Please Share

Friday, November 9, 2007

MIT creates search engine to aid students to find topics within video taped lectures

Wow, 2 AI / natural language recognition search related posts in one day! Faculty and students of MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed a search engine that allows students to search video taped lectures as they might a text based web site. Lectures that have been video taped are fed to speech to text software that captures a transcript, the text is then divided into groups of words, or topics, that are analyzed for key-word density. Students can then query the search engine for topics and be guided directly to the portion of the video that best matches the query, therefor saving the trouble of scanning an entire lecture to find a specific portion. Best of all it's not just for MIT students, in fact its openly available thanks to MIT's OpenCourseWare initiative. Guess I know how my weekend is going to be spent! Give it a try! (Requires Real Player Plug in)