Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

    •  
      CommentAuthorAngus
    • CommentTimeJul 22nd 2015
     
    I meant senior military personnel, of course.
  1.  
    There is a Politics thread, you know. I made that one too.
  2.  
    https://www.youtube.com/watch?v=AbjVdBKfkO0

    Yann LeCun, a neural net luminary
    •  
      CommentAuthorAngus
    • CommentTimeJul 22nd 2015
     
    Posted By: Andrew PalfreymanThere is a Politics thread, you know. I made that one too.


    Are you suggesting that discussion of Politics in a thread called Artificial Intelligence is inappropriate?
  3.  
    You tell me, since you started it
    •  
      CommentAuthorAngus
    • CommentTimeJul 22nd 2015
     
    OK. No it's not.

    This is the moletrap and we don't have Moderator Nazis enforcing kadavergehorsamkeit to the strictest interpretation of the thread title. This is the Free Moletrap, where creativity reigns and reigns and reigns.
  4.  
    https://www.youtube.com/watch?v=W15K9PegQt0
    Go to about 12:00. Rewiring the auditory cortex to see. Amazing.
  5.  
    Offhand thought. Instead of teaching these multi-teraflop GPU-based beasts to more effectively advertise consumer products, why not teach them mathematics and physics in order to explore new theoretical ground?
    •  
      CommentAuthorAngus
    • CommentTimeJul 22nd 2015
     
    Posted By: Andrew Palfreymanhttps://www.youtube.com/watch?v=W15K9PegQt0
    Go to about 12:00. Rewiring the auditory cortex to see. Amazing.


    Experimental proof in humans.. I'm surprised he didn't mention it.
  6.  
    That gels with Ng's point that the brain seems to use the same learning algorithm universally.
    I wonder if he's using his visual cortex?
    But nobody has yet rerouted a human eyeball to a human auditory cortex (with or without LGN).
    Or indeed a human ear to the visual cortex (with or without LGN).

    The question really is - what's the brain algorithm?
  7.  
    This will give you about 8.5 Teraflops (single precision) for around $4000
    http://www.amazon.com/Nvidia-Accelerator-passive-cooling-900-22080-0000-000/dp/B00Q7O7PQA
  8.  
    Posted By: Andrew PalfreymanOffhand thought. Instead of teaching these multi-teraflop GPU-based beasts to more effectively advertise consumer products, why not teach them mathematics and physics in order to explore new theoretical ground?
    I've looked around, and nobody seems to be talking about this. Perhaps it's a premature wish.

    On the other hand, wouldn't it be cool to have a machine give physics a turbo boost?
    • CommentAuthorloreman
    • CommentTimeJul 22nd 2015
     
    Posted By: Andrew PalfreymanBut nobody has yet rerouted a human eyeball to a human auditory cortex (with or without LGN).
    Or indeed a human ear to the visual cortex (with or without LGN).



    But some people end up that way by accident. That's an interesting study for the purposes of this discussion.
  9.  
    I'm not aware of that happening because it requires surgical intervention, does it not?
  10.  
    Chomsky on AI
    https://www.youtube.com/watch?v=0kICLG4Zg8s

    Gets it wrong in the first five minutes.
    • CommentAuthorloreman
    • CommentTimeJul 23rd 2015
     
    Posted By: Andrew PalfreymanI'm not aware of that happening because it requires surgical intervention, does it not?


    AFAIK some people with head injuries consequent on accidents end up being able to smell the colour Red.
    •  
      CommentAuthorAngus
    • CommentTimeJul 23rd 2015
     
    Synaesthesia doesn't even require a head injury. Some are naturally blessed with it. Not sure however what it tells us about learning algorithms.
  11.  
  12.  
    What happens when we combine Deep Learning with 3D printing?
  13.