Location: Olsson Hall 104
Add to Calendar 2019-07-17T15:00:00 2019-07-17T15:00:00 America/New_York Master's Thesis Defense for Martin Bolger A Framework for Creating Text Parsing Dialogue Systems Olsson Hall 104

A Framework for Creating Text Parsing Dialogue Systems

Educational dialogue-based simulations have been used in variety of fields.  Some applications include: teaching medical students effective communication methods for clinical exams [1], exposing military personnel to local culture before deployment [2], helping students practice a foreign language [3], and teaching negotiation strategies [4]. Many of these simulations use a multiple-choice dialogue system to simulate conversation. User input can easily be scored based on the educational goal of the simulation because each multiple-choice option has an associated score. Despite its advantages, multiple-choice training is a poor proxy for the complexity of real-word interactions. Free-input training would be more appropriate, but free-input dialogue systems are difficult to implement and do not offer a clear input scoring mechanism. Many past dialogue-based simulations with a free-input dialogue system have restricted input to specific phrases that users are taught during a training phase [5][6]. Other dialogue-based simulations have used a non-restrictive free-input dialogue system but lack an educational mechanism because they do not perform detailed classification of specific text features in the input to provide a score [7].

The goal of this research is to develop a framework for creating a free-input dialogue system that is able to perform detailed classification of input to provide educational feedback.  Components of the developed framework include a data generation scheme for creating potential input, methods for creating crowdsourcing prompts for collecting data, and a classification pipeline that finds machine learning models and text representations to create classification models for scoring user input. The framework was implemented in an existing dialogue-based simulation to replace its multiple-choice dialogue system. Modeling results show that this approach offers a viable method for building a free-input dialogue system for an educational simulation.

 

[1] Kron FW, Fetters MD, Scerbo MW, White CB, Lypson ML, Padilla MA, et al. "Using a computer simulation for teaching communication skills: A blinded multisite mixed methods randomized controlled trial." Patient Educ Couns (2016)
[2] Zielke M. A., Evans M. J., Dufour F., Christopher T. V., Donahue J. K., Johnson P., Jennings E. B., Friedman B. S., Ounekeo P. L., Flores R., “Serious Games for Immersive Cultural Training: Creating a Living World.” IEEE Computer Graphics and Applications, vol. 29, no. 2, pp. 49-60, Mar./Apr. (2009)
[3] Bernstein, Jared & Najmi, Amir & Ehsani, Farzad. (1999). Subarashii: Encounters in Japanese Spoken Language Education. CALICO Journal. 16.
[4] J. Kim, R.W. Hill, P. Durlach, et al., BiLAT: A game-based environment for practicing negotiation in a cultural context. International Journal of Artificial Intelligence in Education
[5] P. Wik and A. Hjalmarsson, “Embodied conversational agents in computer assisted language learning,” Speech Commun., vol. 51, pp. 1024–1037.
[6] Johnson,  W.L.,  Marsella,  S.,  Mote,  N.,  Viljhalmsson,  H,  Narayanan,  S., Choi,  S. Tactical Language Training System:  Supporting the rapid acquisition  of  foreign  language  and  cultural  skills. InSTIL/ICALL  Symposium, Venice, Italy. (2004).
[7] Mateas, Michael and Andrew Stern. “Structuring Content in the Façade Interactive Drama Architecture.” AIIDE (2005).

 

Committe Members:
Dr. Donald Brown
Dr. Laura Barnes
Dr. Stephanie Guerlain