GSI BOT

The project

This project aims to integrate the advantages of intelligent agent systems, information retrieval and Natural Language Processing in a personal assistance system.

Thanks to this integration, we have been able to provide the personal assistant the ability to solve questions and support the learning process in an e-learning platform. This opens the opportunity for the student to be able to find help in any moment, not having to resort to the teacher. At the same time, thanks to the agents system the learning process is checked and supported to be the correct one. Furthermore, for solving the student questions, our personal assistant has been given the capability to access external sources of data using information retrieval techniques, so he can also improve his existing knowledge base.

During the development, we dealt with one of the main limitations that the natural language processors nowadays have: as there is a wide range of applications for them, they usually are unable to execute orders from the user that require using external modules. We overcome this limitation by implementing over an event-based network several execution modules, as the agent system that is able to support the learning process, or the information retrieval module.

Demo

You can try a live demo from here.

Download

You can download the code from here.

About Us

This project has been developed as the master thesis of Javier Herrera under the tutelage of Miguel Coronado and the supervision of Carlos A. Iglesias at GSI UPM.