3,932 reads

AI on Blockchain — What’s the catch?

by Oscar WOctober 13th, 2018

Too Long; Didn't Read

To date, Artificial Intelligence has become ubiquitous and our reliance on Artificial Intelligence increases day by day. For examples, Amazon’s recommendation system that offers users personalized recommendations based on previous purchases or activity; Uber’s AI model to determine arrival times, pick-up locations, and fare price; and PayPal that uses machine learning algorithms to detect and combat fraud.

Companies Mentioned

Coin Mentioned

featured image - AI on Blockchain — What’s the catch?

What Is Cortex solving?

AI Monopoly

source: https://www.redpixie.com/blog/examples-of-machine-learning

Having a few companies to dominate the AI industry may not be wise. AI models trained by the big corporations often put their own interests before costumers’ privacy or needs. Blockchain technology can change the status quo by decentralizing and open-sourcing AI algorithms to ensure that people could choose from multiple AI providers and are free from the danger of big corporations.

The Cortex blockchain can democratize AI by creating a tokenomic model that incentivizes AI developers to share their models with the world: every time a model is called, the owner of that model will get paid in CTXC. This will encourage more developers to post better AI models on-chain. With more and better models being submitted on the chain, Smart contract and DApp developers will have access to more functional AI tool to include in their development. This creates a new and healthy eco-cycle for the AI industry.

Limited Smart Contract Capability

Smart Contract is one of the best inventions of the century. The idea is that contracts can be translated into code, which is then distributed to multiple nodes for verification in a blockchain. Yet, the current smart contracts are not actually “smart” but more like digital contracts. Without enhancing the capability of the smart contract it is hard to implement real-world decentralized applications on the blockchain. A real smart contract should have AI functionalities.

With the current smart contact technology, developers could only create applications like CryptoKitties that offers users static cat pictures with no interaction or Fomo3D that uses Mathematical calculations to provide a scam-like experience. With Cortex’s AI smart contract, developers could create more complex applications such as a KYC application that uses a face recognition model to provide a more secure process and prevent fraudulent actions.

Source: https://slideplayer.com/slide/14197503/

Unable to Make AI Inference On-Chain

At present, smart contracts and virtual machines are executed on CPUs; however, due to the complexity of AI algorithms, there is no public blockchain that has enough computational power to effectively execute AI algorithms. Current solutions limit the company to execute AI off-chain or on the side-chain then bring the results back on the mainchain which still pose black-box issue. The Cortex public blockchain is built from the ground up, using GPU/FPGA on full nodes to support AI inference. The Cortex Virtual Machine will use GPU/FPGA to execute the AI smart contract which also creates a transparent audit trail to improve the trustworthiness of the result.

How does Cortex do that?

AI-powered Decentralized Applications foresee a trillion-dollar market, and Cortex will be the first infrastructural blockchain to support this ecosystem.

AI Smart Contract and DApp

Allow smart contracts to include AI functions, creating truly intelligent smart contracts and enhanced DApps with more diverse use cases. AI DApps will be executed on each full node using GPU/FPGA and reach consensus on-chain. AI DApps will have enhanced capabilities, such as more vivid NPCs in decentralized gaming, AI-governed stable coins, AI model competitions, Autonomous Driving, Cryptocurrency Lending, Ads platform, and Recommendation System.

AI Inference

With Cortex’s unique approach to bring AI inference directly on-chain, the inference result no longer comes from a third-party Oracle, eliminating the trust problem. Cortex provides “Synapse”, an integer inference engine to do deterministic AI inference with Cuckoo Cycle consensus mechanism, one of the most promising ASIC-resistant frameworks due to the memory-intensive algorithm, yet with instant verification. Cortex also proposes MRT (Material Representation Tool), an open and flexible standard that enables deep learning frameworks and tools to be interoperable. MRT guarantees AI models deployed on Cortex can be executed with the same inference results.

AI Model Platform

An AI model platform on blockchain where AI model developers and providers can store their most advanced AI models on-chain and be accessed by smart contract developers who pay them in CTXC. The platform can boost applications like AI DApps, AI competitions, model zoos, and data stores that help the open-sourcing of AI.

Cortex Virtual Machine

The CVM is fully compatible with EVM but the entire infrastructure runs on GPU. Ethereum developers can migrate their Ethereum smart contracts to the Cortex blockchain and add AI functionalities. Developers can use Solidity to write AI smart contract with infer instruction sets that are intuitive and easy to implement.

About Cortex

Cortex is the first ever blockchain technology that allows the execution of AI algorithms on the blockchain. Cortex provides an AI platform and stack for developers to upload their models on the blockchain that can be integrated with the smart contract. TestNet for mining and AI smart contract is out, Click here to learn more!