Tuesday, April 23, 2019

Neural Networks


Your humble poster here once headed an Artificial Intelligence (AI) software project. One of my best programmers quit because of his frustration over complex nested conditionals (neural networks). The ultimate success of AI will depend on the ability to identify nodes in such networks as well as being able to deal with the almost infinite levels of conditional complexities.

It seems to this author that one of Trump’s great strengths as a policy maker is his ability to include in our country’s decision-making process an uncanny knack to identify previously ignored nodes in our neural network — unfair trade alliances, over regulation, Russian pipeline into Germany, NATO slackers, open borders, steel and aluminum being strategic US native industries, etc.

The jury is somewhat out on Trump’s ability to navigate these new expanded decision trees ... but he is still way, way ahead of our last four presidents.

Back to the subject at hand. Yes, quantum computers may someday be able to handle the huge number of complex conditionals involved in true AI. But, if a butterfly flapping it wings in Japan can cause a tornado in Alabama, I still have some doubts about the logistics of the data amassing project needed to account for all possible inputs into such an ultimate AI breakthrough.

How about it Elon?

4 comments:

  1. Veritas12:22 PM

    Maybe the programmer quit because he realized the boss was a madman, who would one day write a blog entry praising a dingbat autocrat as a deep thinker.

    ReplyDelete
    Replies
    1. What’s the Latin word for sourpuss?

      Delete
    2. Veritas10:40 PM

      Trvmpswab

      Delete
  2. I looked it up ... “veritas” means no Asians wanted at our college.

    ReplyDelete