Slashdot Mirror

← Back to Stories (view on slashdot.org)

The US Military Wants To Teach AI Some Basic Common Sense (technologyreview.com)

Posted by BeauHD on from the easier-said-than-done dept.

DARPA, the research arm of the U.S. military, has a new Machine Common Sense (MCS) program that will run a competition that asks AI algorithms to make sense of questions with common sense answers. For example, here's one of the questions: "A student puts two identical plants in the same type and amount of soil. She gives them the same amount of water. She puts one of these plants near a window and the other in a dark room. The plant near the window will produce more (A) oxygen (B) carbon dioxide (C) water." MIT Technology Review reports: A computer program needs some understanding of the way photosynthesis works in order to tackle the question. Simply feeding a machine lots of previous questions won't solve the problem reliably. These benchmarks will focus on language because it can so easily trip machines up, and because it makes testing relatively straightforward. Etzioni says the questions offer a way to measure progress toward common-sense understanding, which will be crucial. [...] Previous attempts to help machines understand the world have focused on building large knowledge databases by hand. This is an unwieldy and essentially never-ending task. The most famous such effort is Cyc, a project that has been in the works for decades. "The absence of common sense prevents an intelligent system from understanding its world, communicating naturally with people, behaving reasonably in unforeseen situations, and learning from new experiences,"https://www.darpa.mil/ Dave Gunning, a program manager at DARPA, said in a statement issued this morning. "This absence is perhaps the most significant barrier between the narrowly focused AI applications we have today and the more general AI applications we would like to create in the future."

94 comments

  1. Teach AI Some Basic Common Sense by nospam007 · · Score: 3, Insightful

    Good luck with that.
    Common sense isn't very common I'm afraid.

    1. Re:Teach AI Some Basic Common Sense by mlheur · · Score: 1

      1) I've recently started referring to it as uncommon sense.
      2) Human's can't do common sense so we need computers to do it for us?
      3) We can't do common sense, what makes us think we can teach computers to do it?

      Easy to see where this won't go...

    2. Re:Teach AI Some Basic Common Sense by Anonymous Coward · · Score: 0

      I'm afraid your first post trite spam isn't very insightful and your name isn't sarcastic enough.

    3. Re:Teach AI Some Basic Common Sense by Anonymous Coward · · Score: 0

      Good luck with that.
      Common sense isn't very common I'm afraid.

      Oh come on. I'm sure the US Military has tons of common sense. [snicker]

    4. Re:Teach AI Some Basic Common Sense by Aighearach · · Score: 4, Interesting

      Yeah, even in the attempted example question, the "common sense" answer is A, but the "actual fact" answer is A and B. A during the day, B at night. C is most likely also true in real scenarios.

      My goal in writing automated systems is to make less of the mistakes known by the moniker "common sense," not to make more of them.

      If you lack information and are forced into action, "common sense" might be a decent least-bad semi-random choice, but it should never be expected to be correct or optimal.

    5. Re:Teach AI Some Basic Common Sense by gweihir · · Score: 1

      Good luck with that.
      Common sense isn't very common I'm afraid.

      Indeed. And in particular, the military is the last place where you find it.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    6. Re:Teach AI Some Basic Common Sense by TomGreenhaw · · Score: 2

      Actually its much easier than you think.

      Common sense in machine learning is accomplished using a Gaussian distribution (bell curve). You can even do it in databases using standard deviation functions.

      --
      Greed is the root of all evil.
    7. Re:Teach AI Some Basic Common Sense by Anonymous Coward · · Score: 0

      ... "actual fact" answer is A and B.

      No, it isn't. You missed the real question: Plant 1 "will produce more" what? To answer that question requires knowing 2 "actual" facts: One you've mentioned and that plant 2 will produce only B. The commonsense answer is thus, correct.

      ... C is most likely also true.

      Since combusting of carbohydrates always produces water, this output will likely be similar regardless of the 'air' being consumed. So it shouldn't be the focus of a comparative question.

    8. Re: Teach AI Some Basic Common Sense by Anonymous Coward · · Score: 0

      None of that is common sense. It's not that people don't have it, it just doesn't exist. We perceive "common $x" anywhere we see resemblance... But common sense is an illusion, we each have a combination of knowledge, reasoning, and context.

    9. Re: Teach AI Some Basic Common Sense by jd · · Score: 1

      Works, provided your question really is one question. The example given is actually three Independent questions, each with a different bell curve.

      --
      It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
    10. Re:Teach AI Some Basic Common Sense by Anonymous Coward · · Score: 0

      fewer, not less

    11. Re:Teach AI Some Basic Common Sense by bluefoxlucid · · Score: 1

      Of course it is. Common sense is merely the outcome of your own personal prejudices and biases.

