Slashdot Mirror
AI Going Nowhere?
jhigh writes "Marvin Minsky, co-founder of the MIT Artificial Intelligence Labratories is displeased with the progress in the development of autonomous intelligent machines. I found this quote more than a little amusing: '"The worst fad has been these stupid little robots," said Minsky. "Graduate students are wasting 3 years of their lives soldering and repairing robots, instead of making them smart. It's really shocking."'"
It's AI! That's what it does!
It can pick me out in a crowd, and it can show a number of emotions, such as surprise, anger, and boredom.... yawn.
"There are more pleasant things to do than beat up people." --Muhammad Ali
AI Founder Blasts Modern Research
By Mark Baard
Story location: http://www.wired.com/news/technology/0,1282,58714, 00.html
02:00 AM May. 13, 2003 PT
Will we ever make machines that are as smart as ourselves?
"AI has been brain-dead since the 1970s," said AI guru Marvin Minsky in a recent speech at Boston University. Minsky co-founded the MIT Artificial Intelligence Laboratory in 1959 with John McCarthy.
Such notions as "water is wet" and "fire is hot" have proved elusive quarry for AI researchers. Minsky accused researchers of giving up on the immense challenge of building a fully autonomous, thinking machine.
"The last 15 years have been a very exciting time for AI," said Stuart Russell, director of the Center for Intelligent Systems at the University of California at Berkeley, and co-author of an AI textbook, Artificial Intelligence: A Modern Approach.
Russell, who described Minsky's comments as "surprising and disappointing," said researchers who study learning, vision, robotics and reasoning have made tremendous progress.
AI systems today detect credit-card fraud by learning from earlier transactions. And computer engineers continue to refine speech recognition systems for PCs and face recognition systems for security applications.
"We're building systems that detect very subtle patterns in huge amounts of data," said Tom Mitchell, director of the Center for Automated Learning and Discovery at Carnegie Mellon University, and president of the American Association for Artificial Intelligence. "The question is, what is the best research strategy to get (us) from where we are today to an integrated, autonomous intelligent agent?"
Unfortunately, the strategies most popular among AI researchers in the 1980s have come to a dead end, Minsky said. So-called "expert systems," which emulated human expertise within tightly defined subject areas like law and medicine, could match users' queries to relevant diagnoses, papers and abstracts, yet they could not learn concepts that most children know by the time they are 3 years old.
"For each different kind of problem," said Minsky, "the construction of expert systems had to start all over again, because they didn't accumulate common-sense knowledge."
Only one researcher has committed himself to the colossal task of building a comprehensive common-sense reasoning system, according to Minsky. Douglas Lenat, through his Cyc project, has directed the line-by-line entry of more than 1 million rules into a commonsense knowledge base.
"Cyc knows that trees are usually outdoors, that once people die they stop buying things, and that glasses of liquid should be carried right-side up," reads a blurb on the Cyc website. Cyc can use its vast knowledge base to match natural language queries. A request for "pictures of strong, adventurous people" can connect with a relevant image such as a man climbing a cliff.
Even as he acknowledged some progress in AI research, Minsky lamented the state of the lab he founded more than 40 years ago.
"The worst fad has been these stupid little robots," said Minsky. "Graduate students are wasting 3 years of their lives soldering and repairing robots, instead of making them smart. It's really shocking."
"Marvin may have been leveling his criticism at me," said Rodney Brooks, director of the MIT Artificial Intelligence Lab, who acknowledged that much of the facility's research is robot-centered.
But Brooks, who invented the automatic vacuum cleaner Roomba, says some advancements in computer vision and other promising forms of machine intelligence are being driven by robotics. The MIT AI Lab, for example, is developing Cog.
Engineers hope the robot system can become self-aware as they teach it to sense its own physical actions and see a causal relationship. Cog may be able to "learn" how to do things.
Brooks pointed out that sensor technology has reached a point where it's more sophisticated and
The problem with AI, IMHO, is that computers, for the most part, are a just billions and billions of switches. You'll never have real, true intelligence because computers don't 'know' anything except on and off. You can try to simulate that, but so far simulation consists of what amounts to a gazillion 'if' tests, which is how any program works, really. All AI is is a larger, more complex set of 'if' tests than your average program.
My journal has hot
He has a complete disregard for the question of where the AI engine will run. If an AI is to be of any more use than a curiousity then those "little autonomous robots" must function in a viable manner so that the AI has something to do when it comes to "life".
I understand his frustration in general progress. But, those grad students are building a strong foundation for their later work that may very well meet the goals he is espousing. No need to have design flaws in implementation down the road because the engineer wasn't properly educated in physical design as well as logical design.
-Rusty
The Master (Angelo Rossitto) in Mad Max Beyond Thunderdome, "Not shit, energy!"
Yes, the man is quite brilliant, and possibly the most important voice in the field. That being said, he's also a self-important jerk. Intelligent Systems (what people in the field call AI) aren't where *he* thinks they should be, and he regularly complains about it.
Does it bothers you that AI is going nowhere?
Je t'aime Stéphanie
.. it was a really bad movie.
I don't see NON-ARTIFICIAL intelligence progressing a whole hell of a lot either...
It's not the AI which is going nowhere. It's the traditional approaches to AI such as Minsky's symbolic logic which are not going anywhere. Seach google for Henry Markram, Maass, Tsodyks. Their research seems very promising.
I have to agree that some AI (expert systems, etc.) has not progressed very far, but creating human like intelligence is not something that's going to happrn overnight. There have been tremendous leaps forward over the past few decades in things such as agents, however. Have patience.
I don't know if anyone has a google cache of aination.com, but I had a similar comment back in 2000 in the 'Works' section regarding the works of the MIT press which have recently proved as useful in developing true AI as these robots.
For REALLY good insight check out Nick Bostrom's articles on Super Intelligence here: http://www.nickbostrom.com/
Remembering that you are going to die is the best way I know to avoid the trap of thinking you have something to lose.
So hire and pay them money so they do real research instead of having fun. Otherwise quit your bitchin'. I personally think building stupid robots is cool.
-- Thou hast strayed far from the path of the Avatar.
