Framework for Linguistic Applications Development(FLAD)

Abstract

Abstract Natural Language Processing (NLP) is a main area emerged with the Arti - cial Intelligence. Text to Speech (TTS), Speech to Text (STT), Optical Character Recognition (OCR) and Language Translation are four main components of Natu- ral Language Processing software applications which are also known as Linguistic components. There are many service providers currently available providing the services of the previously mentioned four linguistic components. Each linguistic component also has di erent vendors providing the service. These services can be online or o ine. For an example, Google services such as Google Cloud Speech API, Google Cloud Vision API are online services. O ine services can be libraries such as Tesseract or stand-alone servers like MaryTTS which should be installed in machines. Therefore, these service providers maintain di erent interfaces to pro- vide their services which makes a burden to the linguistic application developers. Developers who want to use di erent linguistic components in one application will have to face a lot of di culties in many ways such as installing services, con g- uring services, etc. Developers who want to add more advanced functionalities to their applications by combining two or more linguistic components will have to face many major problems in integrating these components in the same application as these components are not in the same platform to be integrated. Considering all these problems we propose a framework which brings all the above mentioned lin- guistic components to a common platform and exposes the services as web services through REST APIs. Further, this framework addresses the issue of combining the services of linguistic components by exposing REST APIs for the complex services which are made up of services of two or more linguistic components. Therefore, developers only have to send requests to the REST APIs and handle the JSON responses sent by the system. Developers can create projects to access the services provided by the framework through the FLAD Console of the system. Developers can select the linguistic components and vendors of the selected linguistic compo- nents they need in their applications. When considering this framework from the technical aspect, expandable nature of this framework is an important achievement. This framework is designed in a such a way that a new linguistic component or a new instance of a linguistic component can be integrated to the framework with less overhead. However, when considering the overall nature of this framework it is clear that this will ll the gap between linguistic applications development and the linguistic application components which addresses the goal of this project that is to reduce the application development overhead through implementing a framework for linguistic application development.