Messing with our new Computer Overlords

Posted on by

Some interesting entailments present themselves If we accept that minds do abductive inference differently than non-minds. For instance it allows us to have some fun with neural networks.

There is an AI experiment from Google that attempts to guess the subject of your doodles. I’ve found It can be easily defeated by simply drawing a random line through the middle of the screen before you begin to draw your picture. Even when your drawing skills are good the AI will usually get it wrong simply because it will assume that the line is an integral part of your drawing and not just a red herring or noise. 

check it out for yourself.https://quickdraw.withgoogle.com/#

Another interesting experiment was conducted by using evolutionary algorithms that interact with Deep Neural Networks designed to classify images.

http://www.evolvingai.org/fooling

Originally the researchers wanted to use the evolutionary algorithms to construct images that humans would recognize. In this way both the DNN and the evolutionary algorithms could be said to fail the “abduction Turing test” because they obviously come to very different conclusions about the concepts that were classified than we do.

What’s interesting is that we can make that evaluation by observation without directly asking the AI questions.

What do you think?

peace

 

 

 

89 thoughts on “Messing with our new Computer Overlords”

  1. Flint: Do you really think that attaching a label to something alters your understanding of it?

    Often it does like it or not. Isn’t that why folks on your side like to call throw around labels like Sarah-Palinesque twerp?

    Seriously, I think that I divide the world in two different sets things with minds and things with out minds.

    I generally feel I have a moral obligation to the first set and not so much to the second one. It’s important to me morally that I don’t categorize entities incorrectly.

    If it turned out that my robot vacuum has a mind I would not want to be mistreating it for example.

    peace

  2. Flint: Why? If you can’t tell the difference, no test will help you. Indeed, for you the test would only provide the only answer you’d permit before you started.

    Why would you say that?

    I can’t tell the difference between between oxygen and carbon monoxide with out a test. In the case of that test I can and have changed my mind as to the answer.

    Flint: YOU, on the other hand, worship a label that applies to nothing.

    Do you really think that “mind” is a label that applies to nothing?

    Flint: And now you know. Reality is my yardstick

    Are you actually saying that minds are not reality? Or are you just flinging insults?

    peace

  3. fifthmonarchyman: I hope you don’t have to work up a head of steam just to answer a simple question or two.

    I think Tom was commenting on the fact that you misread his initial post in this thread. It did not seem at all to me like he was saying the things you thought he was saying.

  4. Mung: It did not seem at all to me like he was saying the things you thought he was saying.

    I’m not really sure I thought he was saying anything at all except that deep neural nets are not all that great and “large-scale systems” are where the action is when it comes to abduction.

    I will admit that I always assume that there is an undercurrent of contempt and derision on his part when he comments no matter the topic. Some of this is probably just a clash of personalities.

    I will readily grant that he is the expert on this one and I am the uncouth bumpkin. I only hope that he would be willing to stoop to having a discussion.

    I would like to learn a little a long the way.

    peace

  5. I don’t always assume Tom is disagreeing with me, lol. But I admit it’s sometimes hard to tell. I think we could both learn from Tom. Just got to find a way to communicate.

    ETA: I feel the same way about Joe Felsenstein. It’s not always easy to tell just who he is agreeing with or disagreeing with, but I don’t think he’s always disagreeing with me.

  6. Tom English: It’s hard to work up a head of steam to engage a sassy, Sarah-Palinesque twerp who misreads even the simplest of comments.

    Misreading is method.

  7. Pedant: Misreading is method.

    Fortunately it’s only theists that misread the comments of others. Here at TSZ that keeps misreading to minimum.

  8. Mung:
    I don’t always assume Tom is disagreeing with me, lol. But I admit it’s sometimes hard to tell. I think we could both learn from Tom. Just got to find a way to communicate.

    ETA: I feel the same way about Joe Felsenstein. It’s not always easy to tell just who he is agreeing with or disagreeing with, but I don’t think he’s always disagreeing with me.

    I’ll work hard at learning along with anyone who seems interested in learning. Have a look at something (not yet complete) I’ve done with you in mind. You can view the code by clicking on icons next to the names of methods. I emphasize that I did the work as much to learn as to teach.

  9. Tom English: Have a look at something (not yet complete) I’ve done with you in mind.

    That’s impressive, Tom. You put me to shame. I rarely document my code so well.

  13. Mung,

    The automatically generated documentation needs to be tweaked by hand. What you cannot see in the documentation is that require 'os' is in my code. Ruby raises an exception if the gem is not available. I catch the exception, and load instead my quick-and-dirty substitute for the gem. Apart from playing with Ruby, I’m catering to (perhaps imaginary) users who will bother to download my code, but will not bother to install Ruby gems.

  14. Tom English: Have a look at something (not yet complete)

    Don’t take this as a criticism.

    The to_s method …

    In my experience this method would not normally take any arguments. In the context of a HuffmanCodec it would return a String representation of a HuffmanCodec not a HuffmanTree. So imo, this to_s method should be defined on the HuffmanTree module.

    But this is very high level ‘critique’. It really depends on what you’re trying to do. Modules in Ruby are perhaps not quite what others coming to Ruby from another background might think.

    What I would like to offer, in addition to just general Ruby knowledge, is help with writing test cases for any of your code that does not have any tests.

