Semantic Interoperabiliity for SageTea

Renseignements sur le financement
Natural Sciences and Engineering Research Council of Canada
  • Type de subvention: Programme de subventions d'engagement partenariat
  • Année: 2014/15
  • Financement total: $25,000
Mots clés
Chercheur(e) principal(e)
Collaborateur(s)

Aucun chercheur n’a été trouvé.

Partenaires

Sommaire du projet

SageTea® (State-Activity-Group-Element-Transition-Event-Action) is a new technology to design software which is based on the conversion of natural language (i.e. text) into software automatically. At present, SageTea uses a simple approach for language processing It can read sentences, paragraphs and can determine the topic of each sentence. In its current state, it only has a limited ability to handle complex syntax, such as conditional statements. SageTea provides an API to Smalltalk for direct programming of conditional statements. In order to be able to develop fully functional software applications from user textual inputs, the input mechanism should be user friendly, and the translation should be automated using proper syntax, grammar, and, in particular, semantics. During this research project we will investigate the different problems existing for designing and developing such framework. In order to improve the software generation process, we need to tackle important issues, in particular the semantics of the natural language and the identification of mappings between the two language grammars. In this research, we are going to investigate the basic architecture about the translation process. In particular, we will focus on the semantic extraction of the input language as well as the basic solvable mapping rules between two languages (e.g. English vs. SmallTalk). This research collaboration will focus on advanced techniques to improve the semantics understanding of natural text and its translation into software code. It will also focus on the interaction with adnvanced input devices in order to improve the software development process.