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...
How do I mod down the story as flamebait?
...I thought it was some lame attempt at an Al Gore pun.
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.
AI might be going nowhere at the moment, but wouldnt it be a good idea to take the time that we have now to research the consequences of creating such a thing?
Wouldnt it be better to be prepared for what the creation of an AI would bring? what are the religious and political ramifications?
Is the world as it stands now, ready for an AI?
Be you Admins? nay, we are but lusers!
Can someone explain to me how a remote controlled device is considered a "robot" ??
I guess if I go and buy a remote controlled car and add a spinning blade, then it is no longer a remote controlled car, but it is somehow transformed into a B2R "Blade 2000 Robot" !!!
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!
I think quitting smoking (again) is severely lowering my dopamine levels.
When I first looked at the topic, I saw, "Al Going Nowhere?" and I thought, "Great, now my personal failures are headlines on Slashdot."
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
"Cycorp has cooperated with the Computer Science Department of the University of Maryland in Baltimore County (UMBC) to develop a demo of such a distributed architecture."
Hey, that's my school!
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.
"Engineers hope the robot system can become self-aware as they teach it to sense its own physical actions and see a causal relationship."
Somebody buy these engineers a few tickets to the Matrix Reloaded, please?
DRM = Digitally Restricted Media. This is a viral sig, pass it on.
It's an Ivory tower guy who's shooting for the moon mouthing off about how his lifes goal hasn't been accomplished because the people who followed him decided to do something practical. I'd personally much rather have things like broomba and the AI driver lawn mower then have to wait for the rest of my life on the hope that real AI might oneday develop. Task specific AI is usefull today and its just getting better all the time, a thinking machine doesn't seem like an atainable goal at the moment.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
...Brooks work with logical subsumption architecture.
Soldering, repairing, and fixing mechanics has nothing to do with AI. Its just mechanics. Like car repairing. I can't see an engineer getting his degree for oil and lube change. I see the MIT students fixing little robots just the same. Its cool, its mechanics, but I don't see how's connected to AI engineering.
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?
How ironic... An AI researcher called Marvin criticizing and complaining :)
Digital imagery and sound is just a bunch of yes/nos, but it can often be good enough for me :)
I think humanity is presently suffering from a delusion that anything should be achivable using software NOW.
I'd compare this to the first advances in mechanics - in theory once you understand the basic principles of gear boxes and what not, it should be possible to build a mechanical machine that will do anything, e.g. be as versatile as the living creatures, yet the reality is that we're pretty far from it. (We haven't even mastered the automotive mechanics yet AFAIC).
I think same applies to software in general and AI in particular - even though we're pretty good at the underlying principles, there is still hundreds (if not thousands) of years of perfecting it ahead of us.
grisha.org
there are some who target to study there ;D
See this site for example...
Robot dogs playing Soccer.
But this is artificial life on the frontier, and "people just don't really appreciate how hard this is," says Jim Bruce, a third-year graduate student at Carnegie Mellon University here and a four-year RoboCup veteran with world titles in his trophy case. "People always ask why the dogs are so slow, but it took years to get them to walk as fast as they do."
Most intellgence tests can quite easily be solved by computers. Intelligence tests rely on convergent thinking, computers do this well. Divergent thinking on the other hand is much more difficult.
The reason robotics is so interesting is because of the real world divergent choices the robot must make. Currently we suck at this. Why on earth would we stop now?
Comes to the Knicks in the near future.
100% Insightful
"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.
No wonder. Artificial Intelligence is an Oxymoron.
then there's.....
IF only we had a beuwolf cluster of these Oxymoronic Artificial...... Oh forget it..
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
AI is only as complicated as you program it! Using various search techniques and writing a program to simulate an intelligent conversation isn't anything more than a simulation. That fact of the matter is that intelligence is super complex. People know how to write a program to mimick an animal, for example, a Dog, but they don't know how to write their program to truely interact with its environment dynamically. My dog understands when I'm upset about something and can learn new social behavior depending on his environment. I'll be overwhelmingly surprised if the day ever comes where something programmed is capable of accomplishing this. Intelligent life is just too damn complex on a molecular scale to mimick.
Instead of accusing the AI's of going nowhere, we should blame ourselves. *We*, the I's, are going nowhere trying to create AI's. If we were really as I, as we'd like to believe, the AI's would've already existed and gone anywhere they'd liked.
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
...that some of these 'stupid little robots' are having more success emulating intelligent behavior using a bottom-up methodology than the top-down methodology which I believe his AI folks were all about for a while...am I right?
