JSrealB: demo of a text realiser for the web
Natural language generation, a part of artificial intelligence, studies the development of systems that produce text for different applications, for example the textual description of massive datasets or the automation of routine text redaction.
The web is constantly growing and its contents, getting progressively more dynamic, are well-suited to automation by a realiser. However, existing realisers are not designed with the web in mind and their operation requires much knowledge, complicating their use.
We present here a text realiser designed specifically for the web and easy to learn and use. This realiser allows its user to build a variety of French and English expressions and sentences, to add HTML tags to them and to easily integrate them into web pages. This approach was implemented in two phases.
1st phase: JSreal
JSreal was written in 2013 by Nicolas Daoust as part of his master's thesis in computer science at RALI. He showed the benefits of the approach, but the realiser could only generate French sentences. For more information, please refer to the archived page for the project.
2nd phase: JSrealB
JSrealB was developed in 2015 by Paul Molins as part of an internship from INSA Lyon spent at RALI. He came back to the initial idea for the realizer, and generalized it to a bilingual use-case (French and English). He completely overhauled the system's architecture.
For more information:
- GitHub repository
- Documentation for JSrealB users [in French], including examples of syntax representations in French and English, as well as their realisations.
- JSrealB SDK for developers [in French]
- Demo paper presented at the ENLG 2015 conference
- A short paper summarizing the system's features
- Paul Molins' internship report [in French]
- Presentation of this work at RALI-OLST's seminar on 8 July 2015 [in French]
Click on one of the examples above and it will be loaded below.
For more information, contact Guy Lapalme.