Deep Learning: Concepts, Types, Applications, and Implementation

Theory of Approximation and Applications

مقاله 11، دوره 16، شماره 2، دی 2022

نوع مقاله: Applied-Research Articles

نویسندگان

Fereshteh Aghabeigi^{* 1}؛ Sara Nazari²؛ Nafiseh Osati Iraqi²

¹Department of Computer Engineering and Information Technology- Arak Branch- Islamic Azad University- Arak- Iran.

²Department of Computer Engineering and Information Technology- Arak Branch- Islamic Azad University- Arak-Iran.

چکیده

Today, deep learning has attracted attention in various scientific and non-scientific fields. Deep learning is a branch of machine learning that simulates the human brain for various applications like recognizing voice, face, handwriting, identifying kinship, image processing, and etc. In deep learning, a set of representation algorithms is used to model high-level abstract concepts through learning at different levels and layers. Deep learning has become popular due to its capabilities like automatic feature extraction, high extendibility, and wide application in different fields. In this paper, it is tried to describe different deep learning models and architectures, how they are trained, and the required hardware and software structures.

کلیدواژه‌ها

Machine learning؛ Deep learning؛ Neural networks؛ Network training