On-demand service composition based on natural language requests

The widespread and diversity of Web services in the ubiquitous computing era and the impossibility to predict a priori all possible user demands, generates the need of a system capable of dynamically composing new services, based on unrestricted natural language requests. In order to create such a system two problems need to be solved: a) retrieval of required services and b) service orchestration that fulfills the user request. We solve the first problem by using concepts associated with services and we define a conceptual distance between the user request and a service configuration. Retrieved services are then composed, based on some aspect oriented templates (called Aspects of Assembly) that provide the means of adaptation for a composed service. Our system successfully built new functional configurations of devices inside an intelligent house. One test scenario was to create the service that fulfills the following request: "I want to use my phone to turn off the light, turn on the TV and play some music on HiFi". The implementation is based on the WComp middleware for ubiquitous computing. In contrast to other approaches, our method doesn't find services using individual words, nor is using a controlled subset of natural language, but tries to minimize the distance between the user demand and the potential services that can be orchestrated.