What Is Cortex (CTXC)?
In the current blockchain world, a Cortex complete smart contract virtual machine has been widely used and attracted the attention and participation of application developers. As a result, many fields like decentralized finance, crypto art, and decentralized game have made great progress. However, the current blockchain infrastructure requires all nodes to verify and agree on the calculation results, which severely limits the ability of smart contracts.
The existing smart contract languages and virtual machines are limited to writing short programs and accessing very few resources. Yet, machine learning models require a massive demand for computing and storage resources to be applied on the blockchain environment. With the emergence of artificial intelligence services, machine learning plays a huge role in image recognition, natural language processing, pattern recognition and many other fields.
The Cortex project adds AI algorithm support to smart contracts by expanding the underlying instruction set of smart contracts and enhancing the storage layer so that anyone can add AI capabilities to smart contracts. At the same time, the proposed incentive mechanism prompts model contributors to submit and optimize models on Cortex chain and receive rewards.
Cortex Storage Key Points
| Coin Basic | Information |
|---|---|
| Coin Name | Cortex |
| Short Name | CTXC |
| Circulating Supply | 192,960,486.09 CTXC |
| Total Supply | 299,792,458 |
| Source Code | Click Here To View Source Code |
| Explorers | Click Here To View Explorers |
| Twitter Page | Click Here To Visit Twitter Group |
| Whitepaper | Click Here To View |
| Support | 24/7 |
| Official Project Website | Click Here To Visit Project Website |
The Core Framework
To build a public chain that supports AI models, Cortex 2.0 needs to optimize the AI model inference and the underlying blockchain infrastructure. Not only does it needs to improve the on chain model accuracy and consistency, but it also needs to optimize the existing CTXC chain in terms of consensus and performance.
Cortex AI operator library Further improve the underlying operator library of AI model supported by CTXC, so that CTXC can achieve more AI model inference work. Gradually realize the packaging of transfer transactions, smart contracts, and AI inference, and improve the performance of the CTXC chain through zero knowledge proof technology.
Formal Verification Z3Prover
The instruction execution and calculation results in the smart contract virtual machine belong to the consensus mechanism of blockchain, which requires the instruction operation in the virtual machine to be deterministic and reproducible. CTXC 1.0 regards the AI model inference operation as a basic instruction (INFER| IFNERARRAY) integrated into the virtual machine (CVM), and this leads to two important characteristics that AI inference operation should have on blockchain.
Cortex Determinism and reproducibility. The Cortex Labs team pays sufficient attention to the above features, and proposes or plans to propose a series of interpret-able models or methods to ensure the completeness of the inference operation within the deterministic AI framework. Write and publish an MRT quantitative paper to introduce the necessity, completeness of the deterministic AI framework and related model transformation method.
More AI models on Cortex Extended Operator Set
A series of operator sets and their implementation are defined in the CVM Run time project library. And a strict mathematical description definition is described above, stipulating the logic of the operator to calculate under given input and output a deterministic result. The supported operator set refers to the existing main stream deep learning framework architecture and combines the network structure involved in AI models used most often, including necessary operators.
Cortex In addition to the officially defined operators, CTXC 2.0 will also launch a custom operator function. Users can complete the custom operator according to the protocol and tools provided, and upload the operator to the CTXC operator library for the extension. That extends the user defined range from the model level to the operator level. Operators contributed by the community can effectively create a practicable operator library to meet their needs.
Performance Improvement: three phase of Zero Knowledge Proof
In the blockchain field, performance bottlenecks have always plagued relevant researchers to ensure the decentralization and security of the blockchain system. Up to now, there are many solutions to improve the blockchain performance, such as the consensus protocol replacement, DAG, zkRollup, sharding, and side chains [19].
Due to the limitation of the CAP theorem of distributed systems, scaling up the blockchain will be an option, a trade off between system consistency, availability, and persistence. The CTXC Labs team has conducted a series of indepth research on the capacity expansion issue, hoping to improve the network performance without sacrificing core security assumptions. And they finally selected the zkRollup as the solution.