Artificial Intelligence (AI)

Posted on
Page
of 22
  • Guess it depends how you define 'intelligence'. If alphago can beat the best Go player in the world without just calculating all possible moves like that Chess bot big blue, but by machine learning.

  • As for AI being boring, are you joking? This isn't 3D TV.

  • Not by a long mile. I love this biz. I firmly believe we should be pouring resources into this at an accelerated rate for a faster return on the research. Not even financially, but in every avenue of our lives.

    My issues to pick is the broad brush the concept of artificial intelligence is being painted with. When you listen to people in the field talk about 'it', they make a pretty concise distinction between machine learning and strict AI.

    Which brings me to the point I raged on over. TV as predictably as it does dumb down, does imprint people with the idea that we are somewhere nearer to producing some form of AI that is, as the basis for it's idea, self ware.
    I'm not even talking about the turing test. That's been debunked as flawed many times over.
    Am I griping over nothing? Probably, but skipping the 'how' and jumping to the theoretical dramatics of the AI physically being there/becoming aware only serves to cripple itself when people start realising we are, scientifically, nowhere near creating one.

    The hell do I know though, I just fix bicycles.

  • Fair enough. I was more impressed with the DeepMind thing learning to play Atari games than the fake human. I guess the fake human is a nice little fluffy tie in with their show but at least they touched on some actual AI projects briefly and some of the issues we might need to be thinking about before we unleash learning tech on the masses.

  • we don't even understand how ours works,

    I understand how mine works. It's pretty simple. Involves some beer and my missus

  • Basically...we don't know how the fuck to make one because we don't even understand how ours works, and if we do, we don't know how to power it.

    This. I was getting really excited about AI in the pub once and a mate just said that we don't understand our own minds, so how can we create another?
    Sometimes I think of it as a simple input/output/computation problem though. When we create a system that can process as much information (sights, sounds, touch etc) as quickly as a human, will it just 'be' intelligent? Will it need an aim? The atari example gives the algorithm rewards with the in-game points. Scary to imagine more sophisticated bots could have rewards for much more sinister things. Without an aim, will it just try things at random (like a baby)? A robot with the power of a human but the whims of an infant is scary too.

    If all it took was 3D printing my face, I'd have accomplished world domination by now.

    Have done, can confirm I haven't achieved world domination.

  • When we create a system that can process as much information (sights, sounds, touch etc) as quickly as a human, will it just 'be' intelligent? Will it need an aim?

    Will it be self aware, conscious?
    Are we really conscious, are animals or even plants?
    Or is the fact that we feel self aware just an instinct that perhaps humans have elaborated more than other beings?

  • Is artificial intelligence just the copying of our own minds though? Seems a waste, given most of us are fucking idiots.

    We seem to be building working quantum computers at the moment and last time I checked we were pretty vague on how that shit worked too.

    It's not just about processing as much (more) information as a human though, it's about it learning how to do something with the information and changing how it's using the information as it is fed more information. Rather than Big Blue just being really fast at computing ALL the moves on a chess board we have Alphago which learns Go strategies the more it plays.

    Scary, yes, which is where all the ethical discussions come into play - what controls should be in place when testing this kind of thing? What safeguards should be built in? Self-driving cars that adapt and learn road maneuvers seem to be the hot topic on this at the moment. See all the ethical dilemas raised in that thread.

    Red Dwarf new series had them printing a crew out. Add in a "paper jam" for hilarious consequences...

  • we don't understand our own minds

    Anyone who thinks that could do worse than read:

    Plato's Phaedo
    Plato's Republic
    Plato's Parmenides
    Plato's Theaetetus
    Plato's Sophist
    Plotinus' The Enneads
    Kant's Critique of Pure Reason

    Just for starters. :)

    Of course, the fundamental uncertainty at the root of it all persists in the face of the best philosophy, and 'artificial intelligence' is still baloney. Can we create very sophisticated computer systems mocking up how we think? Yes, probably. Can we create something that is alive in the same way we are and hence has 'intelligence'? Probably not.

  • So much this. I've read none of your suggestions unfortunately, but the Philosophy and Ethics behind it are astounding. At this point, I'm not even talking about "I think, therefore I am" marlarkay. I'm talking about...how do we think? Someone out there might enlighten us, but It still floors me how much information we literally dump/ignore.

