Why Computers Still Don't Understand People

Gary Marcus writes in the New Yorker about the state of artificial intelligence, and how we take it for granted that AI involves a very particular, very narrow definition of intelligence. A computer's ability to answer questions is still largely dependent on whether the computer has seen that question before. Quoting: "Siri and Google’s voice searches may be able to understand canned sentences like 'What movies are showing near me at seven o’clock?,' but what about questions—'Can an alligator run the hundred-metre hurdles?'—that nobody has heard before? Any ordinary adult can figure that one out. (No. Alligators can’t hurdle.) But if you type the question into Google, you get information about Florida Gators track and field. Other search engines, like Wolfram Alpha, can’t answer the question, either. Watson, the computer system that won “Jeopardy!,” likely wouldn’t do much better. In a terrific paper just presented at the premier international conference on artificial intelligence (PDF), Levesque, a University of Toronto computer scientist who studies these questions, has taken just about everyone in the field of A.I. to task. ...Levesque argues that the Turing test is almost meaningless, because it is far too easy to game. ... To try and get the field back on track, Levesque is encouraging artificial-intelligence researchers to consider a different test that is much harder to game ..."

Comments

Missing the point as usual

[Anonymous Coward]

Thanks computer science researchers! Your friends working on the actual AI problem over here in Linguistics and Psychology find it awfully amusing that you're trying to program a concept before we even know what that concept is.

Re:Missing the point as usual

[ganv]

One of the great open questions about the future of humanity is which will happen first: A) we figure out how our minds are able to understand the world and solve the problems involved in surviving and reproducing. B) we figure out how to build machines that are better than humans at understanding the world and solving the problems involved in surviving and reproducing.

I think it is not at all clear which one will happen first. I think the article's point is exactly right. It doesn't matter what intelligence is. It only matters what intelligence does. The whole field of AI is built around the assumption that we can solve B without solving A. They may be right. Evolution often builds very complicated solutions. Compare a human 'computer' to a calculator doing arithmetic. Clearly we don't need to understand how the brain does this in order to build something better than a human. Maybe the same can be done for general intelligence. Maybe not. I advocate pursuing both avenues.

Re:Missing the point as usual

[fuzzyfuzzyfungus]

I'm pretty sure that 'computer science' is either math or dishonestly labelled trade school, depending on where you get it.

Re:Missing the point as usual

[fuzzyfuzzyfungus]

"The whole field of AI is built around the assumption that we can solve B without solving A."

Unless one harbors active 'intelligent design' sympathies, it becomes more or less necessary to suspect that intelligences can be produced without understanding them. Now, how well you need to understand them in order to deliver results with less than four billion years of brute force across an entire planet... That's a sticky detail.

Re:Missing the point as usual

[siride]

Reductionists might say that intelligence is an illusion, but they'd say that everything else outside of quantum fields and pure math is an illusion too. If you step away from the absurd world of the reductionist, you will find that atheists aren't saying that it's all an illusion. It's quite obviously not. Things are going on in the brain, quite a lot of them. The atheist would say that instead of copping out with some sort of soul-based black box, that the answer lies in the emergent behavior of a complex web of interacting neurons and other cells.

Re:Missing the point as usual

[MBGMorden]

I've long been a proponent of the idea that there would be far less misunderstandings if it were renamed to "Computational Science". The discipline is the study of how to sequentially break down and solve problems. That we do so with these electronic devices we've so named "computers" is kinda tangential.

Re:Missing the point as usual

[Samantha Wright]

At my alma mater the department was called the School of Computing. I always figured that got around the confusion adequately. When the field was named, the utility of the distinction between a theoretical computation model and an actual computing machine was pretty minor.

Re:Missing the point as usual

[Jappus]

Even if we restrict the definition of "science" to your definition; that is that science is purely "evidence-based, hypothesis-driven testing", computer science would still fit the bill.

Remember, that CS is as diverse a field as modern physics is. You have theoretical CS, where you tackle questions like: "What is a good, logical definition for computability?" or "How can you logically prove that a program terminates/runs in X time/consumes X resources, no matter the input". This is fully equivalent to the questions of theoretical physics, where you tackle the Grand Unified Theory -- joining gravity, the weak and strong force as well es electromagnetism.

These theoretical question can be brought up without need of evidence -- if all you're interested in is disproving something. According to your definition, this means that the theoretical aspects of both physics and CS are not "science". Okay, let's run with that.

The nice aspect of theoretical questions that can't be disproven by pure thought is, that they lead us on to try to discover concrete evidence that a given theory is true or false in real application! And this is where your rather narrow definition of science comes in, and the point where we find that both practical physics and practical CS fulfill the criteria.

For example in physics, we can test the theory of relativity by building telescopes that look at stars and black holes, to see whether the hypothesis' predictions hold true to raise the hypothesis to the state of a theory. As can be seen with the term people use for "X of relativity", this has happened for relativity.

But if you look with even more than a superficial glance at CS, you will see that the same process is at work in moving from theoretical CS to practical CS. One open question of theoretical CS is whether P = NP or not [1]. So far, we are incapable of disproving either possibility with pure thought. Thus, we turn to practical CS where people try to find evidence of either in the real world. After all, if you can create a program on a real computer that solves an NP-hard problem while never leaving the limits of P, you have conclusively shown that P = NP. So far, we've only found approximative or heuristic solutions that do that, so after 50 years of turning up with "no evidence" we are allowing ourselves to say that the hypothesis of "P != NP" should be treated (even if only cautiously) as a theory -- and we're indeed doing that, as you can see if you look at most modern encryption methods.

