several award-winning papers and has more than 10 years R&D experience at Google, Amazon, Facebook, Tencent and Baidu. Over the past four years, Avestimehr and He have worked with nearly 40 collaborators to build FedML’s open source library and commercial software that combines federated learning tools with an industrial-grade MLOps platform and secure data marketplace.
In addition to its breakthroughs in federated learning, FedML believes collaborative AI will be valuable in overcoming the cost and complexity of large-scale AI development. For example, training of GPT-3 would require about $5 million of compute credit over the cloud, and training of more advanced models are often limited to the largest technology companies with massive GPU clusters. And while some progress has been made to simplify how AI models are deployed (e.g. the easy-to-use APIs of HuggingFace), AI training and development is still very complex for many enterprises.
“We allow people to train anywhere and serve anywhere, from edge to cloud, enabling lower-cost and decentralized AI development that’s accessible to everyone,” said Chaoyang He, co-founder and CTO of FedML. “We’re committed to maintaining a strong and vibrant open source community of AI researchers, while also advancing commercial needs for the best and most customized large AI models.”
FedML is a leader in collaborative AI, delivering an open-source library and enterprise software platform to train, deploy and customize machine learning models across edge and cloud nodes at any scale. FedML’s distributed MLOps platform uniquely enables sharing of data, models, and compute resources in a way that preserves data privacy and security. The company hosts the top-ranked GitHub library for federated machine learning, and is currently used by more than 1,900 developers and 10 large enterprise customers spanning multiple verticals.