Abstract:With the popularity of artificial intelligence (AI) and its integration into every field, it has yielded positive results that have increased productivity and aided us in solving complex problems. Deep learning (DL) is a subset of AI developed to mimic the human brain. It is a way for a computer to perform actions that come naturally to humans. It is a tool extensively used to organize unsupervised or unlabeled data and find patterns in them. DL has had a major impact on the field of medical science owing to its applications in medical drug discovery, medical imaging, genome synthesis, disease diagnosis, and much more. The growth of DL in this field has increased significantly owing to the processing and type of data used in the models. Focusing on the type of data, preexisting or curated, can drastically affect the success rate of a DL model. DL is used to diagnose diseases such as skin lesions, brain disease, and chronic disease. To understand the application of DL in disease diagnosis and how it has become one of the most efficient methods for disease diagnosis, this chapter will examine the different architectures of DL, such as convolutional neural networks, feed forward networks, and recurrent neural networks, and how their diagnosing methods compare with traditional disease diagnosis methods. In addition, it will explore advantages and disadvantages of DL compared with traditional methods by comparing various efficiency parameters such as accuracy, specificity, sensitivity, area under the curve, and receiver operating characteristics curve, showing how DL is more effective and produces better results. However, DL faces numerous obstacles. If these are overcome, it will lead to even better results and complete autonomous disease diagnosis.

Deep learning applications for disease diagnosis