Background

Before understanding deep learning, we need to first define the concepts of artificial intelligence and machine learning and know-how artificial intelligence, machine learning, and deep learning are interrelated to each other. First of all, artificial intelligence (AI) can be briefly defined as a research activity to automate intelligent tasks performed by ordinary people. As such, AI is a comprehensive field that encompasses machine learning and deep learning. Machine learning is a little more advanced technology in AI. Machine learning started from the question, "Is it possible for computers to process more than 'know how to command' to make something work? And can computers learn how to perform a specific task on their own? Instead of the data processing rules created by the programmer himself or herself, can computers see the data and learn these rules automatically?" These questions opened up a new paradigm in the AI industry. In the paradigm of AI, traditional programming, entering rules (program) and data to be processed according to these rules outputs answers. In machine learning, the rules are output by entering the data and the expected answers from this data. Computers can create creative answers by applying this rule to new data. Machine learning systems are not explicitly programmed, but trained. Providing many samples related to the task finds statistical structures in this data and creates rules to automate the task. Then, what is deep learning? Like machine learning, deep learning is a subset of artificial intelligence and is a form developed in artificial neural networks. The artificial neural network utilizes information input/output layers similar to neurons in the human brain and is a technique that automatically performs complex mathematical modeling when data is input in the form of a black box. Deep learning learns data through algorithms using these complex artificial neural networks.