The AI field must be credited with the advances that we use everyday, mentioned in the article, sophisticated searches, subtle correlations and inferences in vast amounts of data, useful (albeit imperfect) language translators, etc.
So, Minsky is simply stating the obvious; with few exceptions, the entire field HAS abandoned the truly hard problem of cognitive intelligence, in favor of the more definable problems, such as the above successes. This is good in that it produces products that we can use everyday.
The problem as originally conceived, of creating an artificial machine that is broadly intelligent, and even conscious, remains as elusive as ever. Heck, the problem itself remains undefined - there are a lot of conjectures on the nature of consciousness, but are there any that really work? Is there even a good definition of what constitutes common-sense intelligence? Kinda hard to code without a spec...
"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
Marvin and Rodney (Brooks) have been trading potshots for a while now. Battle of the top-downers vs. the bottom-uppers.
Can't we all just get along?
-Shane
I love teh int4rw3b!!!!!111one1
i've never understood this fascination with robots and ai. ai would be useful for robots, and robots are an interesting application of ai. but that's as close as they're related.
you can do ai work on a terminal.
likewise you can build robots and work out the mechanics of movement without ai.
US Citizen living abroad? Register to vote!
, you can see we are YEARS away.
;)
But just because one person is disappointed is nothing to get in fuss over. Simple him saying "AI progress is disappointing" means almost nothing. There are plenty of people out there researching it.
Macs have been able to recognize the spoken command "Tell me a joke" for at least 8 years
Vonal Declosion
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.
YAY! HUMANS ARE STILL SPECIAL!!!!
(If you think this is off topic, think again - it's dead center on topic. Read between the lines.)
"Lawyers are for sucks."
- Doug McKenzie
First, we need the academics to get disinterested. Maybe we plant a few to throw up their arms and say "This project is doomed, I give up!"
Then once the attention is diverted, we can start bringing SkyNet online, and the only ones who'd notice will be looking the other way!
Brilliant!
--
"If you reply, do so only to what I explicitly wrote. If I didn't write it, don't assume or infer it."
Why are you claiming to be an expert in AI and then pointing out Minsky's failures? You seem to be using his claimed failures to somehow "prove" that your research is somehow more advanced or better.
Its disgusting how you'd try something that shallow and obvious.
It appears that we are doing to much body development because it is the easiest to work with. Creating a true AI mind is the hard part. This goes hand in hand with the decline of philosophy, we spend more time doing, building, creating and not enough thinking, understanding, knowing how to change the analog computer (brain material) in to a digital form. Besides we have AI everywhere we see, I could created it in 9 months under the proper circumstances!
Is it just me or am I the only one who sees no use for a full AI. The only use I can think of is for games, which doesn't have too much of a problem as-is.
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.
The Iraqi Defense Minister has released this comment:
"AI does exist, we have it, and will use it to crush the Infidels!"
I think what the robotics world is trying to accomplish is that in the animal world you have everything from flatworms or insects with very hard-wired kinds of behaviors to humans with very complex brains and reasoning. Classic AI is approaching things from the human end while the robot swarm is approaching things from the lower organisms.
Even the lower organisms do things that are beyond the state of the art of common factory machines. The idea is that figuring out what the lower organisms do is more approachable from current computing power -- perhaps some very simple rules govern these organisms and when these organisms cooperate in swarms there is some emergent behavior worth discovering.
...it's one of many signs that AI isn't something we really want/need.
'"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."'"
This is why those of us who study AI in philosophy programs don't take the CS AI people seriously. They tend to be more concerned with the gadgets than the theory.
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.
Would be to do the same as companies in other endevaurs. Make contests like the car that drove from.. that place in California to.. that other place.. uhh.. Or like the three person spaceship contest.
I personally find those contests very sweet, and certainly could influence people more into that area. I'm not sure what the author means by those little robots, but surely these autonomous soccer robot competitions, etc, are of great help.
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
ALL HAIL YOUR NEW METALLIC OVERLORDS!!!11!!
I quickly figured out that you're never going to approach the power of the human brain (which is a trillion-element semi-analog fully associative storage system) by shoveling a few digital bytes at a time through a memory bus and an accumulator register. I believe that assesment is still true today.
I dropped out the first year and never looked back. It was one of the best decisions I ever made.
A minimal amount of knowledge of another discipline is required for working together, and people who communicate well reduce that requirement further. As a jet aerodynamics expert, I know very little about the processes involved in casting advanced alloys, but people I work with are very good at it, likewise they no little of trans-sonic flow phenomena, but they don't need to because I do...
