3.1 Design concept

In order to support the many complex applications of AI, it is necessary to break through the limitations of program size and programming language of virtual machines on the blockchain. However, for the AI industry application platform, besides considering the importance of data value, it also needs to include developers of AI industry applications, platform resource providers with the resources to run AI industry applications (with a large number of video card servers), and consumers. Ultimately, someone has to use these resources, and it is not enough to just define data objects on the blockchain. ATMT is to split the AI industry application development process and use it into different roles, publish them on the blockchain, provide information release channels, and promote the interoperable flow of resources to form an AI collaborative development application platform.

1. ATMT is the use of external virtual machines, integrated open source Docker containers to solve the limitations of the programming language. the solution of ATMT integrated Docker containers can be gradually improved with the progress of the open source project Docker. Relying on the agreement and release of a unified standard DockerIMG image file enables ATMT's application runtime environment to remain uniform and independently upgradeable. The ability to integrate multiple programming languages in the DockerIMG image file, including complex AI industry application environments such as tenserflow, makes it possible to trigger running AI industry applications on ATMT nodes.

2. ATMT stores the application executable data and resource data outside the blockchain, and stores the descriptive information (or tags) of the application and resource data on the blockchain. It allows larger data size and more complex applications.

3. ATMT provides a blockchain interface dedicated to application data resources deployment, extraction and transaction record query on the blockchain, defines rules for asymmetric encryption and decryption to verify user identity, and enables application providers to easily complete the interface between services and the blockchain without understanding the underlying blockchain technology.

ATMT has agreed on four roles: data provider, application provider, running platform resource provider, and resource consumer. The first three roles all have different types of resources; resource consumers are the people who use these resources. Consumers can be ordinary users or companies that are developing AI industry applications (which require large amounts of data for machine learning).