Productive blockchain architecture based on parallel machine learning

Abstract

Through several effective research and applications, machine learning has proven to have limitless potential. When delving into robust machine learning usage, two essential research questions are how to guarantee that no one tampers with a system's search results and how to prevent other users within the same network environment from easily obtaining our personal information from apps or systems. Regarding privacy and security, this circumstance is comparable to another existing systems of information. An alternative strategy for addressing these two issues is presented by the rise of blockchain technology. It is for this reason that several recent studies have attempted to include machine learning methods or blockchain-based technologies into platforms for machine learning. To demonstrate what the potential of combining blockchain technology with machine learning is demonstrated in this study, where we present a parallel framework for utilizing a metaheuristic algorithm to determine appropriate deep learning hyperparameters in a blockchain environment. The problem is also considered in the suggested framework. reducing communication costs by restricting the quantity of data exchanges between blockchain and miners.