Some Super-approximation Rates of ReLU Neural Networks for Korobov Functions

Introduction In recent years, Deep Neural Networks (DNNs) have emerged as a crucial technology in the fields of machine learning and artificial intelligence and achieved tremendous success. The remarkable success of DNNs has inspired numerous practical applications in computer science and engineering, including image recognition, natural language processing, autonomous driving, scientific computing, etc. Towards an explainable framework of machine learning, understanding the approximation power of DNNs is a fundamental question and remains an active research area.