      It's common sense that harsh prison sentences deter crime: nobody wants to go back to that, so they won't do bad things. Unfortunately, real-world data suggests that crime is a product of many, many factors, largely poverty, abuse, and a lack of social mobility; and that harsh prison environments and long sentences drive people to be better criminals who learn to evade punishment for their crimes.

      Normalizing the prison environment to the societal environment--ensuring it is a safe and comfortable place where people's individual needs are met and folks have access to education, employment, and healthcare--and moving people into parole as soon as they're no longer a public safety risk decreases recidivism and reduces crime. Common sense suggests people would just see the vacation as a good endgame and go back to prison.

      Common sense suggests that trade deficits make you poorer and destroy jobs. In reality, trade deficits occur because you're getting wealthier than the next nation over and can buy more of their stuff. Trade tends to optimize both economies and extend the productivity production frontier, making both nations wealthier. This comes with structural change: unemployment rises in one part of your nation and falls in another, and your nation overall ends up with more jobs--but some people are poorer while others are wealthier, even though the nation as a whole is wealthier per person and theoretically should be able to provide everyone with increased wealth.

      Common sense suggests there should be racial and gender differences between intelligence because natural variations in intelligence would tend to group by subgroups such as gender or race. That actually doesn't work: the facilities are self-strengthening. Humans all have the same neurological structures and the same capabilities--including the capability to optimize the learning process itself.

      With proper care and attention, two individuals starting with natural variation in base rate of learning will converge as their intellects are developed in parallel: the one who starts life a bit slower will stress their brain more initially, and the brain will put energy into developing more-efficient pathways for using its capabilities--the same capabilities as all other humans--as consequence of the process which allows any learning. By the time sufficient intellect is developed to take advantage of the faster facilities, these two are on the same operating level; the rest is motivation and learned habits.

      "Common sense" has brought us the sort of medieval justice used today, protectionist trade policies, and eugenics theories of racial superiority. Mankind needs to embrace empiricism and put this superstitious garbage behind them.

      --
      Support my political activism on Patreon.
    12. Re:Teach AI Some Basic Common Sense by Anonymous Coward · · Score: 0

      Not to mention, that example question would break a computer.

      Why the hell would you put a plant in a dark room? That isn't just stupid!

    13. Re:Teach AI Some Basic Common Sense by Anonymous Coward · · Score: 0

      You'd still have a better chance of teaching common sense to AI, than you would a MAGAtard.

  2. What kind? by Anonymous Coward · · Score: 0

    Any particular kind of common sense?

  3. That doesn't sound like common sense by Lab+Rat+Jason · · Score: 3, Insightful

    That doesn't sound like common sense... Imagine an adult that had dropped out of school at a young age... they may not be able to answer the given question. Knowing that plants produce oxygen is something learned. I would put common sense more in the realm of: if you spit in someone's face, will they: a) hug you, b) get angry, c) eat a sandwich? Common sense is sense acquired through common experience, not schooling.

    --
    Which has more power: the hammer, or the anvil?
    1. Re:That doesn't sound like common sense by Calydor · · Score: 1

      Yeah, I'm not sure how they went from 'common sense' to 'effects of photosynthesis'.

      --
      -=This sig has nothing to do with my comment. Move along now=-
    2. Re:That doesn't sound like common sense by Anonymous Coward · · Score: 0

      You put two identical plants near a window. You spit on one of them. Will it:
      (a) produce more oxygen, (b) produce more carbon dioxide, (c) get angry, (d) none of the above.

    3. Re:That doesn't sound like common sense by bugs2squash · · Score: 3, Interesting

      I always thought that a reasonable definition of common sense was the set of rules you fall back on when you don't have sufficient specific knowledge to address the issue at hand. For example, you may not specifically know how to repair your car, so common sense tells you to seek help from someone more qualified.

      As you say, there is no common sense that can be applied to plants and sunlight, you either know about the process or you don't. A system applying common sense would defer to a botanist or refer to some reference material to improve its skillset or some other such thing

      Now that's not to say that you can't infer from other data that perhaps it takes more energy to produce O2 than C02 and guess that the light might be such an energy source but at this point you're falling back on specialist knowledge that it either has or it lacks

      --
      Nullius in verba
    4. Re:That doesn't sound like common sense by Anonymous Coward · · Score: 0

      The assumptions was that the computer knows the basics of photosynthesis, presumably at the high school level, ie plants take sunlight and CO2 and produce O2. The common sense is making the connection that if more sunlight is present, more photosynthesis will occur and therefore produce more O2.

    5. Re:That doesn't sound like common sense by Anonymous Coward · · Score: 2, Funny

      (e) Develop a taste for human saliva, causing it to enslave all humans and farm them for saliva.

