A Survey on Automated Machine Learning: Problems, Methods and Frameworks

Automated Machine Learning (AutoML) is a research field that automates machine learning processes and optimizes their costs. As machine learning begins to be widely used, many users in industry and academia are paying attention to AutoML. However, to satisfy the different demands in their various fields, the AutoML systems have been defined and studied differently according to several requirements. This paper classifies various characteristics of the AutoML systems into Hyperparameter Optimization (HPO), Combined Algorithm Search and Hyperparameter Optimization (CASH), and Machine Learning Pipeline Creation (MLPC) problems, and formulates their problem statements in a similar fashion. Moreover, we review the methods and frameworks widely used in the AutoML field in order to improve the practical understanding of the future direction of AutoML.