arras@icase.edu (Michael Arras) (03/05/91)
I'm looking for a buzzword that describes problems that at first sound easy to do, but when it actually comes down to writing an AI program, the problem turns out to be very difficult to solve. For example, a planning program to buy food at the store. Anybody can actually do it, but implementing a program that plans the tasks at a very detailed level can become quite a complex task, as I'm sure you know. A professor notioned at "cryptohard". Surely, there must be a better word then that. All suggestions will be carefully thought over. Thank you. Mike Arras arras@icase.edu arras@cs.wm.edu ICASE/NASA Langley College of William and Mary
csinger@cs.ubc.ca (Andrew Csinger) (03/05/91)
In article <1991Mar4.170121.9630@news.larc.nasa.gov> arras@icase.edu (Michael Arras) writes: | | I'm looking for a buzzword that describes problems that at first sound |easy to do, but when it actually comes down to writing an AI program, the |problem turns out to be very difficult to solve. Artificial Intelligence is the canonical buzzword. You need look no farther than that :-) The Phantom. |Mike Arras |arras@icase.edu arras@cs.wm.edu |ICASE/NASA Langley College of William and Mary -- I hate ean.
cam@aipna.ed.ac.uk (Chris Malcolm) (03/05/91)
In article <1991Mar4.170121.9630@news.larc.nasa.gov> arras@icase.edu (Michael Arras) writes: > I'm looking for a buzzword that describes problems that at first sound >easy to do, but when it actually comes down to writing an AI program, the >problem turns out to be very difficult to solve. Why do you need such a word? Aren't ALL problems like that? Name one problem that at first sounds easy, and when it comes down to writing an AI program, it turns out to be easy too! -- Chris Malcolm cam@uk.ac.ed.aipna +44 (0)31 667 1011 x2550 Department of Artificial Intelligence, Edinburgh University 5 Forrest Hill, Edinburgh, EH1 2QL, UK DoD #205
scotp@csc2.essex.ac.uk (Scott P D) (03/05/91)
In article <4110@aipna.ed.ac.uk> cam@aipna.ed.ac.uk (Chris Malcolm) writes: > >Why do you need such a word? Aren't ALL problems like that? Name one >problem that at first sounds easy, and when it comes down to writing an >AI program, it turns out to be easy too! I think it would usually take me more time to solve a cryptarithmetic puzzle than it would to write a Prolog program to solve the same puzzle. More generally, those problems that are easily defined formally and lend themselves to solution by something approaching exhaustive search are readily solved using a computer. Of course, most real problems fail to qualify on either count. I suspect the original poster was seeking some term to denote those problems that humans regard as trivial but which turn out to be far from simple when examined in detail. I think this is a reasonable request. If one were to rank a set of tasks according to (a) difficulty for a human, and (b) difficulty for a computer, I doubt that the correlation would be very high. Why this is so raises lots of interesting questions. Paul Scott, Dept Computer Science, University of Essex, Colchester, UK.
fawcett@unix1.cs.umass.edu (Tom Fawcett) (03/05/91)
arras@icase.edu (Michael Arras) writes: > > I'm looking for a buzzword that describes problems that at first sound >easy to do, but when it actually comes down to writing an AI program, the >problem turns out to be very difficult to solve. > How about AI-hard? ("at least as hard as any other problem in AI")
turner@webb.psych.ufl.edu (Carl Turner) (03/06/91)
In article <4110@aipna.ed.ac.uk> cam@aipna.ed.ac.uk (Chris Malcolm) writes: >In article <1991Mar4.170121.9630@news.larc.nasa.gov> arras@icase.edu (Michael Arras) writes: >> I'm looking for a buzzword that describes problems that at first sound >>easy to do, but when it actually comes down to writing an AI program, the >>problem turns out to be very difficult to solve. >Why do you need such a word? Aren't ALL problems like that? Name one >problem that at first sounds easy, and when it comes down to writing an >AI program, it turns out to be easy too! Of course, lots of A.I. problems are *not* "like that." Many of them look hard when you start. The question as posed by Michael seems to be a pretty good one. Naturally, there are no A.I. problems that look easy and solve easy, else they wouldn't be A.I. problems.... -- vxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvx v Carl Turner Psychology Department x v turner@webb.psych.ufl.edu University of Florida, Gainesville FL 23611 x vxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvxvx
clagett@ms.uky.edu (Bill Clagett) (03/07/91)
a long time ago, someone wrote: >I'm looking for a buzzword that describes problems that at first sound >easy to do, but when it actually comes down to writing an AI program, the >problem turns out to be very difficult to solve. gee. how about "deceptively simple"? -- clagett@ms.uky.edu clagett@ukma.bitnet ukma!clagett bill@ukpr.uky.edu
tesar@tigger.Colorado.EDU (Bruce Tesar) (03/07/91)
>>I'm looking for a buzzword that describes problems that at first sound >>easy to do, but when it actually comes down to writing an AI program, the >>problem turns out to be very difficult to solve. > ... lots of quick suggestions, head-scratching, etc. omitted ... > I hope people realize that this endeavor is in danger of becoming self-referential. :-) -- ========================== Bruce B. Tesar Computer Science Dept., University of Colorado at Boulder Internet: tesar@boulder.colorado.edu