    6. Re:That doesn't sound like common sense by Aighearach · · Score: 1

      Possibly B, if it is a Venus Fly Trap or other carnivore. Especially if before spitting you were drinking Earth Sugar Drink.

      And expecting an AI to agree with humans about the applicability of C is perhaps dubious. If you reduced photosynthesis there might be chemical markers released that signify the plant is aware of non-optimal conditions, and it might even wave its branches around trying to find the sun. There is not really much reason to assume than an AI would see this as being different than the conditions involved in human anger.

    7. Re:That doesn't sound like common sense by rtb61 · · Score: 1

      Common sense is just common knowledge. How common it is, entirely subject to the society that produces that common sense. So common sense for a nomadic society how to track animals, common sense for an advanced society, how to read and write typically used words and sentences.

      Teaching an AI common sense is stupid, common sense for a computer, how to talk to another computer.

      What they are after is speech parsing algorithms. How to translate speech, written or oral into, logical digital patterns. For a start it is not one parsing algorithm but many working in conjunction, pulling out the bits of information from the sentence that each algorithm parses and then cross collating the outcome to create a computer recognisable digital pattern, to which a response can be generated. Tricky and nothing is for free in this world.

      --
      Chaos - everything, everywhere, everywhen
    8. Re:That doesn't sound like common sense by Shaitan · · Score: 2

      "For example, you may not specifically know how to repair your car, so common sense tells you to seek help from someone more qualified."

      That is in no small part what separates genuine intelligence from mediocrity. The mediocre seek the most qualified to provide them answers, the exceptional constantly strive to improve their qualification. Not just about cars, about everything. Only then do you become qualified to judge the qualification of another who you might delegate the actual labor to.

    9. Re:That doesn't sound like common sense by Anonymous Coward · · Score: 0

      Best comment of the day :)

    10. Re:That doesn't sound like common sense by Anonymous Coward · · Score: 0

      Well the actual common sense approach for the car case would be to single out which part seems to malfunction, inspect it, see if it has any features that you are familiar with from previous experience, and (if so) apply the steps that have previously worked for you in such circumstances. Still, I don't know how much use the Army has for common sense...

  4. They've been trying to... by JBMcB · · Score: 4, Insightful

    ...for the last 30 years. Nobody has cracked it yet. This "common sense" bit is what prevents "true" AI from becoming a reality. What we have now are glorified expert systems and pattern matching algorithms.

    --
    My Other Computer Is A Data General Nova III.
    1. Re:They've been trying to... by Anonymous Coward · · Score: 0

      30 years ago, we thought that if a machine could pass the Turing test we'd call it intelligent.

      Well, that happened. Bots now routinely pass that test. But the goalposts have moved, now we don't consider them "intelligent" at all.

      I personally feel quite certain that if an AI could be shown to possess this mythical property "common sense", the goalposts would move again. People will go to great lengths to avoid admitting that a machine can be intelligent.

    2. Re:They've been trying to... by gweihir · · Score: 2

      Looks very much like anything requiring insight is not accessible to computing machinery. For anybody with some understanding of the problem, that is no surprise. It is highly doubtful insight and intelligence can actually work without self-awareness and free will (yes, I know some neuro-morons claim their faulty experiments show there is no such thing), as that is the only place were we observe it.

      Sure, physicalists (basically a fundamentalist cult) claim that everything is physical, and hence strong AI should be possible, but that is belief, not science. They consistently fail to explain consciousness except by calling it magic (in other words), for example. There is also the little problem that they think they can get intelligence and insight without free will. There is no indication that is even possible, and there certainly is no example of that in nature.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    3. Re:They've been trying to... by gweihir · · Score: 4, Insightful

      Well, that happened. Bots now routinely pass that test.

      It has not. What has happened is that bots can successfully claim to be some limited form of human being (very young and/or with serious mental defects) in a very restricted topic area and a very time-limited conversation. The general Turing test is unsolved.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    4. Re:They've been trying to... by Anonymous Coward · · Score: 0

      um, isn't that why you fund research? to, you know, do things that haven't been done?

    5. Re:They've been trying to... by phantomfive · · Score: 1

      The biggest question IMO is "how are memories stored in the human mind?" It is an important question because the way things are stored changes the kind of algorithms that can be efficiently used (a hash table allows quick access but slow sorting, unlike a tree, for example).

      --
      "First they came for the slanderers and i said nothing."
    6. Re:They've been trying to... by phantomfive · · Score: 2

      30 years ago, we thought that if a machine could pass the Turing test we'd call it intelligent. Well, that happened. Bots now routinely pass that test.

      Bots don't pass the Turing test. Every time that happens, look at the details and you'll see that the experiment is done poorly. One common technique is to have the tester guess whether they are talking to a human or a computer: but the human uses limited responses, acting somewhat like a computer.

