Improving Modeling for Autonomous Machine Learning with a Task Ontology

Abstract

Computing, big data, and machine learning algorithms are the building blocks of artificial intelligence, which has recently attracted a great deal of academic interest due to its potential to identify, learn, infer, and act upon external information across many domains. Nearly every sector is making use of AI technology right now, and a large number of machine learning specialists are trying to standardize and integrate different machine learning tools so that even those without a technical background may utilize them effectively. Ontology building for standardizing machine learning concepts and autonomous machine learning are additional areas of interest for the researchers. As part of this study, we outline a problem-solving process and categorize common processes in autonomous machine learning problem-solving as tasks. Our proposed approach to modeling autonomous machine learning makes use of a workflow-like procedure for executing machine learning tasks. The suggested machine learning model is based on task ontologies and proposes a strategy for grouping UML actions based on tasks. Furthermore, it will autonomously construct and augment machine learning models using transformation rules grounded on shared components and structures, including elemental linkages and processes.