The Advancement of Information Retrieval Systems: Spoken Dialog

by Davina V
This article starts starts off with the interest of designing an automatic system to retrieve information available in backend database. Along with the paper’s relevant definitions and the way the paper is organized.  Then it goes into the latest state-of-the-art development platforms and acceleration strategies (and considering that it is currently 2012 it still is) in three different fields: Commercial, Academic, and Research. Or more specifically they discus about common features like the commercial platform has Speechdraw, OpenVXML studio, and Vocalocity App center. The rest of the article discusses about the authors’ proposal in this paper. Concluding that “Thanks to the accelerations, the design time the evaluators spent on performing the proposed tasks was reduced by more than 56%, the number of keystrokes by 84%, and the keystroke errors by 55%. As regards the subjective evaluation all the accelerations were marked over 8.0 and the whole platform was rated with an average score of 8.0. According to this evaluation, the most appreciated accelerations were the automatic generation of action proposals for each dialog and the ease in designing the state flow model” (D’Haro, Córdoba, Montero, Ferreiros & Pardo, 2012), which sounds pretty good.

As a round it all up I chose this article for my last one, because I wanted to do one final blog on the point of the class database. It just helps me that this same article applies some of the topics that we have gone over in class.

Since this is most likely our last time in this class, I thought that I enjoyed the time together, problems and all.

D’Haro, L. F., Córdoba, R., Montero, J. M., Ferreiros, J., & Pardo, J. M. (2012). Application of backend database contents and structure to the design of spoken dialog services. Expert Systems with Applications39(5), 5665-5680. doi: 10.1016/j.eswa.2011.11.104