Neural Fairness Consensus Protocol

Nowadays, the concept of Immutable Decentralized/Centralized Distributed Ledger Technology, otherwise called Blockchain, is rising in many fields of the society. It is easy to hear about it in fields as energy, automotive, economy, voting, law, insurance, cloud computing, supply chain management, etc. However, many problems are present. In particular, from a technological point of view, the problem of scalability-intended as the capability to maintain the network, composed by a very huge number of nodes, up, with a reasonable operational velocity- should be solved in order to be really suitable in production; as well as the question of maintaining a high level of security in terms of data exchanged and identity of members, is an important issue, and has taken central stage thanks to legislation like the European Union's General Data Protection Regulation (GDPR). Finally, it is not always guaranteed that the system is truly decentralized. This is known as the Blockchain Trilemma. It is a fundamental problem that must be addressed before a global adoption of blockchain ecosystems. To face such a problem, a remarkable number of solutions have been proposed in recent literature. In particular, several types of consensus' protocols have been conceived, in order to achieve scalability and velocity. In particular the Proof of Stake consensus approach was the first to show that a balanced combination of all these three properties may be possible. Within this context, the main goal of the thesis' job is to design a Neural Fairness Consensus Protocol (NFCP) to validate transactions while, at the same time, preserving the privacy of the identity members, as well as the data exchanged into the network and assuring a scalability for the entire system. The NFCP is a blockchain-based distributed ledger secured using neural networks and machine learning algorithms, enabling a permission-less participation in the process of transition validation while concurrently providing strong assurance about the correct functioning of the entire network. The protocol is built to guarantee the privacy preserving using cryptography and federated learning techniques in each phase of the transaction and data communication between peers.