Please turn on javascript in your browser to play chess.
Science Forum

Science Forum

  1. Standard member sonhouse
    Fast and Curious
    23 Oct '17 20:14
    https://phys.org/news/2017-08-artificial-intelligence-gravitational-lenses-million.html#nRlv
  2. 23 Oct '17 21:38 / 5 edits
    this is a great example of how artificial neural networks can often process data massively faster and more efficiently than other methods.

    One of the things I hope to eventually get around to researching is representing that kind of neural network-like behavior with software in an ordinary non-neural network computer and have that fully integrated into a knowledge-based system and possibly one designed to use genetic algorithms to design that software itself and even the genetic algorithms themselves with a continuous cycle of self-improvement and learning. This is to get around some of the hardware limitations of neural networks. But I believe before I can even think of doing that, I will need to first produce an extremely impressive knowledge-based system else that would be a non-starter.
  3. Standard member sonhouse
    Fast and Curious
    24 Oct '17 13:54 / 1 edit
    Originally posted by @humy
    this is a great example of how artificial neural networks can often process data massively faster and more efficiently than other methods.

    One of the things I hope to eventually get around to researching is representing that kind of neural network-like behavior with software in an ordinary non-neural network computer and have that fully integrated into a kn ...[text shortened]... o first produce an extremely impressive knowledge-based system else that would be a non-starter.
    I think eventually quantum computers, neural netwoks and traditional computers will be tied together, each one giving results as a result of the collaberation between those three ways to compute.

    Look at my post 'memory that computes'
  4. Standard member moonbus
    Uber-Nerd
    24 Oct '17 19:03
    Yes, this is big news indeed. Analyzing large data sets in powerful new ways holds much promise in medical diagnostics too. We are right on the threshold of discovering pre-symptomatic indicators for various pathologies, including dementia.