Abstract
Researchers across different disciplines develop models to solve various types of questions. Consequently, many of the developed models tend to be implemented using different development tools and adhere to different requirements. Thus, making it extremely difficult to reuse these models. Therefore, the Connected Intelligence Engine (CIE) aims to enable model reusability by providing standards and infrastructure that automate the process of models' deployment and execution. In CIE, we addressed the reusability and accessibility of such models. We adopted lightweight virtual-machine (Docker) to automate the deployment, execution, and access in the simplest form where it would require the least amount of expertise and knowledge.
Additionally, researchers and developers face significant challenges in the ability to integrate and connect models. These challenges becomes more complicated with complex scientific models. Such models are developed to solve various types of complex questions e.g., a traffic model that calculates the time it will take to travel between two geo-locations. Different individuals work with varying types of development tools based on their experience and the needs of the problem. Therefore, many of the developed models tend to have different execution requirements. Furthermore, it will require individuals to have a background in a variety of tools that the models are commonly developed in.