russell@minster.york.ac.uk (03/07/91)
In article <4111@stl.stc.co.uk> "Michele Morris" <mem@stl.stc.co.uk> writes: >In the referenced article csinger@cs.ubc.ca (Andrew Csinger) writes: >>In article <1991Mar4.170121.9630@news.larc.nasa.gov> arras@icase.edu (Michael Arras) writes: >>| >>| I'm looking for a buzzword that describes problems that at first sound >>|easy to do, but when it actually comes down to writing an AI program, the >>|problem turns out to be very difficult to solve. >> > >I think the buzzword you may be looking for is "wicked problems". I can't >remember who coined the term, and I can't remember its exact meaning >ciao ... Michele Morris Yeaow! Wicked! I mean, mega. An excellent word, with splendiferous overtones. A special kind of phrase, for the truly magnificent. Excellent. I am certain it was first coined in "Bill and Ted's Excellent Adventure", though. :-) Russell. (I have got a .signature but I can't find it)
willdye@typhoon.unl.edu (03/07/91)
(in response to a request for a buzzword for 'easy' problems that turn out to be difficult) I've used the term "The Abyss" to refer to what some others call "The common sense barrier". Natural language translation, for example, looked easy in the 1950's, and AI researchers got a lot of money to work on it. They found out that solving the translation problem required having the computer 'understand' what the words meant. In other words, you can't solve this little problem until you have a computer that "thinks". I've found other problems like this, not necessarily involving the common sense barrier. Finding NP-Complete within a problem is an example. It's been shown that to solve ANY NP-Complete problem you would be able to solve ALL of them. It's as if we are working on our little problems, and then we discover that the solution lies in the Abyss. The funny thing is, it seems that there is only ONE abyss, i.e. to solve one is to solve them all. The problem is, we can plunge further into the Abyss, but we'll never hit bottom. There will always be irreducible problems. The Incompleteness theorem indicates that even mathematics r~? has areas that cannot be discerned. Quantum Physics may imply that there is not necessarily a single set of clear laws that determine what happens next. Who says the universe has to be comprehensible, or even simple? There's an ironic bright side to all this. Our own thought process seems to capitalize on this frustrating side of nature. Combinatorial explosion enables us to curse the complexity of combinatorial explosion. Well, I've degressed, as usual. Anyway, what do you think of the buzzword "The Abyss"? William L. Dye willdye@typhoon.unl.edu .
mack@intvax.UUCP (Michael J. McDonald) (03/11/91)
From article <1991Mar6.192044.8055@csn.org>, by tesar@tigger.Colorado.EDU (Bruce Tesar): >I'm looking for a buzzword that describes problems that at first sound >easy to do, but when it actually comes down to writing an AI program, the >problem turns out to be very difficult to solve. > How about non-CS AI for not a computer science AI problem. Or perhaps we could use the term real or real-world problem. I say this because a lot of AI problems are picked as easy simply because the programmer does not have a lot of experience in actually solving the problem in real life and therefore assumes that it will be easy. A buzword that said, in effect, that I didn't know much about the problem and therefore thought that it would be easy to solve might answer the request.
minsky@media-lab.MEDIA.MIT.EDU (Marvin Minsky) (03/11/91)
>From article <1991Mar6.192044.8055@csn.org>, by tesar@tigger.Colorado.EDU (Bruce Tesar): >>I'm looking for a buzzword that describes problems that at first sound >>easy to do, but when it actually comes down to writing an AI program, the >>problem turns out to be very difficult to solve. I've watched this thread for a while. In a certain futuristic sense, all problems so far solved by writing AI programs were very easy -- by historic standards. Thus the probllem of indefinite formal integration was unsolved for some 300 years until decisively solved by Moses-Engelman-Martin-Risch-Caviness, etc. But as I suggested in Society of Mind, the problem is that it is the "humanly" easy problems that turn out hard -- like learning to talk, or to see, or to do "simple, commonsense reasoning". These are the problems that use lots of brain. So, seriously, that buzzword could do more harm than good unless it has some sort of time-scale or dating parameter. The point is "difficult for whom" at what state of technology. Or, to specify to who the problem looks difficult.
gowj@novavax.UUCP (James Gow) (03/21/91)
In article <5491@media-lab.MEDIA.MIT.EDU> minsky@media-lab.media.mit.edu (Marvin Minsky) writes: >>From article <1991Mar6.192044.8055@csn.org>, by tesar@tigger.Colorado.EDU (Bruce Tesar): >>>I'm looking for a buzzword that describes problems that at first sound >"humanly" easy problems that turn out hard -- like learning to talk, >has some sort of time-scale or dating parameter. The point is >"difficult for whom" at what state of technology. Or, to specify to >who the problem looks difficult. I had to laugh when I read this. It sounded like the disarming of the armada. linc James