EXPLAINABLE ARTIFICIAL INELLIGENCE AS A SERVICE: METODOLOGY OF ASSESSMENT AND QUALITY MODELS

The subject of the research is Artificial Intelligence as a Service (AIaaS) and AIaaS quality models. The goal is to define quality characteristics and develop a methodology for evaluating explainable AIaaS (XAIaaS). AIaaS solutions offered by cloud providers are analysed and a classification of AI tools supported by the services is provided. The AIaaS quality assessment methodology is proposed, which is based on the principles of case-oriented assessment and the quality model of artificial intelligence systems. It combines quality models of AI tools and cloud services, and metric evaluation of relevant characteristics using weighted convolutions. Examples of requirements profiling and quality characteristics for two XAIaaS are discussed. Keywords: artificial intelligence, АІ as a Service, cloud systems, explainable AI, requirements profiling.