Data Providers

Providers are crucial to the Agentic Data Coordination Service (ADCS) ecosystem, supplying the essential data that power AI-driven applications and services. Based on the specific data requirements defined within the Adaptor, they can deliver various types of data. However, there are two primary providers: Classic Data and Inference Data (AI Data). Classic Data and Inference Data, each serving different purposes and use cases within the network.

1. Classic Data

Classic data typically consists of two primary types:

  1. Data Feeds: Provides periodic or scheduled updates at specific intervals.

  2. Data Streams: Delivers continuous, real-time data for applications requiring instant updates.

Currently, the Data Stream is not yet supported, but it will be available soon.

Data feeds: consist of pre-packaged, processed data sets that are delivered at scheduled intervals, such as hourly, daily, or weekly. These feeds typically provide structured data, meaning the data is already organized into a clear format—such as tables, rows, and columns—and is often aggregated to offer insights or summary information. Data feeds are particularly useful when regular updates are needed, but real-time data isn't critical. They are commonly employed in scenarios like market updates, financial reports, or periodic weather forecasts, where the focus is on trends and insights rather than immediate reactions.

Characteristics:

  • Structured and aggregated

  • Delivered at defined intervals (e.g., every hour, day)

  • Used for historical analysis, insights, and trend monitoring

  • Easier to process due to its structured format

Example Providers:

  • CoinMarketCap: Provides cryptocurrency pricing, market cap, and volume data at regular intervals, useful for tracking trends in the digital asset space.

  • Binance: Offers structured historical data feeds for cryptocurrency prices, trading volumes, and order books, often used by traders and analysts.

  • Bloomberg: Supplies pre-packaged financial data at regular intervals, including stock indices, commodities, and market insights.

Workflow for Data Verification

  1. Before data providers can contribute to the ADCS as a Data Hub, they must first complete their registration through the dashboard using their wallet or any authentication method capable of signing the data which ensures secure and verifiable identity management.

  2. After signing and registering the wallet, they can start providing data.

  3. Validators evaluate the data submitted by providers by executing the same processes as the providers. If the results match, it indicates that the provider is performing well and will receive a positive score. The reputation score is determined by the number of accurate submissions made by the provider; higher accuracy contributes to a positive growth in reputation. Conversely, if the results do not align, the provider will be flagged as potentially malicious.

  4. Once validators have compared the data submitted by providers with their own results, they reach a consensus on the accuracy of the submissions. After this validation, proof is generated to demonstrate that the data has been assessed and found to be accurate. This proof includes hashes of the validated data, along with relevant metadata, and is subsequently submitted on-chain to ensure transparency and immutability. By storing this proof on the blockchain, the integrity of the validation process is preserved, allowing for easy verification and accountability within the ADCS framework.

ADCS Reputation System

Benefits for Data Providers

Participating as a Provider within ADCS offers several advantages:

  • Monetization Opportunities - Providers can monetize their data by selling or renting their data feeds and streams within the Intel Market, creating revenue streams from their data assets.

  • Enhanced Trust and Reliability - The ADCS Reputation System ensures that only trusted and reliable data is utilized, increasing the value and trustworthiness of the providers’ data offerings.

  • Scalability and Flexibility - Providers can offer a wide range of data types and formats, catering to diverse applications and industries within the ADCS network. This flexibility allows Data Providers to scale their offerings based on demand and market needs.

By contributing high-quality data, Data Providers play a pivotal role in maintaining a robust and reliable data ecosystem, empowering AI Agents to make informed and accurate decisions across various domains.

2. Inference Data (AI Data)

Inference Data refers to the results or outputs produced by an artificial intelligence (AI) model after processing raw input data. This data can be used for various applications, including decision-making, recommendations, and identifying patterns in large datasets. In the context of AI-powered systems, Inference Data is crucial because it reflects how an AI model interprets and responds to real-world data.

Characteristics:

  • It can come in various formats, including numerical values, categorical labels, probabilities, or even more complex outputs like text or images.

  • Driven by Input Data: It is highly dependent on the input it receives. A slight change in the input can lead to different outputs.

  • Dynamic and Adaptive: Inference Data can change as new input data is processed by the model. AI systems adapt to new patterns, and their output may evolve over time as more data is fed into the model, improving accuracy and predictions.

Providers:

For Inference providers, we currently support the following:

  • Llama

  • Gemini

Upcoming providers include:

  • Anthropic

  • OpenAI

Last updated