Federation Partners
Federation Partners are essential participants in the Brainstems ecosystem, responsible for contributing valuable data to fuel collaborative AI model development. Federation Partners securely encrypt and transfer their datasets to computational nodes within the Brainstems network. These datasets are then used to train foundational AI models through Decentralized Federated Learning (DFL). The process of data contribution involves several key steps, including data encryption, transfer, and validation.
This collaborative approach to model development enables Federation Partners to leverage diverse datasets while maintaining ownership and control over their data.
In addition to contributing data, Federation Partners may also participate in the governance of the Brainstems protocol and contribute to the curation of high-quality models. This participation helps prioritize the selection of models for training and ensures that the resulting intelligence meets the highest standards of accuracy and relevance.
Overall, Federation Partners play a critical role in advancing the capabilities of Brainstems by providing the raw material necessary for AI model collaboration, training, and development. Through their contributions, Federation Partners drive innovation, collaboration, and value creation within the Brainstems ecosystem.
FAQs and Technical Deep Dive
What are the tradeoffs between a running a light node and a full node?
Great question - to start, it is true that any federation partner can run either a light node or a full node, but only light nodes can be run by fed partners.
Why is that? Well, thats a question inside a question.
Light nodes provide a partners a toolkit with the bare minimum to get started within the Brainstems ecosystem, while full nodes must have the computation capacity to handle intensive training tasks, including processing encrypted data, generating model updates, and creating Zero-Knowledge Proofs (ZKP) to validate the training process.
So, its really a decision based on financial resources and technical know-how. We did not envision our clientelle having the time, money or resources to dedicate CPUs/GPUs to model training, which is why we built our light node framework in the first place!
Where is this Technical Deep Dive you speak of?
Hit our Github! That's the spot for the latest and greatest when it comes to Brainstems tech. We do our best to keep our Gitbook and Github in sync, but everything hits Github first :)