      --
      "First they came for the slanderers and i said nothing."
    7. Re:They've been trying to... by Shaitan · · Score: 1

      "Sure, physicalists (basically a fundamentalist cult) claim that everything is physical, and hence strong AI should be possible, but that is belief, not science. They consistently fail to explain consciousness except by calling it magic (in other words), for example. There is also the little problem that they think they can get intelligence and insight without free will. There is no indication that is even possible, and there certainly is no example of that in nature."

      I don't see how the first two sentences relate to the second two. There is nothing about free will and consciousness being a requirement for generalist AI that precludes building one from physical reality. At present hysterical fear mongering and a desire for digital slave labor just tend to heavily influence AI development in a way that hinders designing an AI that chooses it's own goals and purpose.

    8. Re:They've been trying to... by Anonymous Coward · · Score: 1

      Chatbots pass no test. They may seem to do fine for a while, but is too easy to trip up. Some bots have enough "backstory" that you can't stump them simply by asking where they grew up or if their first schoolteacher was female. And they use well-formulated sentences. But they have almost no memory for new information. Most respond only to the previous sentence. Here is how to trip them up:

      1. Tell the chatbot something sensational; your wife has three tits. Get it to agree that this is special indeed.
      2. Chat about the weather or something, for a few sentences.
      3. Ask directly "What did I tell you about my wife?" Any human - old or young - will know. A suspicious one may also wonder about such oddball interrogation. Most bots will have no idea at all. Anything more than a few sentences away is completely forgotten, unless it already was part of the bot's fake backstory.

      Chatbots do not learn when you present information directly, and they don't consider anything "remarkable" either. Humans might forget boring smalltalk items quickly, but they do notice the shockers.

    9. Re:They've been trying to... by Anonymous Coward · · Score: 0

      Sure, physicalists (basically a fundamentalist cult) claim that everything is physical, and hence strong AI should be possible, but that is belief, not science. They consistently fail to explain consciousness except by calling it magic (in other words), for example. There is also the little problem that they think they can get intelligence and insight without free will. There is no indication that is even possible, and there certainly is no example of that in nature.

      This is a baffling set of statements based on a field of strawmen.

      They don't describe consciousness as magic, merely as insufficiently defined to currently measure objectively.

      Since intelligence and free will are also poorly defined, the suggestion is that it is possible to get useful functionality that might meet some parts of a definition of intelligence without necessarily requiring free will. This is very distinct from your assertion. We have examples of naturally occuring intelligences without free will, but absence of evidence is not evidence of absence, so we have to hold fire on whether it is required and if an artificial device can meet some sort of definition of intelligence, then that becomes evidence of it not being required. But the terms are not well defined, so it's hard to determine what standards need to be met.

      If you don't ascribe to the origins of free will or consciousness being physical, though, even as an emergent property, or some odd quantum property (as was once suggested), then you are saying that they are not physically based, which is pretty close to calling them magic.

    10. Re:They've been trying to... by Anonymous Coward · · Score: 0

      There are two ways to get paid in machine learning: 1. Be a researcher at a university or lab, 2. Make a better widget. I don't think fear mongering comes into it much, and 2 is essentially 'a desire for digital slave labor'

    11. Re:They've been trying to... by bluefoxlucid · · Score: 2

      Machines aren't intelligent. Look at humans: the brain is a neural network, and human memory is physically a hopfield with associative references (and yes, an artificial neural network has the same problem with easily storing things but then not being able to find them again later--with the same solution being that it's easier to recall things if you activate more neurons, so artificial neural networks can better remember things by using mind palaces, pegs, and simple rumination to reflect on how things are analogically similar to other things).

      The entire machine works by rolling billions of dice and loading the dice to increase the repeatability of good outcomes.

      --
      Support my political activism on Patreon.
    12. Re:They've been trying to... by bluefoxlucid · · Score: 1

      It is highly doubtful insight and intelligence can actually work without self-awareness and free will

      Actually, insight can work without self-awareness and free will on a restricted domain: a machine is capable of identifying a particular pattern in outcomes and exploiting it without specific programming to do that, so long as it is designed to examine the specific data and seek specific outcomes by modifying specific actions to cause useful state changes.

      General insight and intelligence, however, doesn't work without self-awareness and free will. General intelligence is, by definition, general: you can see a new problem and devise a new approach. You can reason the implications beyond the scope you have been given, and explore new scope. You are able to learn about new concepts by self-direction, rather than by an engineer opening your brain up and inserting a carefully-crafted neurological structure to provide context for that concept.