I'd consider that pretty much intelligent, compared to some people I know. Then again, some people I know can hardly be described as sentient, let alone intelligent.
Hate me!
Humans measure intelligence by gauging replys to questions that have quantified answers. Giving an advanced computer a IQ test is sinply a matter of recalling the appropriate information from memory to answer. The true form of AI isnt about intelligence, its about reason. But how do you teach a computer to respond with an answer to a question that the computer has never encountered before... When we build a machine that can answer a question based on incomplete imput we will have made the first step in creating a machine that can "think"
reply to MYCROFTXXX@lunaauthority.com
Fire in the hands of the village idiot is no tool, but a weapon of mass destruction
How about we more thoroughly study and understand how human intelligence operates before we even presume to design something that imitates or rivals it in depth and complexity.
Well, what is a robot?
Courtesy of dictionary.com
robot
1) A mechanical device that sometimes resembles a human and is capable of performing a variety of often complex human tasks on command or by being programmed in advance.
2) A machine or device that operates automatically or by remote control.
3) A person who works mechanically without original thought, especially one who responds automatically to the commands of others
Number 1 is your R2D2 spaceman conventional robot crap.
Number 2 is your R.C. car, your microwave, dishwasher or toaster - all robots in the literal sense.
Number 3 pretty much describes slashdot readers.
I don't need no instructions to know how to rock!!!!
If you think that artificial intelligence going nowhere is a problem, what about natural human intelligence? There's clear evidence that it's going rapidly backward!
I wonder when they'll finally realize that you can't make a thinking machine. It doesn't have a a soul, a consiousness. It just follows some programming. At the most basic level, it's just a binary program. It follows whatever instructions it was given.
I honestly don't think we understand what makes a human consious or what makes someone be that person well enough to try to replicate it in software. You can make the logic more sophisticated, but I doubt we'll ever make them truly "think." And even if we did, how could we prove it? If you think about it, how can you prove anyone other than yourself is consious?
This quote ruined AI:
The wars of the future will not be fought on the battlefield or at sea. They will be fought in space, or possibly on top of a very tall mountain. In either case, most of the actual fighting will be done by small robots. And as you go forth today remember always your duty is clear: To build and maintain those robots. Thank you.
No kidding! Industry came to this conclusion 30 years ago. You can't make things "smart" if you don't know what smart is. Come up with a real useful definition of "intelligence". Apply simple engineering concepts to the problems instead of rushing to the "Statistical Death Spiral" where we generate reports with bad statistics to get paid form some research.
Come off it, stop making escuses. Robitics and AI are two entirely different branchs. AI is almost pure math and computing, and robotics is an engineering discipline. Why are AI researching building little robots?
Building any form of AI system is not easy, but copping out of it by building toys is not the answer. We already have platforms for AI; they're called line terminals. Things like pattern matching do not require a fully autonomous robot, after all.
Minsky is right; whats new to come out of actual AI research in the last 30 years?
Digital imagery and sound is just a bunch of yes/nos, but it can often be good enough for me :)
"As soon as we solve a problem," said Pollack, "instead of looking at the solution as AI, we come to view it as just another computer system."
;)
This is the most interesting comment to me. Because we understand the nature of the process that produces supposedly 'intelligent' results, (and we don't understand the same process in ourselves), we perhaps rightly just view the resulting system as just an application.
Seems like Minsky is throwing all his toys out of the pram because he doesn't want to admit to what everyone else has been saying for a while: that whether a computer can think is at best an astonishingly difficult question to answer and at worst meaningless. I'm a grad student who's just spent a year looking at computational linguistics and semantics (amongst other things), and the most debilitating restriction on the semantic side of things is the problem of so-called 'AI-completeness', which essentially says that if you solve this problem you have a, externally at least, thinking computer. Really simple things like anaphora resolution are AI-complete in the general case. If we could have solved this problem by now, I think it's fair to say we would have done, given its massive importance. However, we know that the brute-force case is ridiculously intractable, and we can't figure out how to do it any more cleverly. Roger Penrose argues that this is due to the fundemental Turing-style restrictions that we place on our notion of computing. Until we get a paradigm shift at that level, we're likely never to solve the general case.
And I'm sure that Minsky is aware that attempts to solve constrained domain inference and understanding have been taking place for a good long time now. I just don't see why he's so upset that the field of 'AI' (which is a nebulous catch-all term at best) has shifted its focus to things that we stand a cat in hell's chance of solving, and that have important theoretical and practical applications (viz. machine learning). Replicating human thought is not the be-all and end-all, and you can argue that it's not even that useful a problem.
Robots, though, I agree with. Can't stand the critters
Henry
i don't do sigs. oops.
I very much enjoy the works of Markram and Tsodyks. What they mainly analyze is how two nerve cells can interact with each other. They showed how they change their connection weights and how the timing of spikes, nerve impulses, affect how neurons connect to each other and how they transmit information.
While these studies tell us a lot about the underlying biology they do not tell us what these modes of information transmission are used for. For years it had been known that synapses have complex nonlinear properties. Biology pretty much does not constrain what functions neurons compute.
Thats why I do not believe that such studies will bring us nearer to real AI anytime soon. The algorithms coming from these systems are severely underconstrained. A lot of modelling has followed the pioneering works of Markram and Tsodyks, one of them being Maas. All these algorithms are very fascinating and might yield insight into the functioning of the nervous system.
The algorithms however are lightyears from being applicable to real world problems. The field of AI is old and in some sense quite mature. None of the "biologically inspired" algorithms today can compete with state of the art machine learning techniques.
Googlefight "Slashdot Troll" against "BSD is dying" 303:229. BSD thus cant die.
Although mentioned in a (lone) paragraph in the article, I don't know why we haven't heard more about the Cyc project. Lenat's premise that you can't have intelligence without knowing the millions of "obvious" things about the world, aka, "common sense" seems intuitively right.
If you reply, do so only to what I explicitly wrote. If I didn't write it, don't assume or infer it.
