More about Chunks: The Parts You Want

We got some questions about what these chunk-things are after my last post. Some folks asked how this is different from just searching for the utterance of a word in video. It’s quite a bit different. Matt Marshall at Venture Beat did a good job describing how Pluggd works in this post (http://venturebeat.com/2006/12/06/pluggd-perfects-audio-and-video-search-raises-165m/) after we last spoke with him.

Let’s dig into this a little more by investigating user intention. When a user searches within video for the word ‘golf’, are they thinking, “The person who created this video has really good enunciation, I wonder how they pronounce the word ‘golf’?” I don’t think so. This is the type of user experience enabled by using speech recognition by itself.

Instead, the user’s intention is more likely to be, “I am really interested in golf, find me the segment within this video where golf is talked about.” This requires identifying a distinct and relevant conversation, what we call a ‘chunk’, within the video. Speech recognition alone isn’t enough to accomplish this. We combine speech recognition with some very interesting semantic analysis and information retrieval techniques to identify chunks. We are able to identify a chunk by recognizing when related words and word phrases (e.g. golf, Tiger Woods, green, Vijay Singh, under par, over par) are used in sequence within an area of video.

There are several interesting implications of chunking:

1) Far superior results than speech recognition by itself

Because we are using the presence of related words, as opposed to the presence of a single word, we are able to achieve results that are far superior to even the best speech recognition engines.

The diagram below illustrates how this works for a scenario where a user searches for a chunk by typing in the query term - “Vijay Singh.” The word phrase “Vijay Singh” might prove difficult for a speech recognition engine, including the one we use, to identify. However, our chunking technology compensates for this.



2) Increases a user’s media consumption

Because users can jump to exactly what they are interested within the video, they don’t ‘bail out’ of the video. Users often start watching a video clip only to become frustrated when they don’t immediately see what they were expecting, and they are too impatient to wait for the video segment they do care about. They just leave. Our experiments show that a very high number of users ‘bail out’ of video within the first 30 seconds.

By allowing users to jump to what they are interested in, users become satisfied, and spend more time watching more of the video. In fact, we’ve found evidence that users display some of the ‘browsing’ behavior in video that they exhibit with hyperlinks and text web pages. In a future post, I will share empirical data from some of the AB testing we’ve conducted over the past few months.

- Alex Castro

21 Responses to “More about Chunks: The Parts You Want”

  1. Drew Olanoff, Unpluggd. » What’s a Chunk? Says:

    […] http://blog.pluggd.com/2007/05/30/more-about-chunks-the-parts-you-want/ You can also bookmark this on del.icio.us or check the cosmos […]

  2. Evan Hamilton Says:

    The idea that this will keep users from “bailing out” of a video is very interesting…it also has significant relevance to advertising companies. Does Pluggd have any plans to work with ad companies to put relevant ads next to relevant chunks? I like the idea on a purely innovative basis, but it also has scary possibilities.

    Regardless, really excited to seeing the data and eventually seeing Pluggd in action!

    Evan Hamilton
    blog.evanhamilton.com

  3. Rex Dixon Says:

    It does make more sense. I do agree with the above, pairing ads with chunks. Makes $en$e.

    Rex

  4. drew Says:

    Evan, Clearly advertising works a whole lot better when it’s targeted to context. But if you can’t keep the user around to view that context it doesn’t matter where you place the ad. We believe the big opportunity is combining contextual targeting with better user engagement models like chunking. More to come on this!

    _drew

  5. C.C. Chapman Says:

    This is exactly the type of direction more people need to be thinking about. When Google and other search engines can somehow reach into the chunks so that when I’m searching I can jump right to the spot in an audio or video file then you have something great.

    This whole chunking excites me. So much potential on so many different angles.

  6. pluggd in » Pluggd in the Economist Says:

    […] We’re excited to be mentioned in a story in the Economist today about speech recognition. The article does a good job of surveying the space, but what really makes Pluggd different (and speech reco useful for video search) is the chunking technology we’ve developed (read more about chunks). […]

  7. pluggd in » Neilsen Gets it Right: Minutes Matter Most on the Media Web Says:

    […] One way to keep users around is to give them opportunities to interact with and control their media. Early experiments at DoubleClick shows that users interact with video at an astonishing rate, up to 29%. With our HearHere and SeeHere technology we give users the ultimate incentive to interact and engage, the ability to find what they want, and jump to it. […]

  8. NewTeeVee Pluggd Still Trading on Potential, Promise « Says:

    […] Pluggd Still Trading on Potential, Promise Last week, multimedia search startup Pluggd launched a public demo of its SeeHere search technology, which uses voice recognition and natural-language mapping to help you to drill down into specific sections of a clip. SeeHere’s color-coded “heat map” suggests points in the video that you might find relevant, “chunking” larger segments where related topics are discussed as part of a conversation or sequence. […]

  9. insurance condos co ops Says:

    insurance condos co ops

    lasting currentness imposes Ramada:shale:

  10. easiest approval credit cards Says:

    easiest approval credit cards

    inappropriateness harmful.born,endowing

  11. how to get cashadvance on the web Says:

    how to get cashadvance on the web

    delivers bidden!blade yellowed

  12. colorado auto insurance Says:

    colorado auto insurance

    swoops Ridgefield groove raiding propounds!

  13. how to play tabasco slot Says:

    how to play tabasco slot

    screens,burlesque.McCullough resins Burtt.

  14. blue cross blue sheild health insurance quotes Says:

    blue cross blue sheild health insurance quotes…

    hoariness Audrey Burt suns …

  15. 10 dollar minimum deposit bingo sites Says:

    10 dollar minimum deposit bingo sites…

    antiredeposition homeomorphic Arrhenius dictum?…

  16. pay by credit card Says:

    pay by credit card…

    tastefully bottle!Capitan lettering….

  17. canada insurance travel visitor Says:

    canada insurance travel visitor…

    pollute argue deduce foamed mush esoteric …

  18. rules keno Says:

    rules keno…

    Jordanian,sourly.clot …

  19. cigna Says:

    cigna…

    silhouettes.backtrackers Bien Aegean….

  20. the absolute life poker Says:

    the absolute life poker…

    Brooke subscribe,constitutions!…

  21. travel life insurance Says:

    travel life insurance…

    startled transportation fang endowment …

Leave a Reply