    Shit...at this juncture, the human brain actually creates content for us to fill in blanks or to save our 'self' from scarring if we experience trauma. We lie, we cheat and we rationalise. How do we (and should we even allow ourselves to?) create an AI with said human facets? Is it even an AI without these things or just a glorified iPhone.

    @hippy Going quantum would create a fair ripple from what I've read. Simply because we'll abandon the rapid task switching mechanism to infinite possibilities and thus maybe come closer to replicating some simplistic version of the human mind. Like I am right now. Hangry. Hungry and angry but at the same time I can't be arsed to go acquire food so I'm getting hungrier and thus madder, which spends energy and makes me hungrier.
    Infinite loop.

    Beer time.

  • (sorry for cross post) Google DeepMind's next target is starcraft apparently, should be significantly more complex than Go

  • Oh, that should be fun.
    The "problem" with Starcraft would be the speed of the decisionmaking.

    If they can have the machine act fast enough to overcome a human player in clicks/minute, then I see little chance on the human players side. If the decisions done by the machine take longer than the decisions made by the human, the human wins.

  • Yeah it should be interesting. To get the algorithm to scale to the number of potential actions at any given time will be a pretty big challenge I reckon. And then like you say to actually decide on which action to take.
    They seem to have skipped forward to a hugely complex game, I'd have thought they would have gone for a racer first (clear goals and limited options for actions) but then basic racing games have already been done by machine learning.
    https://youtu.be/S9Y_I9vY8Qw

  • Black Mirror spoiler alert!

    Black Mirror: San Junipero. At the end the two main characters have their minds uploaded to the cloud. Would this not be a copy of their minds and therefore not really life after death?

  • Douglas Hofstadter on Trump, ornamental elephants, education, twitter and constructivism
    http://www.idsnews.com/article/2016/12/analogies-and-word-choice-shape-political-discourse

  • From The Medium about our relationship with AI/robots

    Here are the Seven Stages of Robot Replacement:
    1.
     
    A robot/ computer cannot possibly do the tasks I do.
    2.
    [Later.] 
    OK, it can do a lot of those tasks, but it can’t do everything I do.
    3.
    [Later.]
    OK, it can do everything I do, except it needs me when it breaks down, which is often.
    4.
    [Later.]
    OK, it operates flawlessly on routine stuff, but I need to train it for new tasks.
    5.
    [Later.]
    OK, OK, it can have my old boring job, because it’s obvious that was not a job that humans were meant to do.
    6.
    [Later.] 
    Wow, now that robots are doing my old job, my new job is much more interesting and pays more!
    7.
    [Later.]
    I am so glad a robot/computer cannot possibly do what I do now.
    [Repeat.]

  • Spent the last couple of months getting trough Nick Bostrom's 'superintelligence' and Martin Ford's 'Rise of the Robots'.
    Whilst neither jump to firm conclusions, both paint a pretty convincing picture of a troubled future.

    My own two pence is that you'd be wasting your time getting philosophical about whether an AI can truuuuly replace a human being. It's a bit like arguing that McDonald's are of no consequence until they are awarded a Michelin star. I reckon your future will be filled with plenty of inane, unsatisfying replacements for genuine human intelligence. But you'll grudgingly go along with it in the same way that you keep ending up on Ryanair flights despite repeatedly promising yourself never to do so again.

  • I reckon your future will be filled with plenty of inane, unsatisfying replacements for genuine human intelligence

    Just like the present only shinier...

  • Woah.. The google language bot invented its own language to aid its ability to translate other languages better...
    ...without being programmed to do this

    https://medium.freecodecamp.com/the-mind-blowing-ai-announcement-from-google-that-you-probably-missed-2ffd31334805#.nka41khb5

  • Noice,
    However, isn't the secondary point of CAPTCHA to use it as training material for a Neural Network.
    The whole text based CAPTCHA thing got infuriating as they made the characters harder and harder to decipher, then boom, that CAPTCHA style completely disappeared and we are asked to point out road signs and cars out of a grid of 9-25 squares( Clearly training material for driverless).
    I'm not going to shit my pants until CAPTCHA changes to 'Point out places on this map which you don't mind testing nuclear weapons on' OR, as with the daily mail article a page back, 'follow the moving bad man target with your mouse cursor and avoid the good target for 15 seconds'

  • “Are you a robot”

    Well yes, thanks for asking.

  • Post a reply
    • Bold
    • Italics
    • Link
    • Image
    • List
    • Quote
    • code
    • Preview
About

Artificial Intelligence (AI)

Posted by Avatar for skydancer @skydancer

Actions