Cortex aims to build a truly decentralized AI autonomous system, offering state-of-the-art machine learning models on the blockchain. Users can leverage smart contracts on the Cortex blockchain to make inferences from these models. Another goal of Cortex is to establish a machine learning platform where users can post tasks and submit AI dApps.

Project Highlights

Cortex introduces the concept of AI DApps, or artificial intelligence distributed applications. By integrating AI into smart contracts, Cortex enables:

• Information services: personalized recommendation systems, search engines, news writing services.
• Finance: credit scoring, smart investment advisory.
• AI assistants: automated Q&A, industry knowledge graphs, voice synthesis, facial attribute prediction.
• Simulation environments: autonomous driving, Go (game), and other reinforcement learning applications.

Cortex allows users to write machine learning programs on the blockchain and submit interactions that depend on other contracts. This is an innovative feature of Cortex, but its unclear whether "other contracts" refer to smart contracts on Ethereum or contracts on different blockchains. If the latter, it would be challenging to implement given the current limitations of cross-chain smart contract capabilities.

Architecture:
Data Layer: The Cortex blockchain does not store models and data directly but rather their hash values. A key-value storage system without a chain is used. Once enough copies are propagated across the network, the Cortex blockchain can incorporate new models and data.

Network Layer: Cortex will add an inference instruction to smart contracts, enabling them to use models hosted on the Cortex blockchain during interactions.

Consensus Layer: Proof-of-Work (PoW) is used to further reduce the disparity between CPU and GPU acceleration ratios.

Incentive Layer: Cortex proposes a collaborative incentive mechanism allowing anyone to submit and optimize models on the Cortex blockchain, with contributors being rewarded for their contributions.

Contract Layer: Cortex will add an inference instruction to smart contracts, making it possible to use models stored on the Cortex blockchain within smart contracts.

Application Layer, Platform Layer: Cortex also aims to create a machine learning platform where users can post tasks, submit models, perform inferences by invoking smart contracts, and develop their own AI DApps (artificial intelligence decentralized applications).

Use Cases

Information services, financial services, AI assistants, simulation environments, and more.