Deep learning distributed computing method and system based on block chain

Abstract

The invention relates to the technical field of block chains, and discloses a deep learning distributed computing method based on a block chain, which comprises the following steps: dividing a deep learning computing task into a plurality of sub-computing tasks, and storing the plurality of sub-computing tasks into different block chain nodes; sorting the block chain nodes by using a sorting algorithm based on ring signatures; performing consensus of the deep learning sub-computing tasks by utilizing a consensus algorithm based on a block chain node sorting result, and storing a sub-task computing result reaching the consensus into a storage node; performing fault-tolerant verification on the sub-task calculation result stored in the storage node by utilizing a fault-tolerant verificationalgorithm based on a block chain data tampering mechanism; and calculating the hash value of the sub-task calculation result stored in the storage node, and taking the hash value as the calculation result of the deep learning task. The invention further provides a deep learning distributed computing system based on the block chain. According to the invention, deep learning calculation based on theblock chain is realized.

Description

technical field [0001] The present invention relates to the technical field of block chains, in particular to a block chain-based deep learning distributed computing method and system. Background technique [0002] With the development of Internet technology, the amount of data existing in the Internet is also increasing rapidly, and the industry has more and more demands for massive computing and deep learning task processing. The traditional Internet industry mainly uses a centralized system to process a large amount of data, but this system has a complex structure, poor scalability, low fault tolerance, and is not easy to maintain, so it is difficult to meet the needs of Internet technology. Using a distributed system to process deep learning tasks has become Hot topics of current research. [0003] The price of building a distributed cluster is very expensive, and there must be no malicious nodes in the cluster, and the nodes must be completely trusted, so the nodes can...

AI Technical Summary

Problems solved by technology

[0006] By improving the public chain consensus algorithm of the existing blockchain technology, the problems of insufficient scalability and difficulty in resisting adaptive attacks in the traditional alliance chain consensus algorithm are solved, and the improved public chain consensus algorithm is used to establish a distributed cluster , so as to use the established distributed cluster to realize the distributed computing of deep learning tasks

Images