FLock.io raises $6 million for decentralized blockchain AI training platform

FLock.ioa startup that provides full-stack distributed training and fine-tuning of artificial intelligence models, today raised $6 million in seed funding to build a blockchain-based platform that enables collaborative creation of AI models. announced.

The round was led by Lightspeed Faction and Targus Capital, with participation from DCG, OKX Ventures, and Volt Capital. The company said it will use the new funding to continue developing its machine learning and decentralized training platform to put AI governance in the hands of users and avoid centralized corporate oversight.

“FLock allows anyone to contribute to AI development by contributing compute, data, and training scripts. The barriers to participation are now much lower,” said CEO Jiahao Sun. says Mr. “The result is a community-owned model built by the many, not just the few, and where data contributors are rewarded fairly.”

FLock provides a decentralized machine learning platform that encourages users to contribute data in a decentralized manner to train and fine-tune AI models. The company says this approach enables scalability, reduces costs, and improves security.

The platform uses a concept known as “federated learning.” This is a method of training AI models where the training data is distributed across individual devices or servers, rather than being sent to a centralized server. Using this model, the data you submit is not exposed to other users or the Internet at large, reducing the chance of your personal data being leaked or misused.

Unlike large-scale, centralized commercial or open-source model training, FLock’s implementation uses distributed learning, where local models are trained on edge nodes and then synchronized with the global model. Users will be awarded “FLock Points” for participating in co-creation and collaboration for AI model customization, knowledge contribution, evaluation, and quality assurance. These points open up the possibility of exclusive rewards in the FLock ecosystem, including future crypto airdrops, perks, and more.

FLock uses zero-knowledge proofs with federated training data in parallel with homomorphic encryption. This allows data to be shared without revealing the information to other parties sending training data to the model. The combination of blockchain technology provides another decentralized ability to robustly track transactions. It also enables secure multiparty computation when accessing data. This prevents your data from being shared if you do not authorize it.

According to FLock, the platform is model agnostic. Developers can choose between simple statistical models or complex neural networks, including large-scale language models such as OpenAI’s ChatGPT and art generation models similar to Stable Diffusion.

“With the rapid growth of AI comes growing data privacy and regulatory concerns,” said Clara Bo, trading partner at Lightspeed Faction. “By combining blockchain and federated learning technologies, FLock aims to solve this problem and enable broader participation in AI model training.”

In the future, FLock plans to enable distributed hosting by allowing users to contribute to distributed computing pools, vector database hosting, and model inference. This significantly increases the scalability and functionality of your AI model storage and computing environment.

FLock itself is designed to be modular and can also be connected to decentralized hosting networks. Ionet Co., Ltd., Genshin and ritual – This allows AI models to access blockchain-based applications.

Image: Production Perrig/Shutterstock