Eigen

Eigen. Eigen means "own". Let's see, EIGEN is a platform that hosts DAOs that mainly focus on training AI models. Our team was really fascinated by the concept of DATA DAOs, the scope for innovative decentralization seemed evident, after reading the prompts that were also provided by ETH Global. We started brainstorming within the team and started thinking about how data is available for users in Web2. Platforms like Kaggle and Usertesting were spoken about. These platforms give open-source data to users but, how can Web3 make this better? Where is the problem being faced? We did have some ideas in our brainstorming session 2 but we wanted to validate them.

We browsed through some projects that were implemented by fellow participants in the last FVM hackathon. We got highly inspired, and our doubts about some of the ideas were also validated because some of them had already been solved by them. So we went on to look for more ideas.

Our brainstorming session 3 went deeper into all the use cases for which people need data and if that data could be provided by volunteers around. The volunteers in turn could also be rewarded for the same. Some of the use cases were :

• Companies need real-time user feedback when they want to test their products
• Researchers always want real-time data about the topics being researched. For eg, air quality in particular areas, and user data. But the question was: will users be comfortable sharing it? which made us drop that idea too.
• Lastly, and the one we went ahead with Train AI models.

Many people/teams craft AI concepts but one thing that stops them from execution was the long hours of data gathering. Entropy Labs has always faced the same problem.

But what if the data contribution could be incentivized up to a level where the contributor could also get a royalty split when the model is used?

This came to be known as EIGEN. Own your contribution.

We crafted a heavy plan for our build stage because there are a lot of actors in the equation. The platform was divided into different flows with 4 actors DAO Owners Storage Providers Volunteers/DAO Members End-user (the user who will run the AI models)

Different flows;- The owner creates a DAO Volunteers request to contribute Volunteers/DAO Members interact with the DAO The owner checks data to run the model Storage Providers claim contributions to store Output is used by end users

As we grow you and I can only imagine the type of models people can make in collaboration! 2 of my favorites -

• Create a repository of all the endangered animals in order to remember
• Collect emojis to create the next best set of emojis

ALL IN CONTRIBUTION. WHILE YOU OWN YOUR SPLIT.

Know more about how it works in the video uploaded by us.

Experience POV - To craft a flow that would explain the story was tough. The challenge was to make it look fun and not intimidating, giving a community vibe to it. Keep the interactions familiar but also exciting. In all, not make it look like work but fun.

Technical POV - This ReactJs powered web application interacts with smart contracts deployed on the Hyperspace testnet. We have used wagmi to connect the wallet and interact with the deployed smart contracts. DAOs can be created using the DAO Factory contract. Once a user joins a specific DataDAO they get assigned a SBT which enables them to participate in the DAO. Users can make contributions to the dataDAO, each contribution is tokenized and users are given contribution SBTs. Storage providers are incentivized to store these data files on the filecoin network. The admin can ingest the data associated with this DataDAO to a Bacalhau job which will fine-tune the specific model associated with DataDAO. Once the data is ingested the model is ready for use. Users can pay and obtain a Model access SBT which enables them to generate AI images calling a bacalhau job using an API.