SyNapse and Blue Brain Projects Update

As noted before I think the two most promising “Artificial Intelligence” projects are Blue Brain and DARPA SyNAPSE and I’m happy to see in this Boston blog “Neurdon” by some of the SyNAPSE project folks a few of the DARPA bucks going to elaborate on some of the technical goals of the SyNAPSE project:

SyNAPSE seeks not just to build brain-like chips, but to define a fundamentally distinct form of computational device. These new devices will excel at the kinds of distributed, data-intensive algorithms that complex, real-world environment require…

It’s very exciting stuff this “build a brain” competition.  Although I think the theoretical approach taken by Blue Brain is more consistent with what little we know about how brains work, I’d guess SyNAPSE’s access to DARPA funding will give it the long term edge in terms of delivering a functional thinking machine in the 15-20 year time frame most artificial intelligence researches believe we’ll need for that ambitious goal.

My optimism is greater than many because I think humans have rather dramatically exaggerated the complexity of their own feeble mental abilities by a quite a … bit, and I’d continue to argue that consciousness is much more a function of quantity than quality.

Another promising development in the artificial brain area is in Spain where  Blue Brain project partner universities are working on the project:  Cajal Blue Brain

4 thoughts on “SyNapse and Blue Brain Projects Update

  1. Hi. In defense of us poor neuroscientists… I think what we are trying to do is rather complex. Disentangling millions of years of brain evolution and translating the main mechanisms in a chip is not exactly trivial. I share your optimism, but I am sure that the whole enterprise will look simpler… once we got it right!


  2. Thank you Max for checking in. I certainly don’t mean to imply these are trivial tasks although I do think there may be too much focus on generating complex “thinking” algorithms rather than working more towards generating an open and flexible substrate (e.g. a mass of neocortical columns or simulated neocortical columns) from which thinking processes will emerge with only a modest amount of intervention on our part. After all isn’t that how every single one of us learned to think – our parents, environment, and introspection shaped the development of our three pound blob of densely packed neurons?

  3. @Joe Hunkins
    hi Joe. Sure, I agree also on this: evolving a cortical column instead of creating an “adult” version of it is probably the way to go. Again, this is not that simple though, since you have to create the simulated environment, and simulated constraints, that will guide the formation of such an adult structure from an embionic version of it. It sounds like the way to go, but it is surely more complex than approximate the structure of an adult system.

  4. Max thanks again for excellent input. I’m glad your excellent blog is keeping people posted on developments over at SyNAPSE project and I’m also encouraged to hear you are looking along the lines of “evolving” a cortical column rather than trying to come out of the chute with a fully functioning one.

    I think we all need to remind ourselves that very capable, mechanistic intelligences come online every minute all over the world from some very basic biological and sociological processes. These intelligences, aka “newborn babies” take only a few years to become very capable thinkers. My kids are both very sharp, and they were built almost from scratch by me and my wife. Of course we did have 4 billion years of DNA evolution to help lay the groundwork, but… one could argue that our job is less to “invent” AI than to engineer copycat systems based on largely observable biological analogs.