⚙️Artificial intelligence(AI) computation network based on the Utility blockchain

The operation model of the Utility network al- lows the emergence of AI computation service providers. AI computation service providers build their own AI computation centers by leasing IDC server rooms, and through the Ute blockchain can fairly authenticate the type of AI training and infer- ence chips, their computational size, whether they are online, and other information on the chain, thus confirming their total computational power on the Utility network.

However, AI computation AI computational services. What a computation order buyer needs for AI model training and deployment is AI com- putation service with comparable service quality.

The online AI computation in the chain is only one of the production factors of AI computation service. The computational service provider also needs to purchase CPU servers, hard disk, rent net- work bandwidth, maintain stable power supply of AI servers, and other production factors to provide AI computational service.

How to measure and guarantee the quality of service of AI computational services? An efficient solution to this problem will determine whether the Utility blockchain is productive and whether Utility coins can form a closed loop of circula- tion. We have designed an under-chain AI com- putational network solution to efficiently connect AI computation centers around the computational service provider to achieve efficient sharing and circulation of data, computation, algorithms, mod- els, services and other elements. And with the AI computational service quality scoring mechanism based on Utility blockchain, the chain can trace and publicize the transactions between buyers and sellers of computation services. The combination of on-chain and off-chain provides efficient produc- tivity in the AI era. The combination of on-chain and off-chain can maximize the incentive for AI computational service providers to join the Utility computing network, reducing the cost of using com- putation, promote the unified settlement of hetero- geneous computation in different places, make the production, distribution, circulation and consump- tion of AI computation smoother, and improve the efficiency of market operation. The goal is to build a worldwide AI computational infrastructure with easy access to computing resources, efficient and unified scheduling of tasks, convenient and inex- pensive use by users, and a sustainable operation model and mechanism.

The platform layer of Utility AI computa- tional network is designed with four central sub- platforms based on Utility blockchain, namely, the AI computational network task unified scheduling and management platform, the AI computational network data unified storage and management plat- form, the one-stop platform for debugging and training deployment of AI computing centers, and the AI computing network data market.