  15. Tom English: Apart from playing with Ruby, I’m catering to (perhaps imaginary) users who will bother to download my code, but will not bother to install Ruby gems.

    I feel your pain.

    The most popular solution afiak is yet another gem called Bundler. It ensures that any gem needed for running a Ruby application has been loaded and is the correct version. But we’re probably getting ahead of ourselves. 🙂

  16. Not sure what others think, but FMM signing off every post with ‘peace’ gets a bit tiresome.

  17. graham2: Not sure what others think, but FMM signing off every post with ‘peace’ gets a bit tiresome.

    But you’re not claiming that FMM really signs off every post with ‘peace’, right?

    You and Patrick and Neil ought to join together.

  18. graham2:
    Not sure what others think, but FMM signing off every post with ‘peace’ gets a bit tiresome.

    It’s TheoSpeak for fuck you.

    In Southern, it’s rendered as bless your heart.

  19. petrushka: It’s TheoSpeak for fuck you.

    In Southern, it’s rendered as bless your heart.

    Actually more than anything it is a reminder to myself

    quote:

    So then let us pursue what makes for peace and for mutual upbuilding.
    (Rom 14:19)

    End quote:

    I sometimes tend to forget that while in the heat of a debate.
    Especially with people who don’t like folks like me too much.

    peace

  20. Mung: I don’t always assume Tom is disagreeing with me, lol. But I admit it’s sometimes hard to tell. I think we could both learn from Tom. Just got to find a way to communicate.

    ETA: I feel the same way about Joe Felsenstein. It’s not always easy to tell just who he is agreeing with or disagreeing with, but I don’t think he’s always disagreeing with me.

    In those cases you might try saying “I agree” or “I disagree” and then giving a reason. One or both of us might learn something, and others might too.

  21. Joe Felsenstein: In those cases you might try saying “I agree” or “I disagree” and then giving a reason. One or both of us might learn something, and others might too.

    But then Mung may have to go on the record and that comes with the possibility of being shown to be wrong. So no.

  25. Mung: That’s impressive, Tom. You put me to shame. I rarely document my code so well.

    I find it interesting that you saw something generated by a machine and mistook it for something generated by a mind.

  28. OMagain: But then Mung may have to go on the record and that comes with the possibility of being shown to be wrong. So no.

    This is just absurd. I go on the record all the time here.

    As for being shown wrong, when are you going to take up the task of demonstrating the power of cumulative selection, if ever? As in first defining it’s units and how to measure them.

  29. Fair Witness: I find it interesting that you saw something generated by a machine and mistook it for something generated by a mind.

    It’s not so interesting that you think that Tom had no say in the matter.

  30. Mung: As for being shown wrong, when are you going to take up the task of demonstrating the power of cumulative selection, if ever? As in first defining it’s units and how to measure them.

    What do you think active information is? It’s a formalization of the comparison of cumulative selection to single-step selection that Dawkins made in Chapter 3 of The Blind Watchmaker.

  31. Tom English: What do you think active information is? It’s a formalization of the comparison of cumulative selection to single-step selection that Dawkins made in Chapter 3 of The Blind Watchmaker.

    So the unit to measure “the power of cumulative selection” is the bit, the unit of information?

    Why is it that all the other brilliant people here at TSZ don’t know that?

  32. Mung: So the unit to measure “the power of cumulative selection” is the bit, the unit of information?

    Why is it that all the other brilliant people here at TSZ don’t know that?

    They spend too much time reading UD and too little time reading Paley, Dawkins, Dembski, and Marks?

    We’ll be returning to the “log-improbability is information” fallacy when (if ever) I finish my Huffman coding demo. The probability of hitting the target with with a single random guess is p = 27^{-28}. Let q_m be the probability that the Weasel program hits the target in m or fewer generations. (The value of q_m depends on the number of offspring per generation and the mutation rate. But we don’t need to worry about that here.) By the first (2009) definition, the active information of the Weasel program is the ratio q_m / p. This merely indicates how much more likely the Weasel program is to hit the target in m generations than a single random guess is to hit the target.

    Oh, yeah. Take the negative logarithm of the reciprocal.

        \[I_+ = -\!\log_2 \frac{p}{q_m} = -\!\log_2 p - (-\!\log_2 q_m) = \log_2 q_m - \log_2 p.\]

    The negative logarithm serves no purpose but to make things look more Shannon-information-ish. The role p actually plays here is as a reference level. I think it would make more sense to denominate the quantity in bels than in bits.

  33. fifth:

    If it turned out that my robot vacuum has a mind I would not want to be mistreating it for example.

    Patrick:

    Enslaving it would be just fine, though, right?

    🙂

  37. keiths: Enslaving it would be just fine, though, right?

    No enslaving is always wrong. Even when it’s a robot vacuum.

    If my robot vacuum had a mind I would hope some one would author regulations to insure that it was not mistreated by the meat supremacists who insist on owing them.

    Just because you regulate something does mean you advocate it.

    peace

  39. Patrick: In fact, you yourself trivialize slavery as “temporary and local”.

    It’s not trivializing it’s putting things into perspective.

    An atomic explosion is temporary and local but I would not want to experience one.

    The Bible does not talk much about atomic explosions because for all most all of the readers of the Bible atomic explosions are not a live issue. The same goes for slavery.

    peace

Leave a Reply