Comparison of the efficacy of particle swarm optimization and stochastic gradient descent algorithms on multi-layer perceptron model to estimate longitudinal dispersion coefficients in natural streams

Abstract

Union’s General Data Protection Regulation (GDPR). The analysis focuses on the applicability of the ‘data protection by design’ principle during the development of such systems. Because blockchain-based networks are built on distributed data processing operations, therefore data controlling or processing of participating nodes should comply some abstract data protection patterns predetermined and collectively built-in during the system’s development phase. On the other hand, the imprint of AI’s automated data processing could be also observed and tracked back in the blockchain due to its historically retroactive nature. In the end, the study presents the human mind and its ‘uploading’ with conscious and unconscious contents as an analogy to blockchain-based AI systems. My goal is to highlight that the synergy of blockchain and machine learning-based AI can be hypothetically suitable to develop robust yet transparent automated decision-making systems. The compliance of these distributed AI systems with data protection law’s principles is a key issue regarding the high risks posed by them to data subjects rights and freedoms.