Application of Artificial Intelligence (AI) has revolutionized many sectors like healthcare, agriculture, finance, computer vision, ecommerce, social media, data security, and education. AI plays a vital role in the health sector, like detecting, diagnosing, predicting diseases in advance to reduce the suffering and mortality rate. Besides, application of AI techniques improves hospital management and detection of health insurance fraud. With increased automation in the medical sector, advancements in image acquisition devices and availability of personal wearable devices at affordable cost, voluminous data are being generated. Deep learning techniques can leverage this big data with powerful Graphical Processing Unit (GPU) based systems to analyze and detect hidden patterns in the data and gain insights. Deep learning techniques can learn features from big data sets to get insights that will assist doctors in early diagnosis and treatment. Medical data analysis faces many challenges like limited data availability due to privacy issues, unbalanced data sets for diseases like cancer and rare disease, unavailability of specialists for labeling the data, variation in the experts’ opinion in decision making, variability in genes, environment, and lifestyle of individuals. This chapter discusses the techniques for dealing with the challenges of medical data processing. It also presents the AI techniques for identifying and predicting different types of cancer, diabetes, cardiac, and rare diseases at an early stage using data sets of different formats like clinical data, gene expression data, and medical images. The chapter includes the sections that discuss the methods to deal with unbalanced, small, and high-dimensional data sets, data and label denoising methods, and feature representation learning using neural networks.
