Envisaging an Intelligent Blockchain Network by Intelligence Sharing

Blockchain Technology is gaining popularity throughout various industry verticals due to its data decentralization and tamper-evident nature. Machine Learning (ML) is all about embedding a learning capability to computing machines so that the machine can learn based on historical data in a way how human beings learn things. An important part of ML is the process of learning which needs humongous processing capability and hence it is time-consuming. Significant benefits have been predicted from the integration of these two technologies. Making a complete blockchain network intelligent in a simple and efficient way is a major challenge. In this work, a Multi Layer Perceptron (MLP) model is implanted in every node of the blockchain network. An efficient technique is proposed to make an intelligent blockchain network in minimum possible time and using minimum processing power. During the network formation, every node of the network has knowledge of the model architecture. At some point in time, the model of the randomly selected node gets trained. After completion of the training of that node, the intelligence is replicated to the entire network.