May 29 2008
3:30pm - 5:00pm,
CERAS, Room 513
Professor, Department of Psychology,
University of Colorado at Boulder
PhD. 1960 - Harvard University
Author of award-winning book:
"The Trouble with Computers: Usefulness, Usability, and Productivity"
Professor Landauer teaches cognitive psychology at the CU boulder campus. He came to the University in 1994. Prior to CU he worked at Bell Labs and Bellcore for 26 years and served as director of cognitive research at the New Jersey Laboratory for 10 years. Currently Mr. Landauer is part of the University of Colorado's LSA project that researches latent semantic analysis.
For more information on Tom Landauer, please visit his Home Page
The last decade has seen the emergence of ways to build mathematical models that closely mimic human understanding of the meaning of text. The opportunities this opens for devising new ways for educators to help students learn are wide and deep. Current examples include automatic essay and summary evaluation that is as reliable as humans but returns scores and pedagogical feedback in seconds, replacing multiple-choice exercises with open-ended answers, and computer-based tutors that interact in natural language. The presentation will begin with a brief and high-level introduction to the computational language modeling, then describe and demonstrate some applications currently in evaluated use as well as some still in the R&D pipeline. The latter include reading and summarizing with automated feedback in Spanish—or with reading and writing summaries in any other combination of the two, “Open-Cloze” items where students put in their own words, and automatically monitored, mentored and enriched collaborative teaching and learning environments. I am hoping to get frank reactions and advice on these tools and ideas.