That's all there is to it. Minski is just sore that his theories from 30 years ago aren't proving themselves, and the decentralized models being implemented by his rivals at MIT (e.g. Rod Brooks and his graduate students) are demonstrating remarkably sophisticated behaviors and advancing the state of the art.
Maybe it's time to start encouraging open source projects and development in this field. Seniors could network with students from other universities to make more comprehensive and meaningful final projects.
Using open source development, a project to establish a tool kit for AI programming fundamentals could be born. It'd definitely be cool to have something like that available. I'm not sure if MIT has anything like this going yet, but they could easily whip up the brain power to get it started (and started right).
Ask some 3-year-old kids which way is up, and they will all know, but they won't be able to define it. Yet, since computers don't have bodies, they don't have anything like the semicircular canals that we have, which act as gyroscopes and give us an intuitive, non-intellectual sense of which way is up.
Trying to program intelligence purely with software puts the researcher at a disadvantage, since even the most fundamental rules and attributes of things (fire is hot, water is wet) have to be explicitly entered as constant variables.
Once robotics advances to the point where mobility, vision, and speech recognition can be taken for granted, then AI can be programmed as an add-on.
Body first, mind second - That's how animals evolved on this planet, and it's how, I believe, Rodney Brooks approaches this field.
every stain tells a story
We do not understand how to control (or if it is even ethical to control) the billions of automonous intelligent agents roaming this planet... so why should creating a whole new class of intelligent automata be a priority.
AI today has nothing to do with intelligence. Its all basically rule-based procedural programming. While this allows us to make some really neat applications like automatic vacuum cleaners and pool scrubbers, it has nothing to do with "intelligence".
The human mind is not rule-based -- we impose a framework of rules to allow everyone to live together in relative harmony. The core of our being -- how our mind actually works -- remains an absolute mystery.
Conformity is the jailer of freedom and enemy of growth. -JFK
"Graduate students are wasting 3 years of their lives soldering and repairing robots, instead of making them smart. It's really shocking."
Yeah, much more shocking than the -- decades -- he (and others in the 'hard AI' camp) have been spending? They've made oh-so-much more progress, haven't they?
Rodney Brookes made more progress with his robots in the early nineties than the whole hard AI camp did in 3 decades. I remember seeing a documentary once comparing this huge robot which used a traditional procedural program to navigate through a boulder-strewn field. It took about 3 HOURS to decide where to put its foot next. Meanwhile, little subsumption architecture-based robots were crawling around like ants, in real-time. (Oh, and some of them had to learn to walk from first principles every time they were turned on -- only took about half an hour!) That's the most damning evidence of the failure of hard AI I can think of.
As others here have said, what good is a brain until we get a useful BODY working? Manueverability and acute senses are a must before an artificial intelligence can do anything useful, or learn from its environment effectively.
ERROR 144 - REBOOT ?
Paraphrased, but the spirit is there: "And when a robot actually does succeed and walk down the hall and through the door, or whatever it's supposed to do, you've learned absolutely nothing because it may not do it again the next time. This is why mechanical engineers love their videos so much. With video they can say 'See, it really worked once!'"
www.HearMySoulSpeak.com
It drives me crazy that people are so concerned about possible technologies, that they want to "slow down and think about the consequences of xxx".
This is really just unfounded fear. While we still don't know if something is possible, is not the time to worry about what problems we can concieve that it will bring. Knowledge is more important than worrying about some issues that may or may not arise if we are able to do something. It is good to ask "If we cause this atom to split, will it kill us?", but I do not think there is any value in saying "Maybe we shouldn't find out what happens if we split this atom, because if it causes an explosion, someone might use that knowledge to build a bomb..."
One of my favorite quotes is from Isaac Asimov:
I'm sure a lot of people will disagree, but to me, knowledge is most important.
If Marvin wants to know why AI hasn't made major strides in the past 30 years (and, by the way, I would say that it has) he should look no further than his own bullying, arrogant approach.
Promising subfields like perceptrons were intentionally quashed by him... he went out of his way to strangle investment and research in areas he considered to be a dead end. We're not literature majors: we can't just all say the same thing in a party over wine and cheese and call it progress.
Even bad ideas, when well explored, can give new meaning and better approaches to a field. And since this is research, noone knows the correct answer: even a dumb-seeming idea may turn out to be the right one-- or give us clues about features the right answer needs to have.
Of course we've had major advances in AI. One of the challenges of AI, as the article points out, is that once something is well understood, it is defined as being outside the AI field. Computer vision, face recognition, voice and speech recognition. Conversation engines like SmarterChild. No, this isn't HAL, but they are good, positive steps in the right direction.
All of the spiking networks I've seen were nothing more than state machines that depend on numeric comparisons.
But I'm not an expert, and that's just my personal opinion.
What's this Submit thingy do?
I took Minsky's class last year, and let me tell you, the article couldn't print 75% of the irate stuff he has to say about AI, MIT, and life in general. We once spent an hour class session listening to Misky rant about modern science fiction and random things he didn't like about his Powerbook. In fact, most of his classes were extended rants about something or other (you zealots will be happy to know that he too, hates the Microsoft).
He comes across as affable but bitter. I found it strange that though he cointually complains about the leadership of the AI lab, he and his protege Winston were in control of it for some ~30 years without making any groundbreaking progress. In fact, Minsky's latest work "The Emotion Engine" is simply a retread of his decades-old "Society of Mind." I suspect that now that Brooks and the new guard are moving in, the old guard is looking for someone blame its lack of results on.
I personally built and programmed one of these "stupid little robots"; it's a wheelchair programmed to navigate in an office environment, using vision to determine where in the office it is. Nobody asserts that it can "reason". It navigates using a collection of local effects, in much the same manner that simple creatures operate. Watch the film "Baraka" for some rather amusing examples. At one point the film shows a bunch of caterpillars, each following the scent trail of the next --- unfortunately someone flipped the first one around, so it follows the last, and the whole colony just moves around and around until they die of starvation.
I think you would be surprised how easily remarkably complex behaviours can be achieved by a collection of very simple responses. Try fiddling around with Rossum's Playhouse, and read Brooks' book Cambrian Intelligence.
Please mod the parent up.