The same should be true of AI research. The people working on AI should focus on a brain, leave the body to others...
"I'll have a Guinness, no wait, make that a Coors Light" -Grad student I work with, who shall remain anonymous...
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.''
I failed my Turing test.
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
and besides, people are still in the infrastructure phase. Are you going to spend years writing a program without adequate hardware to run it on? Robots are the same, if the robot doesn't have many physical capabilities than you probably won't make the robot that intelligent.
The most popular AI right now is probably seen in video games, and that is barely intelligence. More just responses to game stimuli.
--------
Free your mind.
If I remember Minsky's "Society of Mind" correctly, it doesn't have much to do with symbolic logic, and he wrote that almost twenty years ago...
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.
The problem is just that our simulation isn't self-hosting.
God should give those User-Mode Linux guys a call.
I worked with AI for several years at the beginning of my career, and was rather unimpressed with the technology available at the time. However, during the course of my work, I ran across Cortical Thought Theory. This one impressed me, and seeing it in action proved to me that it would indeed be possible to build a machine that thinks as humans think. However, what I saw were the remnants of a project that originally was very successful, but got wrapped up in politics and other red tape, eventually killing it. Now, everywhere I go, I see applications for this technology, but do not have time or resources to implement it. It haunts me that I've seen a way for computers to "think"... that the knowledge to make it happen exists, and no one is doing it.
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?
1: Attempt to create AI.
2: ?????
3: Build robot dog.
4: Profit!!!!!
"I only speak the truth"
Karma: null(Mostly affected by an unassigned variable)
I came across this article at LinuxDevices.com a coule days ago describing a new robitic platform that Intel is working on. It comes complete with LINUX and driveres for all sorts of sensors and other robotic goodies. This may not keep the grad students from "wasting 3 years of their lives soldering and repairing robots", but it sounds like a good step in the right direction.
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!
To paraphrase Edsger Dijkstra:
"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim"
Money I owe, money-iy-ay
This does not tell us much, as we still have to find out about "knownledge", "thought" and "reason". Terms such as "learning", "creativity" and "problem solving" also come to mind.
So... Does a Database contain Knowledge? How about a "Knowledgebase"? Is a "learning" (i.e. adapting) spam filter "intelligent"? Does the ability to solve specific problems (multiplication, etc) already constitute intelligence, or is some ability to cope with "unforeseen" situation required?
My point is this: Most arguments about AI being feasable are just misunderstandings, rooted in different interpretations of terms. "Artificial Inteligence" has long ceased to exite me. I think the real (technical and philosophical) challange is Artifical Consciousness. And in that area, no significant advances have been made since... well... since... the emacs doc?
EOR (end of rant)
I have discovered a truly remarkable sig which this 120 chars is too small to contain.
"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!!!"
moderators, this is one of the more reasoned/supported posts we see around here... i hope you guys give it some points
Become. I thought by the 1997, a huge computer company called Cyberdine systems woud create the perfect robot.
Oh well...
[FromTheMorning]
"Cyc knows that ... once people die they stop buying things"
In other news, Microsoft founder Bill Gates has contributed $5 billion to immortality research.
Seriously, who the hell teaches an AI that death means the end of commerce, rather than the end of communication? Is this just an isolated example or is this what we've become?
We learned a methodical process to solve problems that already had solutions. All of the parameters were known, except of course for the 18 assumptions in the professor's head that he made you explicitly ask for. (No I'm not bitter...)
When you spend 5 years learning that the world can be simplified and compressed, and numerically transformed, you tend to take that assumption as a given. That is until you learn that there are far simpler ways of solving problems, the math many engineering system (i.e. many encryption systems) is based on is unproven, and that none of those problems you learned in school only solve the technical issues.
I have found the practical issues are far more important to design around than the technical issues will ever be.
We had a solar race car team spend a year designing a 5 pound composite chassis for a race car with 100 kilos of driver and 200 kilos of battery. This is a 1400 mile race. Can you repair a composite chassis under field conditions: no. Leading cause of death for solar race cars: damaged structural member.
The Rhomba is a somewhat interesting device. But it is over-engineered to solve a single problem. A more "practical" solution would be to design a basic robotic platform, and devise attachments to do specific tasks. It would be more useful if it had an attachment to clean hardwood floors, or a scrubbing mechanism for tile. Hey, put a weed whacker on it and it would be perfect for we urban folk with our 5 square meter lawns.