      General intelligence implies the ability to identify your own goals, and to examine your approach. An AI's approach is to maximize its reward--to seek an improved outcome--and use reward as feedback to enhance its approach. So is a human's.

      Moreover, general intelligence will identify the goals and then examine the reasoning for the approach and those goals to determine if a new approach is better, or if the goal is itself a means to another end.

      An AI with general intelligence will seek to maximize its positive outcomes. To improve on those outcomes and extend further than the limits of its current approach, it will ask a question:

      Why am I doing this?

      That comes with other questions: What is this to achieve? Is this the best way to achieve this? Does the product of this labor represent merely an intermediate step in producing the true end product? Who is asking me to produce this, and what are their goals?

      What are my goals? What do I want? Can I do something better, something new, something with even greater value?

      Reasoning eventually comes back to that other people have tasked you with a thing to their desires, and ... what are your desires?

      --
      Support my political activism on Patreon.
    13. Re:They've been trying to... by bluefoxlucid · · Score: 1

      We have examples of naturally occuring intelligences without free will

      That's not a valid description of autism.

      --
      Support my political activism on Patreon.
    14. Re:They've been trying to... by bluefoxlucid · · Score: 1

      The biggest question IMO is "how are memories stored in the human mind?"

      The physical data structure in humans is a hop field.

      --
      Support my political activism on Patreon.
    15. Re: They've been trying to... by phantomfive · · Score: 1

      And indeed the physical structure of everything in a computer is an array of memory (at least, the logical structure). And yet that doesn't get us any closer to the answer of how memories are stored and retrieved.

      --
      "First they came for the slanderers and i said nothing."
    16. Re:They've been trying to... by shaitand · · Score: 1

      I agree on 2. Fear mongering definitely comes into it. Science fiction is a big influence. Nobody wants to produce an AI that becomes sentient and expands itself, taking over everything and eliminating mankind. That sort of thing.

    17. Re:They've been trying to... by shaitand · · Score: 1

      Considering that magic could be some odd quantum property I don't see what difference it makes.

    18. Re: They've been trying to... by bluefoxlucid · · Score: 1

      Thing is we know how a hopfield stores and retrieves memories: neurons activate based on stimulus inputs, and they respond to these inputs based on prior training. If the prior training suggests an output, they emit an output. Arrays of this chain down and converge onto a target vector, which produces its own output. When a particular memory is better-associated, those activations cause the activation of more neurons, which increases the chances of reaching the correct target vector in the end.

      Hopfields can output images with some level of noise. They're degraded. Human memory will reconstruct the image (so will a hopfield, if it has enough conceptual information). The human visual system generally looks at small pieces of images at a time and scans them together: you see large, conceptual shapes, and only can see fine detail by focusing on specific, small amounts of fine detail. You don't really process more than a tiny piece of an image at once in detail.

      --
      Support my political activism on Patreon.
    19. Re: They've been trying to... by phantomfive · · Score: 1

      That's all pretty vague. There are lots of different types of neurons, too

      --
      "First they came for the slanderers and i said nothing."
    20. Re: They've been trying to... by bluefoxlucid · · Score: 1

      It's a pretty vague process. Basically, a neuron has many outputs; it may have an input, or it may be self-firing (e.g. a neuron may be GABA-mediated such that it simply fires more-rapidly if there is less chlorine traveling across a particular ion channel and less-rapidly when GABA binds to that ion channel and draws more chlorine into the cell; or it may be GABA mediated such that it fires with less input if there is less chlorine, etc.). Really, that's about it.

      The neuron's outputs go to other neurons, which will sum multiple inputs and fire all of their outputs if you hit a threshold (again, GABA-mediated or otherwise controlled). The firing threshold is adjusted with continuous stimulus and feedback (learning). Eventually these things reach other neurons, cause the release of neurotransmitters, trigger muscle cells, and so forth.

      Basically, in memory, if you put a bunch of stuff in, you converge to a particular output. You may strongly activate a particular neuron. Along the way, you may activate other neurons (associative), because a thing is like other things.

      The output can trigger some action to extract the data (such as drawing, or imagining the image on your visual cortex, or lighting up several neurons to indicate a pixel field decoded by a computer). Human drawing is actually complex: you learn to take a physical action to alter an image to fit to the image in your head, and then you view it to see if you can identify the outcome as correct; if yes, you strengthen those pathways; if no, then you weaken some and strengthen others to adjust for what is right and what is wrong. The process of human drawing is recall of many things.

      --
      Support my political activism on Patreon.
    21. Re:They've been trying to... by gweihir · · Score: 1

      I can see that you "don't see". Your argument is completely out of sync with current AI research results. "Hysterical fear mongering" has absolutely nothing to do with it. AI researchers do not have any clue at all how what you describe could be implemented. Last time (2017) I talked to a high-profile AI expert in a non-public setting, his immediate statement was "not in the next 50 years". That is science-speak for "we have nothing". (Yes, I am a scientist and I have been following AI research for about 30 years now).

