This article was first published on Stories by Cortex Labs on Medium
With the launch of the Cortex MainNet at the end of June, the Cortex team has enabled the execution of AI smart contracts on the blockchain. For the first time, AI models can be realistically executed on the blockchain in a decentralized, transparent and secure way. In addition to enabling on-chain AI inference, the Cortex infrastructure also encourages machine learning developers to upload their models to the blockchain for all kind of DApp utilities, aiming to create a decentralized open-sourced AI ecosystem.
To connect AI smart contracts to real-world data, a trustworthy middleware (an oracle) is needed. Using centralized oracles provides a single point of failure and defeats the purpose of running smart contracts on the blockchain; therefore, Cortex has chosen to integrate Chainlink, a highly reliable decentralized oracle network to guarantee the integrity of input data used in the execution of smart contracts as well as the outbound execution results on Cortex. With Chainlink, Cortex can make its trustless AI inferencing ability available to a wide variety of chains and unlock access to a lot more data.
Cortex deeply values the integrity, security, and decentralization of smart contracts, already striving to ensure them by utilizing an ASIC-resistant PoW algorithm. The integration with Chainlink further expands this integrity to the entire smart contract execution process, covering the security of inbound external data and outbound results.
AI smart contracts on Cortex can now pull in trusted real-world data and use them in the execution of on-chain AI inference. As an example in DeFi, if we want to create an on-chain Robo-advisor AI that makes investment for you based on the market data of BTC, we can now trust that the whole process is securely decentralized, as the program pulls in market data from an external API via Chainlink, uses that data for executing ...
To keep reading, please go to the original article at:
Stories by Cortex Labs on Medium