Features of Hybrid Machine Learning in Modern Computational Linguistics: Navigating Systemic Risks, Turbulence and Uncertainty Through the Lens of SHIVA and WARMI Paradigms

Abstract

The article explores the features and specifics of hybrid machine learning in modern computational linguistics in conditions of dynamic uncertainty, turbulence and systemic risks. The main directions of hybridization of neural, symbolic, statistical and linguistic methods are considered, as well as the possibilities of integrating large language models, knowledge graphs and adaptive learning mechanisms. The prospects of hybrid architectures for increasing the accuracy, stability, adaptability and comprehensibility of intelligent linguistic systems in conditions of SHIVA and WARMI paradigms of the external environment are substantiated.