:-(
I am an AI researcher and the parent poster is speaking truthfully.
The main challenges in AI at the moment do not concern building the physical robots -- e.g. a piece of kit on wheels with IR sensors or such things.
The main challenges in AI concern applying some very complicated math to solve problems like pattern recognition, density estimation and other forms of machine learning.
It seems to me that a large number of AI PhD students spend their lives tinkering with the mechanics and electronics of the robots that will ultimately be used to test their algorithms. This is wasted time; a good electronics graduate should be able to do the tinkering, it shouldn't require a prospective AI PhD student to do it.
I can see the point in the PhD student learning a little about the hardware that they want to run their algorithms on (so that they know the limitations and common problems with real hardware), but they should not spend all their time doing that and wasting the opportunity to spend their time contributing to their field (i.e. AI, not mechanics or electronics).
That said, many AI labs do not have the funding to be able to pay full time hardware technicians, so in many cases the PhD student *has* to do the tinkering
"The noble art of losing face will one day save the human race"---Hans Blix
The problem with AU lays in poor biology. As long as it is based on pre-cybernetic (i.e. traditional, neodarwinian) biology, AI will never go anywhere. The only known intelligent systems are biological systems. To create AI, you need to imitate biology, you need to reverse-engineer what it is exactly that make biosystems special. But traditional biology has totally misled computer science. Pre-cybernetic biology, the biology you find in most books and the one taught in almost any classroom, cannot even define life. This pseudo-biology is the `biology' of the non-living, and as such, of the non-intelligent.
To create AI, you need to understand natural intelligence (NI) and for this you need to understand life. What is life? Cybernetic biology defines life as molecular autopoiesis. Which is interesting, since this definition of life is based on computation. Autopoiesis is the key here. The self-re-computation of a system is the key to life, and the key to intelligence, because you need a self to be intelligent. With an artificial self, we could have AI, and probably self-awareness. But good biology is the key.
? Unfortunately, it's not going to happen anytime soon. Biology is totally stagnant, and the Neodarwinian Cabal precludes any progress and silences any dissent (sort of a M$ of the science market). `Official' bilogical sciences just won't deal with life. And that's not going to change for a while, I'm afraid, no matter how hard sone of us try.
``L'imagination au povoir.''
You can't get that much out of most humans!
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Here's some perspective from an MIT AI lab grad student who's been inspired by both Minsky and Brooks. (Minsky is on my Ph.D. committee.)
"AI has been brain-dead since the 1970s."
I agree, unfortunately. At least, what was traditionally meant by "AI" has been brain-dead. There is very little focus in the field today on human-like intelligence per se. There is a lot of great work being done that has immediate, practcal uses. But whether much of it is helping us toward the original long-term goal is more questionable. Most researchers long ago simply decided that "real AI" was too hard, and started doing work they could get funded. I would say that "AI" has been effectively redefined over the past 20 years.
"The worst fad has been these stupid little robots."
Minsky's attitude towards the direction the MIT AI lab has taken (Rod Brooks's robots) is well-known. And I agree that spending years soldering robots together can certainly take time away from AI research. But personally, I find a lot of great ideas in Rod's work, and I've used these ideas as well as Marvin's in my own work. Most importantly, unlike most of the rest of the AI world, Rod *is*, in the long run, shooting toward human-level AI.
Curiously, just last month I gave a talk at MIT, tited "Putting Minsky and Brooks Together". (Rod attended, but unfortunately Marvin couldn't make it.) The talk slides are at
http://www.swiss.ai.mit.edu/~bob/dangerous.pdf.
In particular, I shoot down some common misperceptions about Minsky, including that he is focused solely on logical, symbolic AI. Anyone who has read "The Society of Mind" will realize what great strides Minsky-style AI has made since the early days. I also show what seem like some surprising connections to Brooks's work.
- Bob Hearn
Think of it. Only a stupid robot is a good robot.
Before you yell flamebait or troll, let me explain.
I have been following the progress of various AI technologies, including neural nets and adaptive logic networks, for many many years now. Years ago perceptions were first developed and it was shown that they could learn simple patterns. Perceptrons were basically two layers of software simulated neurons. They worked, and researchers were fascinated and worked on them regularly.
Minsky, being the "highly regarded" and "leader" in AI, wrote a paper that proved that these perceptrons could never learn more complicated patterns, and threw a bunch of math at the reader. So people stopped. After all, there was a mathematical proof that perceptrons weren't going anywhere. Research skidded to a halt for decades because of Minsky.
Of course, then someone developed the (gasp!) THREE layer perceptron/neural net and sure enough with the right formula it could learn much more complicated tasks.
Minsky, in my opinion, does this regularly. The problem is, that he has a reputation in the industry as being a leader (I'm not sure why).
He's already lost us two or three decades of research because of his "leadership" -- I am terrified that he might cost us more development into the future.
Where could we have been if Minsky wasn't always going around half cocked, screaming that he is right? "Robots are useless!" is history repeating itself and him trying to get more press. Keep developing guys, just ignore the peanut gallery. There's always someone who says it can't work (ahem, Minsky) -- it can and it will.
The baby's fine -- please stop sending business cards.
Universities are doing just that in the various RoboCup events.
Rodney Brooks (head of MIT's AI lab) has to say about this since he is the genesis of the "cheap, fast, and out of control" school of AI that Minsky is deriding here. But since rantage like this from Minsky isn't new, I bet Brooks takes it in stride.
About 20 years ago AI was going down the craperroo until folks like Brooks decided that the AI field would be better served by moving it from the more theoretical GOFAI method to a more applicable style. Revitalized everything.
What is music when you despise all sound?
People who do perform illusions and escape tricks have been doing things mostly the same way for decades. Magic tricks may change slightly, but all the basic principles and tricks are the same. There's no real evolution, just adaptation to please the crowd.
Now, with that in mind, let's look at artificial intelligence. AI has always been about trying to convince an audience that a machine is thinking. This is demonstrated by the very existence of the Turring test and many products (such as the Aibo, Furby, etc) that try to mimick emotions. If the audence is entertained, amused, or convinced, the AI is considered good. Bad AI is when the audience can see right through it.
