From Brains to Machines : Aug 2, 2011 in Silicon Valley


I am an invited speaker for at an NSF-sponsored special symposium, From Brains to Machines, to be held in San Jose, CA. This symposium is part of IJCNN 2011, the flagship conference of the International Neural Network Society.  The official title of my talk is How to work towards a mathematical understanding of the brain, but that is really the best academic sounding title I could come up with.  What I actually want to talk about is How to build a brain without solving it.  I hope I get to show some of our recent work at Vicarious Systems
This entry was posted in Uncategorized. Bookmark the permalink.

8 Responses to From Brains to Machines : Aug 2, 2011 in Silicon Valley

  1. Dave says:

    Cool! Please post your talk (or at least your slides) when you get a chance. Congrats on the invite!

  2. Deepak Nath says:

    Congratulations Dileep.! Really feel that California is THE place for such research! Too bad not able to find enough researchers in this field in india.

  3. Sean says:

    I echo the last comment. Is there any publicly available information from the presentation that you made?

  4. Dave says:

    Sean, I miss your blog.

  5. Sean says:

    Dave, thanks for missing my blog. Things died on there for two main reasons. First, I don’t feel that I have sufficient technical knowledge of the AI/machine learning field to do a semi-technical blog on it, and second, my interests span too far beyond Numenta specifically to have a blog about it.

    What interests me about Vicarious over Numenta is that Hawkins seems to care more about emulating/understanding the brain, while Vicarious is more interested in using knowledge of the brain to create AI. Hawkins doesn’t seem to have much interest in AI per say.

    Dileep: Having given you that compliment, I have got to say that I thought your Singularity Summit talk started out well, but became very, very lame when you said that you could say absolutely nothing about your new technology because of your lawyer. That sounds rather implausible, and leads people to believe that your technology is smoke and mirrors rather than something real.

    • dileep says:

      Sean, Dave: My IJCNN talk had pretty much the same content as the singularity talk, sans the Wright brothers analogy. So if you watched the singularity talk, that is pretty much it.

      Sean: The IP concern is real, especially because what we are building is easily analyzable and understandable. I understand that it could come across as smoke and mirrors to some. Unfortunately, until we demonstrate something in public, I don’t see another way out. Right now only the technically inclined will be able to understand the progress we are making (because it is all in graphs and metrics) and if we give out the technical details we will be giving away too much. Creating a public demo of the technology is a distraction for now because we have several more problems to solve. What we wanted convey through the talk was not that we have solved all the problems, but that we have a good way of approaching the problem and that it seems to be bearing fruit. Do you think you would have got the same impression if I had just left out the lawyer comment (I actually went off script) but presented the same content?

  6. Sean says:

    The lawyer comment actually didn’t bother me that much (i am one, after all). It was more that I did not get any sense at all of what you are developing or how well it works. Even if you don’t give away the tech details (quite understandable, its not as though Apple is about to tell us exactly how Siri works), even computer vision benchmarks would add a lot. If you can show that you are beating Hinton, Ng, etc at vision tasks you would start to really impress people. Still, up to that point it was very interesting.

  7. Pingback: max

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>