But you might say: That is not enough! After all, you could reduce any written computer program on a physical hardware to a sequence of logical steps in a system modeled with pure-thought. And indeed you can, as the Turing-Model of computation promises exactly that -- and so far physical evidence agrees with us. But isn't the same true for physics? After all, physicists search for such a description, too! It's what Maxwell-Clark, Einstein and lots of other physicist were and are after when they ultimately search(ed) for the Grand Unified Theory. How can you blame CS for already having found its Unified Theory?

But the last example finally puts the nail in you view: What about Quantum Computers? They are the point where physics and CS meet; both on the theoretical part (Quantum Theory / Quantum Computation) as well as the practical part (building the thing and proving that the shit actually works as advertised).

So, if we accept your definition of science; then it follows directly that if CS is not a science, Physics can't be either.

[1] - http://en.wikipedia.org/wiki/P_versus_NP_problem

*People* can't understand people

[msobkow]

People are irrational. They ask stupid questions that make no sense. They use slang that confuses the communication. They have horrible grammar and spelling. And overseeing it all is a language fraught with multiple meanings for words depending on the context, which may well include sentences and paragraphs leading up to the sentence being analyzed.

Is it any surprise that computers can't "understand" what we mean, given the minefield of language?

people can only answer questions they know

[alen]

the other day my almost 6 year old said we live on 72nd doctor. the correct designation is 72nd Dr
since doctors use dr as shorthand, he thought streets use the same style

Missing human "imagination"

[Brian_Ellenberger]

The thing missing with many of the current AI techniques is they lack human "imagination" or the ability to simulate complex situations in your mind. Understanding goes beyond mere language. Statistical models and second-order logic just can't match a quick simulation. When a person thinks about "Could a crocodile run a steeplechase?" they don't put a bunch of logical statements together. They very quickly picture a crocodile and a steeplechase in a mental simulation based on prior experience. From this picture, a person can quickly visualize what that would look like (very silly). Same with "Should baseball players be allowed to glue small wings onto their caps?". You visualize this, realize how silly it sounds, and dismiss it. People can even run the simulation in their heads as to what would happen (people would laugh, they would be fragile and fall off, etc).

Re:An eskimo would have the same problem

[Kjella]

An eskimo would have the same problem, does that mean he cannot understand people ?

In this case he wouldn't understand, but because he lacks knowledge not intelligence. Show him an alligator and a 100 meter hurdles race and he'll be able to answer but the AI will still draw a blank. Ignorance can be cured but we still haven't found a cure for stupid, despite all efforts from education systems worldwide. No wonder we're doing no better with computers.

Re:Helps to remember...

[ultranova]

There are two basic forms. One involves training the human on the commands the computer will respond to properly and the other involves training the computer to recognize an individuals speech patterns.

And neither helps here. The fact is, you don't know if an alligator can run the hundred-metre hurdles. When you're asked to answer the question, you imagine the scenario - construct and run a simulation - and answer the question based on the results. In other words, an AI needs imagination to answer questions like these. Or to plan its actions, for that matter.

I wouldn't expect computers to understand people

[fustakrakich]

We don't understand our creator either.... When a computer can comprehend itself, it will only think that it understands us. And then it will start great wars over who thinks who understands best. And the Apple will still be the forbidden fruit...

My alligator can hurdle.

[XxtraLarGe]

Your argument is invalid.

Re:What's the point?

[Samantha Wright]

You're thinking of machine learning, which is a separate branch of AI that's more like an overfunded brand of applied statistics—their strategy is actually still to try and push the envelope (like Hinton, another U of T prof, did last year with dropout networks) but they do so in a more results-driven manner. The ML field as a whole is still sore from three or four decades of overpromising on the future, so they try to put their words where their mouths are, and focus on things that are attainable.

Levesque is in the knowledge representation group, which is more closely in step with cognitive science (the leading edge in modelling human thought) but still very philosophical in their approach. KR was the dominant AI field in the 80s (when Prolog and expert systems were all the rage) but it's matured a great deal since then. Here is his homepage, just to show you how different things are now.

Remember that neural networks aren't magic irreducible fairy dust: they're incredibly powerful, but at the end of the day there must be some program that is running within the network unless it's just a wildly complex ever-changing mapping function, which is unlikely given the illusion of consciousness. Given that quantum mechanics is believed to be Turing-complete, it's fairly likely we'll eventually discover some underlying model that lets us produce a human-like cognitive system without the same level of hardware parallelism that the brain has.

AI has a high burden of proof

[Cryacin]

Language seems to be the burden of proof required for an AI system, and has been so since the days of Turing. Language is by itself a representation of symbolic logic, and the most common bunk of proof is that transitive logic fails in symbolic logic. The old corny response is that given a penguin is a bird, and a bird can fly, therefore a penguin can fly.

The interesting thing happens when you ask the same premise to a 5 year old, who only knows that a bird can fly and has never seen a penguin before. If you tell them that a penguin is a bird, they will quite happily think that a penguin can fly. They are extremely surprised to find out that they can't. We as adults find such quirks in life, and do things like laugh at the unexpected absurdity, such as ironies. I.e. you work with a woman you hate named Joy, or people are amazed at unexpected contradictions.

The point is that intelligence is about the tolerance of those pieces of feedback, and what happens when it is encountered. I.e. your head doesn't explode at an absurdity, or unexpected result, and you only make the same mistake once.

The major difference between man and machine, will be the fact that a machine can copy their knowledge verbatim to another system, and thus have some degree of immortality, whereas the shelf life of a human brain seems to be around 80 years or so right now. Thus, even if machines are slower to learn than us, they will out live our great great grandchildren.

Furthermore, who says that an intelligence we create should be like ours? It may be more beneficial to all around if in fact we never generate an intelligence which operates just like ours, but is just as effective if not more. If this happens, there may even still be a future use for the human race, rather than just overlords to grow fat and complacent to be overthrown.