Artificial intelligence is magic. It's a trick. It's an illusion.
It is no surprise then that AI hasn't really advanced. The trades of showmen are practically unchanged for hundreds of years. Razzle-dazzling an audience involves technological advances, but it remains unchanged. Even in the cases where "artificial intelligence" is used to aid in medical diagnosis ("expert systems") or manufacturing are really only following man-made logical structures. The computers aren't thinking, they're only doing what they're told to do, even if indirectly. The end result is impressed people who think the machine is smart.
Of course, you don't have to take my word for it. If you want to see how badly AI is going nowhere, I hightly recommend reading The Cult of Information by Theodore Roszak. While his focus is not on the fallicy of AI, it covers it in context with the much broader disillusionment of computers by society.
Now, what does AI need in order to progress? Probably AI creating other AI. Something with a deeper embodiment of evolution. As long as it's man-made, it will never be intelligent, just following a routine. Of course, I am going to stop right here... I am not qualified to offer a solution these obstacles.
Join Tor today!
"How the hell am I ever going to be able to download my brain into these damn little robots if they don't hurry up and make them smarter? I running out of time, dammit!!!"
I hear you, brother. I gave up a week ago. Right now I think I'm gonna take up shooting all the stupid people who are pissing me off. It's like, "hey, that dude just gave up smoking, let's bug him until he snaps"
A good argument can be made that a polecat (wild ferret) is more intelligent than many humans. For example, the polecat can survive outdoors with no assistance. The polecat can eat, sleep, have babies, and be more or less comfortable.
Where does human intelligence come in then? Human intelligence is learned. Of course a polecat at 4 months is more capable of surviving than a human at 4 months. Does this make the polecat more intelligent? But let's try and remember that the polecat is done developing, while the human has about 20 more years until full maturity.
So the human learns then. Plainly, the human learns more than the polecat over the course of 4 years than the polecat. So is the human more intelligent? I think we can unequivically say yes.
But what is it that makes human intelligence, and how is it different from a polecats? The answer is learning. But how does learning work?
Learning is a specific thing. People learn by rote. (Don't let someone tell you otherwise.) It is mimicry that teaches morals. Logic teaches ethics, but logic is learned like morals. This means that, basically, we learn everything.
The point is, if you think there is any difference between you and a polecat, I would like to point out that there is less difference between you and Alicebot.
If you want proof, look at how musicans or epic lyrists work. They learn specific phrases and use them over and over. Listen to your own speech or read your own writing. You'll find that you use plug in words and phrases. They'll be similar to your friends and parents, btw.
Hoist Number One and Number Six.
People justify their robophobia by pointing to these fictional examples, but if recent murder statistics are to be believed, the score is a bit lopsided.
This kind of prejudicial attitude must end.
> Cyc knows that trees are usually outdoors, that once people die they stop buying things...
Hey, that thing is already smarter than the companies that continue to send junk mail to my grandfather who has been dead for 22 years, now. Maybe they should get that software to manage their mailing list?
Just be sure to wear the gold uniform when you beam down -- you know what happens when you wear the red one.
Human Level AI's Killer Application - Interactive Computer Games, John E Laird and Michael van Lent American Association for Artificial Intelligence AI Magazine Summer 2001 pp 15-25
My summary of the above - the AI in games might not be too hot (some would dispute with the academics about that but let it go), but game environments themselves are complex enough to pose a challenge for state-of-the-art AI researchers.
AI seems to be going nowhere because the complexity of the problem hasnt been thoroughly discussed yet. Computers are designed to be strictly deterministic(otherwise it would be impossible to use them, look at wind*ws).
When we try to emulate a system with an other system that is different in nature a lot of capacity is wasted.
That said genetic programming is one of the fields where we actually see truly intelligent solutions to problems completely generated by computers. Problem is the algorithms need computational power beyond our wildest dreams to even be comparable to single cell organisms in ingeniousness.
After all the nature has had 50 gazillion years to evolve.
(Bits On Our Mind, an exhibition of some undergrad and graduate computer science work) ...and I headed STRAIGHT for the nematode booth. You see, I had heard that some clever Cornellian had created a simulation of the entire neural network of a nematode. The way I saw it, there was nothing else there that could possibly be more interesting than that.
:P
So I found myself standing in front of a computer screen. It was a worm swimming through water! In 3D! In real time! After I pushed my jaw shut, I began to ask the genius student some questions...
"Is that real-time?" "Well, actually, no, that is a 10 second looping clip that took a week to calculate."
"Well, I see a neural map there. Is that complete?" "Well, actually, no, that is a simplified version of the real nematode nervous system, on the order of about 1 simulated neuron to 10 actual neurons."
"So you simulate neurons! That's awesome. Let's see the code." (He proceeds to flip through 4-5 pages of very sophisticated-looking mathematical equations to describe the behavior of ONE neuron.)
What a let-down! No wonder Minsky is pissed, real AI is HARD!
We will soon have hardware that has the number of connections or processing power of a human brain. The problem is nobody's come up with the software to run on it. In humans this is what makes the brain more than big organ ... the "soul" if you are religiously inclined. Maybe a human soul can be reduced to nothing more than a program with an enourmous propesity to learn and adapt over years of training / habituation ... say from the years 0 to 18.
...starts by modeling the neurons of the brain dircetly as cells (implying a thorough understanding of the proteomics involved) instead of as a neural-net or some other high-level abstraction, perhaps the results will be more interesting.
Such a model is years off, though, AFAIK.
Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
There are many authors that have written and demonstrated that the brain probably doesn't function as a mass of context-free predicate logic rules -- including my favorite, Hubert Dreyfus.
Dreyfus argument is old, and its rebuttals are well-known. Consider that symbolic systems are not limited to context-free predicate logic.
The progress of AI is uncertain, but it is certain that there's no future for symbolic logic AI.
It is not certain for me.
Both connectionist and symbolic approaches may succeed if given enough time. However, I think that obsession with neural nets of many people here is of the same nature that obsession of numerous early aviation enthusiasts with wind-flipping devices. Certainly you can mimic mechanics of nature with some effort, but there are usually better ways to do the job.