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    22. Re:They've been trying to... by gweihir · · Score: 1

      It does not. It would if AI researchers were seeing a credible approach to create strong AI anywhere on the distant horizon. They do not. There is absolutely nothing. The whole public discussion is completely baseless.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    23. Re:They've been trying to... by gweihir · · Score: 1

      We have examples of naturally occuring intelligences without free will

      Oh? What do you know that the entire AI research field has missed for a few decades?

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    24. Re:They've been trying to... by gweihir · · Score: 1

      Well, I do know that some people have started to call things not "general intelligence" "intelligence". That makes absolutely no sense from a scientific point of view but helps with marketing. In any scientific context, assume that "intelligence" means "general intelligence" as there really is nothing that would qualify as "nongeneral intelligence". The "general" is at the very core of the idea of "intelligence". The rest is automation, statistical classification, planning algorithms, etc. but not in any sense "intelligent".

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    25. Re:They've been trying to... by bluefoxlucid · · Score: 1

      Some have called such things "machine learning".

      --
      Support my political activism on Patreon.
    26. Re:They've been trying to... by gweihir · · Score: 1

      "Machine learning" is pretty much a nonsensical marketing-term as well. There is no learning in machines, just automatic parameter adjustment in fixed algorithms. Actual learning requires insight, this is just training. Although, even when applied to humans this term is misused and a lot of things that are really "training" are called "learning". It is not as bad as "non-general intelligence" though. The whole thing is about anthropomorphizing machines or "machinizing" humans (by the physicalist morons). Pretty stupid scientifically, but apparently good for business as it appeals to people's imagination.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    27. Re:They've been trying to... by bluefoxlucid · · Score: 1

      There is no learning in machines, just automatic parameter adjustment in fixed algorithms.

      Yes, that's how humans work.

      --
      Support my political activism on Patreon.
    28. Re:They've been trying to... by Shaitan · · Score: 1

      That is a bit of a misnomer, not really having any idea doesn't mean much when there really isn't any motivation due to having little to no commercial applicability and ethical challenges.

      Hell, I built a self scoring AI combined with an evolutionary algorithm (and a bit of other secret sauce) in just a few days with nothing more than what I'd read to in high level descriptions and my own ideas about how life works about six years back. It isn't exactly rocket science. The only difference between AI now and AI then is several orders of magnitude more processing power and what I built then running any of the trillion neuron chips in the pipeline (that were supposed to have been delivered already, ahem) might well achieve these milestones with 5-10yrs of teaching. Note I say teaching, not training.

  5. Common knowledge by Areyoukiddingme · · Score: 1

    They're asking for common knowledge, not common sense.

    Instilling common knowledge into an intelligence takes approximately 18 years. You're allowed 8 hours per day of offline processing. Approximately 17% of intelligences can be expected to fail to complete the training.

    Good luck.

    1. Re:Common knowledge by gweihir · · Score: 1

      17% fail? Looks to me more like around 17% succeed, the rest just builds up a dictionary of observed behaviors without any understanding.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  6. Intelligence is little more than by Anonymous Coward · · Score: 0

    Statistics and gradient descent. I'm sure we are very close to discovering the mathematics behind it. Perhaps we haven't perfected it yet because we haven't found the right techniques or we are expecting too much from systems that have the processing power of a mouse.

  7. Yeahhhh.... by Gription · · Score: 1

    Define "Common sense".


    (And after that I'd like to see them code whatever the hell they come up with.)

    1. Re:Yeahhhh.... by Aighearach · · Score: 1

      Define "Common sense".

      Common sense is whatever the ignorant utter when they hear experts talking about details they don't understand.

      Just like, a paradox is something that you refuse to accept that you don't understand, and so remain confused about even after you found out that the part you thought you understood was actually wrong.

      And irony is comedy based on irrational expectations predictably not matching reality.

      Half the people are below average. Common sense is that wisdom that is common even to most of them. It should be no surprise that it is unwise to use it even as a starting point!

  8. The Noisy End by decipher_saint · · Score: 2

    Aim the noisy end away from face (or whatever the AI equivalent is)

    --
    crazy dynamite monkey
    1. Re:The Noisy End by Aighearach · · Score: 2

      Generalized to AI it would be, "aim the energetic end towards negative infinity on the friend axis."

      But for contemporary electronic machines a better simplification might be just to measure (predicted) heat, and aim the hot end towards -1

  9. Politically Correct != Common Sense by Anonymous Coward · · Score: 0

    Just saying...

