To Örebro University

Practical applications of artificial intelligence frequently benefit from the strengths of multiple individual AI approaches. However, these approaches use different representations for data and models and thus are often difficult to combine. To address this gap we created the AI Domain Definition Language (AIDDL) language and framework. The language allows to express AI models, data, and problems, as well as intermediate representations tailored to specific applications. The framework, on the other hand, allows us to define translations between models and data and offers a variety of solver for AI problems. As a result, the AIDDL framework allows to build integrated AI systems tailored to complex problems and composed of well understood and studied AI algorithms.