Proof of training (POT)

Our primary objective is to establish a robust consensus protocol called proof of training (PoT) that lays the foundation for harnessing the power of crypto mining for distributed artificial intelligence (AI) training. The development of this protocol is crucial for enabling the efficient and secure utilization of computational resources across a decentralized network, with the ultimate goal of advancing AI model training. In this section, we will concentrate on the functions and utilities of the protocol, abstracting away from specific network designs.

Conceptual figure depicting on-chain / off-chain components in the Decentralized training network (DTN), which consists of two major components: an on-chain component SC , resident on a mainstream blockchain, and an off-chain component exec that executes on a DTN. The DTN serves as a bridge between the two components as well as connecting the system-level contract with off-chain resources such as service providers, validators, decentralized storage, etc.