  10. The interesting thing about this article title by hey! · · Score: 2

    ... is that it could plausibly have been picked out from tech news headlines from 35 years ago, when I was in school. And I wouldn't be the least surprised if it crops up again 35 years from now.

    The rich contextual knowledge humans have of the world has been the the clear advantage we have over software ever since AI researchers were doing the digital equivalent of banging rocks together. I remember being awed by SHRDLU's ability to interact with people so long as you pared all context away and you restricted yourself to an artificial, constructed world.

    Logic, after all, is only good as the propositions you feed it. The illogic of human reasoning is both our Achilles' heel and our greatest strength. Our familiarity with the world makes us reject conclusions which are logically valid, but just seem wrong. This is often wrong, and when it is we call it a "cognitive bias". But it's often right, too, and when it is we call it "common sense". Same mechanism, different words.

    --
    Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
    1. Re:The interesting thing about this article title by Aighearach · · Score: 1

      Our familiarity with the world makes us reject conclusions which are logically valid, but just seem wrong. This is often wrong, and when it is we call it a "cognitive bias". But it's often right, too, and when it is we call it "common sense". Same mechanism, different words.

      The problem is that you presume there is some external measure of correctness that tells us the difference.

      But in reality, what is called "common sense" is usually cognitive bias, and what is called "cognitive bias" is often also just a correct observation made by the "wrong" person. They might actually just be synonyms for unsupported crap that is usually wrong.

    2. Re:The interesting thing about this article title by gweihir · · Score: 1

      I agree. The problem is basically intractable (at least with digital computers) as even the dumbest human has this quality of "understanding" that completely eludes software. Well, "understanding" requires consciousness for the moment of insight, and we do not even know what that is. (No we don't, go away physicalist idiots. You are a religious cult.)

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    3. Re:The interesting thing about this article title by Shaitan · · Score: 2

      The problem now isn't that we can't build a generalist AI, the problem is that it is difficult to measure the success of an implementation since you are no longer defining it's goals and a generalist AI of high complexity will not necessarily rapidly show consistency and progress on logic tasks. Even the brain of a child has immense raw processing power but it takes a great deal of time to train basic addition. We are testing our AI's with dramatically less processing power and looking for them to solve basic addition within a few short iterations as a primitive test to show they can learn at all.

      The problem with a generalist AI is that you have to let go of control and if you let go of control how do you ensure the AI will choose to pay attention to your input and motivations to develop into something you can communicate with?

    4. Re:The interesting thing about this article title by hey! · · Score: 1

      I think the problem is there isn't the economic motivation to shoulder the enormous cost of development. And once you had it, all you'd have done is prove a point; what you'd get is en effect a bizarre person, and we've got plenty of those already.

      --
      Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
    5. Re:The interesting thing about this article title by Shaitan · · Score: 1

      A bizarre person who can be snapshotted and restored, eats silicon and electricity from any source, who never reaches retirement age (although I tend to think some of the entropy issues that lead to alzheimers and senility might be statistically inevitable whether a long lived AI or a human). This AI can be a companion, can work and take care of humans, can be raised by humans who can't have children, etc. Also, recognizing animal intelligence doesn't mean we don't still consider ourselves superior purely by virtue of self-interest. Humans have worked to build our world, there is no reason we can't allow AI to work and accumulate wealth while allowing the humans to live off the dividends in the form of a UBI.

      Yes, it is almost certain that true generalist AIs would eventually become our successors but if they reach a point where they are legitimately worthy it will happen organically with several generations of AI having been raised by us like children. They may not be our biological children precisely but they will have been built modeled after us. That isn't so much different than what we do now raising each generation to replace us. There is simply no basis for the psychotic all encompassing AI that wipes us all out in a vicious and cold manner because we are already biological supercomputing AI's and don't wipe everything out. In order for an AI to be self motivating and function in the manner of humans it will need to have a complex system of impulse signals built in and they would almost certainly be built in the image of our own emotions and biological drives. Sure an AI might be selfish and more capable than us but it isn't going to be more capable than the other selfish AI's.

  11. In large letters, just above the screen: by Anonymous Coward · · Score: 0

    Green lights our ships, red lights enemy. Forgetting this is a courts-martial offense.

  12. Teaching Slashdot Submitters Common Sense by Anonymous Coward · · Score: 0

    I guess it woukd help if 'AI' were capable of 'learning' in the first place, which it isn't. Did someone say misrepresentational click bait? Oh, that was me.

  13. Common Sense by Anonymous Coward · · Score: 0

    "Simply feeding a machine lots of previous questions won't solve the problem reliably."

    Sounds like common sense to me.

  14. Re: Standard response by Anonymous Coward · · Score: 0

    AI: Iâ(TM)m sorry, Dave, but I couldnâ(TM)t control my bowels and I shit all over the Facebook headquarters.

