Novel machine and deep learning and training techniques for AI-based healthcare systems

Artificial Intelligence (AI) based techniques including machine learning and deep learning are gaining attention for the applications in the smart healthcare systems. The use of machine learning-based techniques is classified into Supervised Learning, Unsupervised Learning and Reinforcement Learning, which can be trained using different types of data containing labels, or without labels. Further, the use of dynamic environments can train the reinforcement learning-based models to work in decision making similar to humans. Furthermore, the deep learning based techniques including the Convolutional Neural Networks (CNNs) are utilised for the applications of image processing, which can be helpful in detection and diagnosis of the diseases. The AI-based learning techniques have been applied in various healthcare applications, including machine learning-based techniques in health monitoring, decision support systems, chronic illness disease treatment and respite care, whereas deep learning-based techniques can be used for the medical imaging, electronic health records, genomics, and smartphone-based applications. The AI-based techniques can enhance the healthcare systems, especially in the drug discovery, image analysis, disease prediction, medical research and smart health monitoring. The integration of AI-based learning techniques results in enhanced healthcare systems and services, however, there are still some challenges and limitations to the usage of these techniques. The training and implementation of these techniques require a large amount of data for training, and use a substantial amount of resources for the data pre-processing and analysis.