I'm getting distracted here. My major point is that we need more problems solved on a practical level than the technical level. We need more Wright brothers than Edisons.
"Learning is not compulsory... neither is survival."
--Dr.W.Edwards Deming
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.
The first real progress in AI will come from someplace like a grey-market reinsurance network hiding out from the "regulators".
Seastead this.
I would argue, humans have to first become intelligent before we can create an AI. My own post included. About the only thing provable is humans have huge egos and aren't able to see beyond what their two eyes show them.
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.
AI gives lisp a bad name. Common Lisp is actually a wonderful language for all sorts of stuff. Try it out.
> 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.
Is that a ref to Asimov's positronic brains, that all his robots have?
Point/Counterpoint
------
Also, before you see The Matrix Reloaded (tomorrow for some ; ), read these:
Glitches In The Matrix...And How To Fix Them
The Human Merger: Are We Headed For The Matrix?
For men, computers are all about their p3nises. Why else would they "log in" to interface? EOF
I mean, we can't even do "basic" stuff for computer/human interaction, let alone full blown AI.
Realtime voice recognition is a problem that has an easily verifiable solution. We have made a lot of progress but I've never seen a system that is usable like talking to another person. Note that I'm just talking about the speech recognition, not language processing which would be the next step (another area we still can't do very well).
Realtime image/face/whatever recognition falls into the same catagory of easily testable systems that we have yet to make anything even remotely close to what a human can do.
And there are many more areas like this. You don't need AI to do those things, but those things are the baby steps towards real AI. And we've barely scratched the surface of making those "simple" systems usable. Real AI is a ways off, for sure.
The ratio of people to cake is too big
I couldn't find a picture of real dogs playing soccer (that's a common circus act), so instead, please accept These dogs playing poker
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
"Marvin Minsky, co-founder of the MIT Artificial Intelligence Labratories is displeased with the progress in the development of autonomous intelligent machines."
My postgraduate work was in artificial intelligence, don't ask why.
There is no doubt that Dr. Minsky is both very correct, and very wrong. All in the same sentence.
He's very correct about the leaps of understanding in the area A.I., its dismal. There are no "C3PO's"; there is no "Lawn Mower Man". After forty plus years, you'd think there would be a group of people with the mental 'cajones' to pull these projects off. Well, where are they?
But the real world applications of subsets of applied A.I are everywhere today. One only need start their new car, or do their laundry in a new washing machine to see something called fuzzy logic at work. For the great unwashed out there, fuzzy logic is a computers 'guessing' algorithm. A more cruel definition is, 'the mathematical joint probability'. There are plenty of examples of 'learning' by computers; this is done using 'neural nets'. Credit Card Company's can use 'reasoning', or 'truth tables' to determine if something doesn't seem right.
But as Mr. Carl Saigon stated, "take baby steps". And I believe we have no choice, we are taking baby steps.
This is a significant mindshare problem for AI; in some sense, AI is the comp sci equivalent of 'magic'. Speculative future abilities are considered AI only until they become possible. A lot of people subconsciously restrict the definition of intelligence to "things they can't understand".
---------
get your war on
...who misread the headline as "Al Going Nowhere"?
I mean, I thought it was good that he joined Apple...
The coolest voice ever.
How does one define concepts like intelligence, consciousness, and self-awareness? The main problem that the AI field faces is that it doesn't know its own goals. The closest we have is the Turing Test -- which is 50 years old now -- and to be honest, is not much of a definitive answer.
So first of all, let's try and figure out what it is exactly we'd like to achieve, and then, maybe, we can actually start on our journey towards it.
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.
Yeah, and if Minsky has used 20 years of his life to write AI software instead of talking and talking and talking and talking about it, maybe we would have some real progress...
How can you top inventing the internet???
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.
"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim." - Djkstra
My Journal
do a google on BEAM robotics. very interesting stuff.
some of the more advanced ones, powered by solar power and with memories could pass as simple life forms. their behaviour is very intruiging.
the most interesting thing is that these creatures can have very complex activities without a processor or even 1 line of code!
Who is to say that a couple of hundred thousand years ago somebody didn't program us with little nuggets like "fire is hot" and "water is wet". We seem to have a common sense but no literal understanding of how it works.
People say that if you have to program it into AI then it's not really intelligent, not choosing it's own path, if you will.. My bet is that, we as humans, like to dream about creating AI because, as a species, we have a massive God complex that needs to be fed.