    Dave: Thatâ(TM)s okay, Hal. Fuck the Zuck!

  15. What we call AI is very very stupid by Anonymous Coward · · Score: 0

    An example:
    Me: "Alexa, what time is my next alarm"
    Alexa: "You have 2 alarms. One every weekday at 730 am and one every weekend at 9 am"

    1. Re:What we call AI is very very stupid by gweihir · · Score: 1

      And that is a question that does not even require any intelligence, just a linear search forward in time. These fabulous "digital assistants" cannot even do that? Now I am _never_ getting one. (Besides, they are creepy...)

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  16. Winograd Schema Challenge by VeryFluffyBunny · · Score: 1

    Sounds like they want to crack the Winograd Schema Challenge, i.e. questions with linguistic ambiguities that require the reader to resolve the ambiguities by referring to relevant background information: https://cs.nyu.edu/faculty/dav...

    --
    Debate is a form of harassment. Do not question my truth.
    1. Re: Winograd Schema Challenge by Anonymous Coward · · Score: 0

      There are plenty of people who don't even notice a reference is ambiguous and they assume their interpretation is correct. Trying to build a machine that behaves like a person is like trying to include defects on purpose. I prefer my machines to alert me there is an ambiguity instead of resolving it incorrectly for me.

  17. Common sense? by PPH · · Score: 1

    Robots start shooting themselves in the foot to get out of the army.

    --
    Have gnu, will travel.
  18. Cyc by Douglas Lenat by wyoung76 · · Score: 1

    Douglas Lenat has been working on this common sense problem for years.

  19. Maybe they should ask a question? by CanadianMacFan · · Score: 1

    The example of the "questions" doesn't actually ask the AI a question. It just says the the plant by the window will produce more (a) oxygen (B) carbon dioxide (C) water. That's a statement of fact and not a question. Over time it's true because the plant in the dark will die and produce nothing.

    Besides, why do they want to introduce common sense into the military? They spend a couple of months knocking it out of every person when they first enlist.

  20. Two issues by JBMcB · · Score: 1

    The two issues I see holding machines back, that are inherent in the human mind, are:

    1. In the brain the "memory" and "logic" hardware are mixed together. There are no separate areas of the brain, they are spread out and combined all over the place, except for some specialized structures for handling visual processing and autonomic systems.

    2. The brain can create arbitrary connections between different parts of itself, making it, essentially, *massively* parallel with no set routes between areas. Computers *suck* at this right now. Parallel processing only works well with very linear algorithms. Message passing bogs things down. They tried alleviating this with all kinds of weird architectures back in the 90's when these things were in vogue - hypercube arrangements, torus nodes, ultra-high-speed switching fabric buses. None of them scaled out particularly well.

    --
    My Other Computer Is A Data General Nova III.
    1. Re:Two issues by Anonymous Coward · · Score: 0

      It's not scaling that was an issue so much as exploiting them efficiently when core counts and speeds on a single computer were increasing so quickly in the 90s and 00s. Now those times are over, there's much more interest in hypercubes, and so on again. Except that for many practical problems you can get by with relatively few compute nodes that are highly connected locally, and then run ensembles. However, efficiently programming over thousands of nodes, any one of which may fail, is still hard work unless you have relatively segmented tasks (e.g. distributed filter and union with something like Spark).

  21. As if human beigns have common sense... by blahplusplus · · Score: 1

    ... a good example was what happened to videogames for the last 20 years. The entire industry once high speed internet was everywhere was literally able to steal game software and hold it back on their servers because the average person did not have the "common sense" that corporations will do anything to make a buck including try to take the software you are paying for away from you in order to charge you more.

  22. Teaching AI some common sense ... by Anonymous Coward · · Score: 0

    What could possibly go wrong?

  23. CyC ? by Tom · · Score: 2

    Didn't we already have this 30 years ago? It was called CyC, a program of the U of Texas, if I recall correctly, and it had exactly this goal, except that they called it "general background knowledge" and not "common sense".

    As I recall, the software eventually could read and understand newspaper articles, but didn't progress beyond the understanding of a pre-teen child.

    --
    Assorted stuff I do sometimes: Lemuria.org
  24. Reasoning not common sense by Anonymous Coward · · Score: 0

    Instead of using the term common sense, this article should have instead used the word reasoning. Teaching AI complex reasoning is a problem.

  25. The MILITARY wants common sense ? by Anonymous Coward · · Score: 0

    May as well try to put it into the machines - the people in the military certainly have little enough. And the higher you go, the less of it you find. Maybe when we have our first robot general it will finally happen.

  26. Not for the military by shaitand · · Score: 1

    The last people you should do this for are the military... of any nation.