NSF Grant Supports Innovative Computer Language Modeling Research

Dr. Nigel G. Ward

Dr. Nigel G. Ward

The National Science Foundation awarded The University of Texas at El Paso College of Engineering nearly $450,000 over three years for a computer modeling project that will explore human dialog and interpersonal communication dynamics.

The project entitled, “RI:Small Time-Based Language Modeling” will be led by Dr. Nigel G. Ward, Associate Professor of Computer Science; Dr. David G. Novick, Professor of Computer Science; and Dr. Olac Fuentes, Associate Professor of Computer Science. Graduate and undergraduate students will gain hands-on research experiences by completing data collection, programming, and analysis for the project.

Based on past context, speech recognizers can predict what words are likely to appear next. The quality of these predictions is critical for achieving high recognition rates. The predictions made by today's language models use the information in the previous words. This innovative research will examine how to use new sources of information, such as the local speaking rate, local pitch range, and other indications of the speaker's cognitive state, to make the predictions.

Dr. Ward notes that the likely project impacts will extend beyond engineering applications, “A model of the cognitive and communicative processes underlying dialog may lead to better methods for training people to be more effective speakers and better listeners, with the potential to improve effectiveness and harmony in personal life, in organizations, and beyond.”