Tuesday, July 15, 2008

"In their own words": political videos meet Google speech-to-text technology



(Cross-posted from the Official Google Blog)


In this U.S. election year, what information could be more important than the candidates' own words to describe their views, actions and platforms?

Our teams have been working to develop tools to make it easier for people to track election-related information. A few months back, YouTube encouraged everyone to participate in the discussion process through the CNN/YouTube debates, Google Checkout offered an easy and fast way for individuals to make contributions to political candidates, and the Geo team created maps and layers to inform voters during elections.

Today, the Google speech team (part of Google Research) is launching the Google Elections Video Search gadget, our modest contribution to the electoral process. With the help of our speech recognition technologies, videos from YouTube's Politicians channels are automatically transcribed from speech to text and indexed. Using the gadget you can search not only the titles and descriptions of the videos, but also their spoken content. Additionally, since speech recognition tells us exactly when words are spoken in the video, you can jump right to the most relevant parts of the videos you find. Here's a look:


In addition to providing voters with election information, we also hope to find out more about how people use speech technology to search and consume videos, and to learn what works and what doesn't, to help us improve our products.

The gadget only searches videos uploaded to YouTube's Politicians channels, which include videos from Senator Obama's and Senator McCain's campaigns, as well as those from dozens of other candidates and politicians. It usually takes less than a few hours for a video to appear in the index after it has been published on YouTube. Candidates can control the videos that appear in the gadget by managing the content they upload to YouTube. While some of the transcript snippets you see may not be 100% accurate, we hope that you'll find the product useful for most purposes. Speech recognition is a difficult problem that hasn't yet been completely solved, but we're constantly working to refine our algorithms and improve the accuracy and relevance of these transcribed results.

To try it out, just visit our iGoogle gadget page. We welcome your feedback, so please feel free to leave a comment while you're there.

No comments: