full transcript
From the Ted Talk "Margaret Gould Stewart: How YouTube thinks about copyright"

Unscramble the Blue Letters

But how do we know that the user's video was a copy? Well, it starts with content owners delivering assets into our database, along with a ugeas policy that tells us what to do when we find a match. We compare each upload against all of the reference files in our tbaaesad. This heat map is going to show you how the brain of the system works. Here we can see the original eferrceen file being compared to the user generated content. The system compares every moment of one to the other to see if there's a match. This means that we can identify a match even if the ycpo used is just a portion of the original eilf, splay it in wols motion and has degraded audio and video quality. And we do this every teim that a video is uploaded to uoybuet. And that's over 20 hours of video every minute. When we find a match, we apply the policy that the rights owner has set down.

Open Cloze

But how do we know that the user's video was a copy? Well, it starts with content owners delivering assets into our database, along with a _____ policy that tells us what to do when we find a match. We compare each upload against all of the reference files in our ________. This heat map is going to show you how the brain of the system works. Here we can see the original _________ file being compared to the user generated content. The system compares every moment of one to the other to see if there's a match. This means that we can identify a match even if the ____ used is just a portion of the original ____, _____ it in ____ motion and has degraded audio and video quality. And we do this every ____ that a video is uploaded to _______. And that's over 20 hours of video every minute. When we find a match, we apply the policy that the rights owner has set down.

Solution

  1. slow
  2. plays
  3. reference
  4. database
  5. youtube
  6. file
  7. time
  8. usage
  9. copy

Original Text

But how do we know that the user's video was a copy? Well, it starts with content owners delivering assets into our database, along with a usage policy that tells us what to do when we find a match. We compare each upload against all of the reference files in our database. This heat map is going to show you how the brain of the system works. Here we can see the original reference file being compared to the user generated content. The system compares every moment of one to the other to see if there's a match. This means that we can identify a match even if the copy used is just a portion of the original file, plays it in slow motion and has degraded audio and video quality. And we do this every time that a video is uploaded to YouTube. And that's over 20 hours of video every minute. When we find a match, we apply the policy that the rights owner has set down.

ngrams of length 2

collocation frequency
rights owners 3
id system 3
content id 3

ngrams of length 3

collocation frequency
content id system 3

Important Words

  1. apply
  2. assets
  3. audio
  4. brain
  5. compare
  6. compared
  7. compares
  8. content
  9. copy
  10. database
  11. degraded
  12. delivering
  13. file
  14. files
  15. find
  16. generated
  17. heat
  18. hours
  19. identify
  20. map
  21. match
  22. means
  23. minute
  24. moment
  25. motion
  26. original
  27. owner
  28. owners
  29. plays
  30. policy
  31. portion
  32. quality
  33. reference
  34. rights
  35. set
  36. show
  37. slow
  38. starts
  39. system
  40. tells
  41. time
  42. upload
  43. uploaded
  44. usage
  45. user
  46. video
  47. works
  48. youtube