DEEP LEARNING: A SURVEY OF RECENT ADVANCES AND APPLICATIONS IN MACHINE LEARNING

Deep learning is one of the most promising paradigms of modern artificial intelligence (AI) and it has a great impact on different computation tasks. The present paper provides a comprehensive and organizationally consistent overview of recent developments in deep learning including algorithmic enhancement as well as pilot roll-outs. The study was conducted via systematic literature review from the leading scientific databases such as Scopus, IEEE Xplore, ACM Digital Library, Web of Science and arXiv using a list of keywords to target the most important core architectures, applications and research frontiers. We synthesized the development of popularities leading deep learning libraries links from well-known models like CNN and RNN to today’s state-of-the-art transformers, graph neural networks and diffusion-based generative models. Selected applications in other areas are reviewed and their performance, limitations, and interdisciplinary impact are addressed. It specifically includes persistent challenges such as data scarcity, explanation, computational burden, ethical challenges and adversarial robustness issues, and points out to key research gaps that still define the field. I