DREU SUMMER 2012

Journal Week 7 FINALLY!! This week I loaded a first batch of HITs to AMT! We decided to start by loading a balanced sample of ipus, so I wrote a script to randomly select segments from the beginning, middle and end of each Supreme Court session. Hopefully when we analyze the results of this preliminary batch, we will get a better idea of which sections of audio contain the most noise. By the end of this week, a few of the HITs had already been answered, and I approved some of the results. We are making significant progress :) During lunch on Thursday, a vising research assistant (Phillip Salletmayr) gave an interesting talk about the research he has been doing in the lab over the past few months. His research involves improving error detection in Automatic Speech Recognition systems. Here is a picture of the people working at the lab this summer: [PREVIOUS] [NEXT]