Current applications of machine learning in dentistry

• Main Authors: Ghazali, Ahmad Badruddin, Reduwan, Nor Hidayah, Ibrahim, Roliana
• Other Authors: Yusoff, Yusliza
• Format: Book Chapter
• Language: English
• Published: UTM Press 2022
• Subjects: R Medicine (General); RK Dentistry; RK318 Oral and Dental Medicine. Pathology. Diseases-Therapeutics-General Works; RK529 Oral Surgery-General Works; T Technology (General)
• Online Access: http://irep.iium.edu.my/99803/5/99803_Current%20applications%20of%20machine%20learning%20in%20dentistry.pdf; http://irep.iium.edu.my/99803/
• Institution: Universiti Islam Antarabangsa Malaysia

Artificial intelligence (AI) is the general description given to computer systems that can perform tasks and mimic the requirement of human intelligence input (Pesapane et al., 2018). Machine learning (ML), a subset of AI was described as an algorithm with the ability to "learn" by identifying patterns in a large dataset (Rowe, 2019). ML programs can improve from experience automatically, unlike traditional computer programming, where every step of the program requires a written code (Mayo & Leung, 2018). The process is similar to a human expert that can learn by repeated training (Hung et al., 2019). The quality of the output depends on the quality of data used to train and validate the algorithm (Rowe, 2019). Additionally, deep learning (DL), which is a subset of ML, was inspired by the structure and function of the human brain called artificial neural network (ANN). ANN contains multiple layers of the network that receives the output of the previous layer, computing a task and sending it to another layer, and the structure is able to teach itself by reviewing a large amount of data (Mayo & Leung, 2018). Convolutional neural network (CNN) is commonly applied in computer vision research. The difference between ANN and CNN is that in CNN, only the last layer is fully connected, but in ANN, each neuron is connected with the other (Kumar, 2017). Most mathematical models were developed to find the relationship between input data and output data. However, a complex real-world phenomenon cannot be described easily from a closed-form input-output relationship. Thus, ML is an automated process to build a computational model of these complex relationships (Bastanlar & Özuysal, 2014). This chapter is organized by firstly presenting potential use of ML in dentistry in Section 1.2. Section 1.3 describes the methodology for ML research while Section 1.4 explains the applications of ML in dentistry. Section 1.5 discusses the available ML products and studies in the field of dentistry. Section 1.6 of this chapter provides the limitation and ethical consideration of ML research in dentistry, and finally Section 1.7 concludes the chapter.