부산대학교 도서관

This paper contains the design and development of an adaptive systems ( AdaptiveSystems Domain-Specific Language - DSL) framework to assist language developers and data scientists in their attempt to apply Artificial Intelligence (AI) algorithms in several application domains. Big-data processing and AI algorithms are at the heart of autonomics research groups among industry and academia. Major advances in the field have traditionally focused on algorithmic research and increasing the performance of the developed algorithms. However, it has been recently recognized by the AI community that the applicability of these algorithms and their consideration in context is of paramount importance for their adoption. Current approaches to address AI in context lie in two areas: adaptive systems research that mainly focuses on implementing adaptivity mechanisms (technical perspective) and AI in context research that focuses on business aspects (business perspective). There is currently no approach that combines all aspects required from business considerations to an appropriate level of abstraction. In this paper, we attempt to address the problem of designing adaptive systems and therefore providing AI in context by utilising DSL technology. We propose a new DSL ( AdaptiveSystems ) and a methodology to apply this to the creation of a DSL for specific application domains such as AdaptiveVLE (Adaptive Virtual Learning Environment) DSL. The language developer will be able to instantiate the AdaptiveSystems DSL to any application domain by using the guidelines in this paper with an integrated path from design to implementation. The domain expert will then be able to use the developed DSL (e.g. AdaptiveVLE DSL) to design and develop their application. Future work will include extension and experimentation of the applicability of this work to more application domains within British Telecom (BT) and other areas such as health care, finance, etc.