Lisp is the Tengwar of programming languages.
So Minsky has an opinion, he expresses it, he provides data, calculations and evidence to support it.
People just accept it, and progress is delayed.
Why is it his fault that there are so many followers? If anything is to be blamed is that these researchers just blindly follow whatever he's saying rather then take a good critical look at what is going on.
If his math and theory "proved" that an area of AI was a dead end, and it wasn't, his math/theory was wrong. It is a sad state when nobody dare challenge the status quo.
I suspect part of Minsky's problem is frustration that his ideas about AI aren't bearing fruit, so he's going to take it out on other people's different approaches. It's not much different from the Perceptrons paper he co-wrote in the late 60's that nearly killed Neural Network research for most of the next decade. Never mind that there was plenty of useful neural network research to be done that avoided the failings of the perceptron model.
In my opinion, if we had the wholistic understanding of intelligence that would let us use Minsky's type of approach to AI, there wouldn't be anything left to do but implementation! One cannot just a priori assume all principles of intelligence by self-examination, and that's where he fails. There are interesting things to learn in that approach, but a large number of them have already been learned, so people are turning to other means (bottom up approaches focussing on self organization are doing well and leading to new discoveries) to get a broader understanding of what is involved in intelligence.
Just because Minsky has sour grapes doesn't mean that the robot people aren't doing useful research.
7 November 2006: The day Americans realized corruption and incompetence weren't addressing 11 September 2001
What Minsky is actually flaming about here is the damage done to the field by the original "AI winter" and the real possibility of a new AI winter starting in a few years.
;-)
.sig below).
"AI winter" is the name given to the collapse of strong AI as a business model in the mid 80's - expert systems and symbolic AI in general didn't deliver on their promises, and so the money went away. As a guy who got his doctorate in AI in 1985, I can tell you all about it.
One of the major causes of AI winter was researcher hubris - lots of people hacked up systems that appeared to solve 80 percent of certain complex problems and then said "all that stands between us and a complete solution is money and time". For many of those systems, solving the last 20 percent would have taken 2000 percent of the time, if it could have been done at all. The tragedy of AI winter, though, is that basically all of symbolic AI was abandoned, though some of it is creeping back out into the light with obfuscated syntax (see my
What Minsky sees here is a lot of people heading down the same path, but with neural nets and small robots instead of expert systems. The new systems are doing some interesting things relative to the old symbolic AI systems (though they do have the advantage of 20 years of Moore's law to help them). But, will they scale up? Right now, nobody knows. If they don't, the last thing the field needs is another cycle of overpromise/underdeliver/abandon.
Maybe AI is just plain hard, and cracking it will take longer than one or two computer industry business cycles.
To a Lisp hacker, XML is S-expressions in drag.
"Al Going Nowhere?"
well duh!
what idiot made the lowercase L and uppercase I look the same?
but, since we're on the subject, did you know Al Gore invented the field of AI?
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it
Actually, they have made significant strides already in figuring out how the brain works. Check out the Levy Lab at the University of Virginia. They've trained a computer model of a rat's hippocampus to do all sorts of intelligent things, such as transitive inference, sequence completion/combination/disambiguatuion, goal finding, etc. While these are not difficult problems for humans to solve or hardcode into a program, the fact that a single network can do these different and sometimes contradictory things represents something that I would call intelligence. As far as I know, they don't plan on having a model of a human brain very soon, since U.Va. lacks NSA-scale compute servers, but even rat-level learning is pretty cool.
WARNING: there is a trojan on your
Minksy belongs to the old school of AI thinking. These guys believe that it is possible to make statements about intelligence itself, without considering the interactions of the organism/agent with its environment or the underlying architecture of the brain/CPU. I think that the total failure of this style of thinking to produce anything interesting in 50 years proves that this approach is sterile. Minsky laments the fact that graduate students build robots, but this activity exposes students to the challenges of constructing a device that must actually interact with the environment. It is ridiculous to assume that you could design a system capable of intelligent behavior without ever confronting the problems of sensors and actuators. Almost every part of the brain is devoted to processing raw sensory input or generating motor output. One cannot simply design an intelligent system without worrying about sensory input and behavioral output. The CYC project of Lenat has the laudable goal of teaching a machine "common sense" by hard coding a vast database of simple statements like "Trees cannot walk". This is a totally wrongheaded approach to learning and reasoning, and is typical of old school, hard AI.
We will only make progress in engineering intelligent, adaptive systems by studying actual examples of intelligent, adaptive systems, namely animals. Neuroscientists and psychologists are beginning to embrace the tools of mathematical modeling and simulation wo help explain nervous system structure and function. Computer scientists would do well to similarly embrace the work of experimental neuroscience.
Minsky is a dinosaur.
Seems like AI has progressed about as far as it can go inside the box, only machines that can proactively interact with an environment as people do will learn to think like people.
Imagine what kind of thing you would be without vision, touch, smell, hearing or the ability to move and change your environment. Without these forms of interaction where would human intelligence be?
Seems that a Budhist philosphical approach is most helpful here, ie we are our parts, not more and not less. We are what we are. If you wish to create something that is like a human, you should take an inventory of our parts figure out how they fit together and try to find analogous electronics, software and hardware.
Which is precisely what a lot of the robot folks have started doing. Except that most have started a bit smaller and have modeled insects instead. Finding that they can model seemingly complex insect behavior with simple algorithms and machines.
Although, perhaps the next best step isn't building real robots at all, which can be expensive, error prone and time consuming, but building virtual robots that can be placed in virtual environments of our invention, somewhat like a "Matrix" virtual reality with intelligent agents that can learn. This approach is more computer intensive, since the environment as well as the agent would require large amounts of computing resources, also, the agent would have to perceive the "environment"
Seems that many more forms of human nature could be investigated in this way.
