English subtitles

← 02_Supervised Learning

Get Embed Code
3 Languages

Showing Revision 1 created 01/19/2014 by Cogi-Admin.

  1. This class is divided into three subclasses
  2. >> Mm-hm.
  3. >> Three parts. They are supervised learning
  4. >> Yeah.
  5. >> Unsupervised learning, and reinforcement learning. So
  6. what do you think supervised learning is?
  7. >> So I think of supervised learning as being the problem of taking labeled
  8. data sets Gleaning information from it so that you can label new data sets.
  9. >> That's fair.
  10. I call that function approximation.
  11. >> Mm.
  12. >> So, here's an example of supervised learning. I'm
  13. going to give you an input and an output, and
  14. I'm going to give them to you as pairs, and
  15. I want you to guess what the function is.
  16. >> Sure
  17. >> Okay? Okay. 1:1.
  18. >> Uh-huh. Two four.
  19. >> Wait. Hang on. Is one input and one the output?
  20. >> Yes.
  21. >> And two the input and four the output?
  22. >> Correct.
  23. >> Alright, I'm on, I think I'm on to you.
  24. >> Three.
  25. >> Uh-huh.
  26. >> Nine.
  27. >> okay.
  28. >> Four. Sixteen.
  29. >> Nice.
  30. >> Five. Twenty
  31. five. Six. Thirty six. Seven. Forty nine.
  32. >> Nice. This is a very hip data set.
  33. >> It is. What's a [UNKNOWN]?
  34. >> It's hip to be squared.
  35. >> Exactly. Maybe. Oh.
  36. >> So if you believe that's true, then tell
  37. me if the input is ten, what's the output?
  38. >> 100.
  39. >> And that's right, if it turns out, in fact, that the function is X
  40. squared. But the truth is we have no
  41. idea where's the function's X squared, not really.
  42. >> I have a pretty good idea.
  43. >> You do?
  44. >> Well, just.
  45. >> Where did that idea come from?
  46. >> It
  47. comes from having spoken with you over a
  48. long period of time. And plus you know, math.
  49. >> And plus math, well, I'm going to
  50. >> You can't say I'm wrong.
  51. >> You're wrong.
  52. >> uhh, you said I was wrong.
  53. >> No, you've talked to me for a long time and plus math, I
  54. agree with that, but I'm going to claim that you're making a leap of faith.
  55. >> Mm-mm
  56. >> Despite being a scientist, by deciding that
  57. the input is 10 and the output is 100.
  58. >> Sure, I would agree
  59. with that.
  60. >> And what's that leap of faith?
  61. >> Well, I mean, from what you told me, it's still consistent with
  62. lots of other mappings from input to output like ten gets mapped to 11.
  63. >> Right, or everything's x squared except ten.
  64. >> Sure.
  65. >> Or everything's x, x squared up to ten.
  66. >> Right, that would be mean.
  67. >> That would be mean.
  68. >> But it's not logically impossible. What'll be the medium.
  69. >> [SOUND]
  70. >> Thank you very much. I was saving that one up.