Open Knowledge Network

Open Knowledge Network

Democratizing Knowledge by Decentralizing Language Models

436
Built at ETHIndia 2023

The problem Open Knowledge Network solves

Our project tackles several key challenges with Distributed computing development:

  • Democratizes Distributed computing creation. By allowing anyone to contribute specialized models, we empower domain
    experts rather than just big tech companies to drive innovation. A teacher could build an education model; an accountant
    could create a tax prediction model.
  • Achieves deeper expertise. Narrow models outperform generalized models in their field. An agriculture model trained on
    crop data will be better at harvest forecasts than a catch-all predictive model.
  • Incentivizes sharing of knowledge. Model builders earn rewards when others use their models, encouraging experts to
    contribute their hard-won knowledge rather than hoarding it.
  • Enables collective intelligence. Rather than relying on a single, centralized system, our distributed network combines
    thousands of specialized contributors into an intelligent whole that far exceeds capabilities of any individual model.
  • Accelerates breakthroughs. With more participants working on models tailored to their field, we will push boundaries of
    what Distributed computing can achieve across every industry, application area and discipline.

By connecting decentralized expertise models rather than generalizing, our project takes innovative approach to developing Distributed computing that is more collaborative, targeted and impactful. We are building future of Distributed computing - please join us! This technology has potential to greatly benefit society by making Distributed computing accessible and meaningful to common people. Let us work together to create positive change.

Challenges we ran into

Here are the 3 key hurdles we faced and how we overcame them:

  1. Communication and routing of data - Different models were not able to talk to each other to pass data around. We solved this issue by using a "middleware" decentralised communication protocol called Wako that enables interoperability between models.

  2. Incentivizing knowledge providers - We had to figure out how to provide right incentives to experts who spent time building specialized models so they would share them on our platform. We designed smart contracts which automatically give rewards to model developers when others start using their models.

  3. Directional flow of data - It was complex to manage how data flows between so many different models. To address this we created detailed flow charts mapping out how each model takes in data, processes it and passes output to the next model in an easy to understand visual way.

By tackling communication barriers through the Wako protocol, by guaranteeing model developer earnings via smart contracts, and by visually mapping model workflows using flow charts - we were able to solve key technology and business hurdles in making decentralized distributed computing using shared knowledge models work effectively.

Tracks Applied (3)

Arbitrum Track

Arbitrum Provides several advantages like rollup and scalability. We have launched OKN Token on their network considerin...Read More

Arbitrum

Waku Track

Our project aligns with the Waku track as it is built on Waku's decentralized communication protocol. In our initiative,...Read More

waku

Alliance Track

Alliance supports innovation and good business model. Our Project being a social-good project we are bringing together 2...Read More

Alliance

Cheer Project

Cheering for a project means supporting a project you like with as little as 0.0025 ETH. Right now, you can Cheer using ETH on Arbitrum, Optimism and Base.

Discussion