We’re excited to announce the launch of the BEB Score experiment. Imagine it as a rating system, similar to a credit score, but specifically designed for addresses. It is based on an open-sourced machine learning (ML) scoring model that anyone can utilize. Our goal is to assess the credibility of an address in a decentralized and unbiased manner. To accomplish this, the final score is calculated by averaging ratings from various sources, known as reputation providers. Anyone can create an NFT at no cost to register as a provider.

You can get started by checking an address’s score at beb.xyz/reputation. Developers interested in accessing the hosted address score API should contact us.

In this post, we'll delve into the BEB Score, how it works, and the possibilities it creates to establish an onchain reputation system.

What is a decentralized reputation score?

Let's understand the concept of a decentralized reputation score by comparing it to a familiar centralized score, like your credit score.

In a centralized system, selected providers report your history to a credit bureau, which uses a scoring system like FICO to compute your final score.

In a decentralized system, anyone can join the network and become a provider without needing permission. The list of providers is stored in a smart contract, not controlled by any centralized entity. Each provider can publish scores for addresses on the network. The final score is then calculated by different groups called aggregators, which average the scores from various providers.

Why is a decentralized address score useful?

A reliable on-chain reputation score has been a hot topic in recent times. Achieving this goal requires an open system that welcomes anyone to participate as a provider and contribute to the scoring process. Emphasizing accuracy and impartiality in the address score, while maintaining decentralization, ensures a fair and trustworthy assessment for all.

Why choose a decentralized reputation score over a one-size-fits-all score? We believe in giving users the power to create their own score aggregator, which combines scores from various providers. This allows different apps (or Dapps) to pick and choose the scores they need, ensuring a fairer and more reliable assessment. For example, a social app may not require an address's financial score as much as a Defi app would. This flexibility ensures that everyone gets the most relevant and accurate information for their needs.

How does the protocol work?

We propose a simple, decentralized network, where providers register in an on-chain registrar:

• The base score comes from a downloadable, self-hostable pre-trained ML model

• Reputation provider information is stored on a fully on-chain registrar similar to DNS, where unique NFTs represent each provider

• Providers can register their own scoring server URIs onchain, ensuring that anyone can build a scoring aggregator

How is the final score calculated? How is the BEB Score calculated?

The final reputation score is calculated by score aggregators, which take an average of scores from multiple providers. This approach ensures a well-rounded evaluation of an individual's reputation.

Think of reputation providers as separate websites that contribute their unique perspectives to the scoring process. Anyone can become a reputation provider and develop their own algorithm, from straightforward community-based rewards to intricate financial risk assessments.

For now, the BEB Score focuses on the base machine learning model as its primary component.

How do I become a reputation provider?

It's easy to become a reputation provider. First, create an NFT (e.g. myprovider.beb), which is used to register your provider's host server in an on-chain registry.

Currently, you'll need to self-host the server with the ML scoring model to offer custom scores. Simply visit the open-source server implementation and copy the code.

We're actively developing a managed provider host solution that includes a set of user-friendly APIs, allowing you to dynamically query and update an address's score. If you're interested, please get in touch, and we'd love to hear your specific needs.

I am a developer, what's next?

Developers interested in using integrating BEB Score have two options: download and self-host the ML model or contact us to access the hosted API. You can also start exploring right away by checking an address's score at beb.xyz/reputation.

Connect with us

If you appreciate what we're doing, think about registering for a provider NFT! This enables you to contribute your own scores to various addresses and help the network grow.

We're thrilled to pave the way for a decentralized yet reliable method of evaluating addresses. Stay up-to-date by following our Twitter and joining the conversation on our forum!