automatica
M. Hadjiski. Мutual Penetration and Enrichment as a Bilateral Accelerating Factor for the Development of Machine Learning and Automation

Key Words: Artificial intelligence; control theory; deep learning; industrial automation; machine learning; reinforcement learning.

Abstract. The study shows that the existence of fundamental similarities between control theory and machine learning is a real basis for productive interpenetration and enrichment with new concepts, methods and tools, which are mutually beneficial for overcoming a number of serious modern challenges such as the control of complex and autonomous systems, cybersecurity, intelligent robotics, bioautomatics. The continuous development of the control theory as a result of its own progress and under the influence of the ideas of artificial intelligence will not allow the transformation of automation into a routine engineering discipline. In turn, the systems based on artificial intelligence and machine learning are enriched with well-developed methods and procedures from the control theory in order to improve and create new algorithms and to ensure a higher speed, robust stability and optimality of the learning process. Industrial automation systems will absorb the innovative results of the interpenetrating development of artificial intelligence and control theory improving both the quality and scope, safety and security of operations.