From 1983 through 1991, George Heilmeier was the Chief Technical Officer at Texas Instruments. He pushed TI into massive investments in AI R&D. Some of the best technical people I knew at TI thought the AI stuff was a waste of time, but it was being pushed by Heilmeier and the executives. Marvin Minsky was one of the experts brought in as an AI consultant, and he appeared in various TI propaganda. At the time the Japanese were pushing "fifth generation computing" which included AI, so there was a push to compete with the Japanese. TI developed AI hardware and software and tried to force fit it into various applications. They claimed various successes applying AI to industry problems, but eventually is all collapsed into a big waste of time and money. Heilmeier left at the end of the collapse.
Today you can find Heilmeier all over the place on various corporate boards and winning various awards for technical excellence. It is interesting that in most of the the bios that you can find on the web about Heilmeier, you don't find references to how he lead TI down the AI path to a deadend.
Marvin Minsky considers intelligence is something like the set of absolute rules, which can be separated from both the subject posessing this feature ('intelligence') and the real world. The reality however is that intelligence is simply a tool for survival in a hostile environment, just another stage in evolutionary process. It has been evolved in specific circumstances, and will continue evolving. It *may* appear as the absolute set of rules (different religious believes is a good example), but this is appearance only. The bottom line is that you have to build the creature (robot or whatever) which interacts with other like creatures and the environment and try to adapt to both. Intelligence may appear as a by-product of this adaptation process. So -- building the robot with intelligence of worm is more productive than trying to simulate Einstein in some computer program. The AI as defined by Minsky and Co. is an emperor's new cloths. This is entirely *his* fault AI didn't go anywhere. The whole concept of separating the intelligence from the background it has evolved is wrong. Only someone who believes that 'mind', 'intelligence', etc is something which was given us by God ('mind' vs 'body') can continue insisting on this approach, but it is no longer a science if you wish, but religion or methaphysics.
> He's right. Theoretical work has ground to a halt in the U.S.
.1 * 1000 = 3,000 TB read - modify - write cycles per second. Modern busses are only hitting a few tens of Gbit.
No, it hasn't.
Those who know what we're doing just aren't advertising it, and on the most part we have to wait for the raw aggregate processing power of readily available computer technology to reach much higher levels before we can put our theories to nontrivial test.
Consider the raw data complexity of the human brain: About 30 TB of synapse states, about 10% of which is actively being read and applied to change the states of other synapses at any given moment, at a rate of up to about 1000 times per second. 30 TB *
If we assume that someone's working theory of sentience requires levels of data and data processing comparable to the human brains', it's going to be several years before it's feasible to put together a computing cluster with that much aggregate main memory bandwidth, and a few years more for the nodal interconnect, even with SpringOS-style duplication of information across the network. Multiple 100 Gb interfaces per node at least.
Right now, the optimal behavior is to sit on our pet algorithms, read up on the progress of others, and try to make lots of money while waiting for commodity computer hardware to get a couple orders of magnitude more powerful.
I switch off between hoping the world doesn't bomb itself back to the stone age before then, and hoping that it does -- there's really no way of knowing whether a successful artificial sentient is going to be our benefactor or a monster, despite the best efforts of its keepers.
In most cases, the hardware and its limitations can be simulated. The only reason that most robotic AI projects are embedded in hardware is because it makes good eye candy for the science press, funders, etc. If you have a good simulation of the environment and the platform, you no longer need to build the hardware for AI research to proceed.
Now that makes a great deal of sense. When I was at university, I did all of my VAX work through either a terminal session or, more commonly, an emulator. It would seem to be a very worthwhile grad project to devise a robotic simulator to be used for future research. Naturally, any half-way decent implementation would allow for plug-in modules to simulate different types of robots.
It should also be able to cope with a variety of different scenarios, to focus on what the AI/robotic research in question is aimed at. Are you trying to cope with terrain, such as spider or walking robots? You should be able to simulate grass, soft/wet grass, rocks up to a certain size, hills with specific angles, etc. Pattern recognition? You must be able to simulate the article to be recognized in many complex scenarios -- rotated, in a crowd, light/dark confusion, etc. (I imagine a good gaming terrain engine could provide a good start here.)
There would be lots of possibilities for future students to extend such a simulator by adding new modules, etc., and the AI researchers/students wouldn't waste nearly so much time playing with cogs, but instead could get down to do their real work.
After all, that's the point, right? AI researchers want to work on AI -- even if it isn't as glamorous as, say, walking talking dancing robots. Right? I mean, I know that would be my dream job, to just be able to knuckle down and work on pure AI.
What's the difference between thinking and fooling people into believing you're thinking?
The same difference between cheating on a test in a subject you've never taken and actually knowing the material: as soon as you're asked a question that isn't on your crib sheet, the charade is over.
An AIBO has a pre-programmed set of behaviors, and any stimulus it isn't programmed to respond to will not have a realistic effect. The same is true of Loebner Prize winner Alice -- the only difference is the size of the pre-programmed response database. Ask it something it's never heard, and it will choke.
When these AIs are able to produce convincing responses to new stimuli, then I'll say that the difference between "fooling" and actually thinking has become irrelevent.
The enemies of Democracy are
OK... Maybe smart robots are the real goal, but without a machine to embody the "intelligence", what is it supposed to do? AI needs a purpose and robots fit that bill really well.
We already have lots of smart people all over the place - what we need is smart robots that can do things that people can't.
Imagine if you could get a whole slew of robots to sort a landfill into elementary components. Imagine if you could get robots to put out fires and rescue people. Imagine if you could get robots to sew any garment you wanted at the download of the latest fashion trend. Just Imagine!
Without extremely advanced senses and mechanisms and the all important control of those things robots will never be able to do these things. Marvin Minsky is right in that those graduate students shouldn't be spending 3 years just getting the machine to work. They should buy the robot and spend 3 years programming it and outfitting it with new sensors. Robot companies should be more common. But the robot market is still in its infancy. Once it gets jump started, it'll be brilliant.
If you were put inside a little white box where you had to flip millions of switches on and off according to certain simple rules, you would look like an idiot next to a computer. A computer can't walk around and recognize things, and doesn't know what an apple is, so what? In my opinion, machine intelligence should be focused on making computers able to make themselves better at what they do best. I'm not sure what a super intelligent computer system would be used for, and I don't think that I would even be able to imagine what would be possible. I would be interested to know what other people think about this idea. Most of the things that I can think of tie back into the "real" world somehow. What would a self-organizing non 3-dimension oriented intelligence be able to do?
