Phonetic searching applied to on-line distance learning modules

In recent efforts, tools have been developed for efficient and rapid production of high-quality, computer-based education materials. Two main properties are to be optimized: ease of production by the preparer of the material, and ease of use by the e-learner. One particular implementation of this strategy is the Infusion production tool. The e-learner is presented with material via download, streaming, or CD which can be accessed in a variety of ways. Often the student would like to search this material for topics over single lectures, entire courses, sets of multiple courses, multiple instructors, or a combination thereof. Since the goal of Infusion is to relieve the instructor of such time-consuming tasks as inserting hand edited meta-tags and time indices, an automated solution is desired. To this end, a procedure was formulated based on an automatic speech recognition technology referred to as word-spotting, in which arbitrary words, phrases, and Boolean combinations thereof could be used for queries. The integration of such technology with the Infusion tool and the lessons learned in the process are reported.