Saying that AI is impossible because computers can't come into "our world" of three dimensions, or understand our literature is kind of intelligence chauvinism.
Stallman and co. hadn't spent so much of their time rewriting unix *GNU), instead of working on AI, like they were supposed to be doing. :-)
Vote for Pedro
That's what my thesis work is about; see my web page for details.
The thing about "Society of Mind" is that it's very difficult to take literally. Each page is its own concept - there's not a lot of high-level organization to the book. The concepts interrelate, of course, but formalizing and implementing them is tricky.
The book has certainly served as high-level inspiration for quite a lot of people. A couple of examples would be Michael Travers's LiveWorld and Mark Humphrys's "World-Wide-Mind" project.
But as far as I know nobody prior to me has really tried to make K-lines, polynemes, pronomes, frames, etc., and hook them all together, as described in "Society of Mind".
As an AI researcher and someone who's read Minsky's books and listened to him talk - I can say that he doesn't know what he's talking about. He was big in his time, but things have moved on and he hasn't. He is an old, pesimistic, armchair AI 'researcher' who still thinks AI is easy. He doesn't understand why AI needs to be embodied and situated.
Having said that, I do agree that AI is almost going nowhere (anyone can see that). But I don't believe Minsky understands why.
Those 'stupid little robots' are the best thing to happen to AI - unfortunately most AI 'researchers' don't really understand what they're doing. Consequently, 97% of the time and effort purported being spend on AI research, isn't.
With a few exceptions, the main reason for the 'advances' we're seeing in AI/robotics now, is that algorithms are riding the wave of advances in computing power.
My guess is that you'll see most of the advances in AI coming as more and more 'real scientists' from other disciplines - such as ethology, biology and neurology - get involved in it.
Keep in mind that this is my opinion - shared by an increasing number of people in the field, but still a small minority.
/..sig file not found - permission denied.
A Israel based research company is claiming to have created a computer a-i program that has passed the turning test for 15 month old baby: i.e when talking to it, you can not tell the difference between the computer and a 15 month old baby. By using a behaviorist approach (positive reinforemecent on desirable responses) they are planning to simulate and grow and adult mind in the next 10 years. The have video press releases here.
i worked in a robotics lab for a few years and people did spend all their time doing that with robots. but that's because it had to be done.
what many people do in the robotics community is not AI, which is considered crap by many. just trying to get a robot to localize itself in a room is hard and is not and AI problem, but a complex application of statistics and other mathematics for making use of sensor data (from sonar, radar, cameras, lasers). unlike minsky there are some people out there who know what good engineering and science is, and it doesn't include "AI".
This is a classic battle between Minsky and Brooks. Heck, we had the same battle in our labs (not MIT). I believe that the Brooks response is along the lines of "sure you'll take an extra year to graduate with me, but you'll have one hell of a demo tape." I agree with Brooks. I still show people videos of one of my robots years later. I've never shown anyone any of my simulated robot work afterwards.
The real money will begin to flow once the humankind will stop being scared of direct integrating of humans into computer networks.
I am not sure when, but ultimately all keyboards, mice and screens will take their places in musiums. People will communicate with computers and each other by connecting computers directly to their brain. Thus, the solid knowledge of natural intelligence will be required.
I think first researchers are already working on it in military-sponsored labs. Of course volunteers realize that they can be seriously damaged or dye, but death is natural in military industry. Military industry operates with huge amounts of money. But that's often not exactly a "free" market - all contracts are signed through lobbiing and bribes.
Once first "Unisolders" will be available on the market (sorry, on the job market), then next to militaries there will be strong demand from real-time traders. And that's a real market. Traders will line up to make a neuro-surgery to be connected to those-days electronic stock markets.
I am not sure when such "UI" will be available on the market, but once it will be there, at some point geeks will buy it. The rest of us will be in the front of the tough choice: to stay 100% "natural" or to win a better job contract.
Now, where is AI? The answer is simple: ultimately there will be nor AI neither NI (N as natural), there will be SAI: Semi-Artificial intelligence. No need to think in English letters if UI can get concepts you think of. No need to count numbers if software can do it for you *AND* some AI can do reasoninig about when, why and how you want it done. The trick is that no need to automate the reasonining 100% as your brain is already connected and can do part of the job in that reasoning.
For example, no need to create a very complicated DB query as SAI can use part of your brain to post-filter a small set of data after the pre-filtering of a big set data is done automated in DB engine.
Many problems of software development can be solved if, in addition to humans using computer, computers will use human brains.
That's what i call SAI.
Less is more !
The nineteeth century debate between two camps of biologists, "Vitalists" and "Mechanists," is very similar to the debate between those who think machines can eventually have intelligence and those who think only biological systems can possess intelligence.
Vitalists believed that living beings had something more than their physical and chemical composition which differentiated them from non-living matter. This difference was a "vital spark" or elan vital which made them innately different from ordinary or "dead matter." Their opponents, the "Mechanists" believed that living things were essentially no different than non-living things, at least in terms of what they were composed of. That there was no "vital spark" which separated living and non-living things but rather only a difference in their physical and chemical compositions.
Obviously the "mechanists" won since no modern biologist believes in the elan vital.
In a very, very similar fashion, Minsky and his supporters seem to be making the same type of argument. They seem to want humans to still have a "soul," called intelligence, something that "dumb" matter can never have. Whether they argue for a mysterious quality that only biology systems seem to possess or for mystical "quantum processes" that seem to only take place in brains and not in machines I still call this vitalism and I don't think its scientific at all. It's more like an intellectual retreat to defend some deep seated emotions about humanity's place in the Universe.
Debunking the "59 Deceits"
All this talk of AI everywhere and no one mentions that there is a place to talk about just this kind of stuff at intelligentlabs.com.
That Dell ad at the top of this page IS GOING TO SEND ME INTO EPILEPTIC FITS. I hope they don't mind when I sue.
You are all fartheads.