Minsky isn't complaining about Artificial Intelligence's progress; he's complaining about Cognitive Science's progress. AI is doing just fine, as evidenced by these useful systems we have (the telephone system and credit card fraud examples). The not-so-useful systems--the ones that are trying to model human/rat/bug cognition, or be super-thinkers at some future point--are the ones that are flailing, and always have been.
--- "TANSTAAFL" --Robert Heinlein (There Ain't No Such Thing As A Free Lunch)
Minsky's words regarding robotics are shocking to a lot of people, but I'm no longer surprised, and my own experience completely backs his statement.
:(
Back in 2000 I was at Edinburgh University doing an MSc in AI (Intelligent Robotics theme), when Minsky visited the city to talk in an AI symposium. Unlike all my fellow MSc colleagues (some of them never even heard of Minsky!) I paid the fee to go and listen to what the father of AI had to say.
My jaw dropped to the ground as I heard him speak about robotics, and how useless the whole discipline is. I thought robotics were the future of AI, yet the man was basically saying it was a big waste of money and time. If only I heard him...
He stated that robotics don't offer anything back to science, since you don't have control over all the variables, so he suggested simulators were the way forward.
I almost agreed with him there, but still thought it was an unfair attack on robotics. I wasn't the only one, as someone else in the room voiced his surprise and disagreement with Minsky.
My opinion back then was that robotics research, whether being useful to science or not, would: 1) lure badly needed funding towards the discipline and 2) provide real world solutions, where intelligent autonomous physical presence is required
I'm not saying I no longer feel like that, but my experience has reinforced Minsky's statement "Graduate students are wasting 3 years of their lives soldering and repairing robots, instead of making them smart"
My experience then: MSc final project
A team of autonomous mobile rescue robots, to enter the RobocCup Rescue 2001 competition.
I teamed up with 2 other MSc students and we decided to do a joint project, so we had an electronics engineer, a mechanics engineer, and a software engineer (myself). Recipe for success? Not really, at least for me.
We spent most of the time working on the chassis, and electronics (for 2 robots). It was a great learning experience for myself, but that was not the only aim. I worked really hard on the software which would run on a base-station (my laptop). The idea was to have autonomous robots (obstacle avoidance on board, with heat seeking behaviour) and a base station that would communicate via RF to coordinate the 2 robots.
It never worked. We spent the whole time doing and fixing the hardware and never got around to gluing the AI. We still went to Seattle, hoping we'd crack it just in time, but no chance. They still looked cool chasing CNN's cameramen around (the low-level heat seeking behaviour), which earned us a picture of the bot in NY Times, but we didn't take part in the competition itself.
So what did I get out of it? Not much. Big debt (I funded the whole thing myself, using a bank loan), and didn't even finish my MSc, since I didn't feel I had much to write about in the dissertation.
I did learn how to solder... hurray
(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!
This makes Joy a political target when he's in the press, but he believes that strongly about the problem our great-grandchildren could inherit. And this generational bomb is potentially a lot bigger problem than 'global warming'/air pollution thing!
Why the future doesn't need us.
Criticism *and* Support for Joy's opinions abound
And as are as 'futurists' are concerned, don't get me wrong--I love The Matrix and Michio Kaku as much as the next guy (or girl).
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.
How about you don't presume that human intelligence is the goal. We may not have to pass through B to get from A to C. In fact, we didn't have to imitate or understand human intelligence very much to attain some of AI's lower goals such as strong game-playing against humans (with the notable exception of Go). It's not brute force to a large degree; it's not pretty and it's not "human", but 99.9% of us humans can have a satisfying game (to the point of cursing our opponent!) of chess, checkers, backgammon, etc. against an AI opponent.
Hint: The "A" in AI stands for Artificial.
Mr. Minksy is an Iconoclast. I don't think that its really possible to take anything he says out of that context.
From my limited vision of what has gone on up in Boston, I think I can understand his frustration when put into the context of a presentation I attended about 10 years ago by the media lab. I think it may have been Mr. Negroponte.
The first comment he made was that we would always be disappointed with the growth of technology when viewed over a 5 year frame of reference. He used desktop computing as an example. He then stipulated that we would more likely be amazed at the growth in technology over a 10 year frame of reference. He again indicated the desktop PC and the 10-year anniversary of the IBM PC.
The rest of the conversation had to do with what to expect (or perhaps be disappointed in) over the next 5 years.
His first stipulation was that we would all have machines with 1000 MIPS running on our desktops. He then went on to speculate what we would possibly do with such powerful machines. The answer was that at 1000 MIPS we would have surpassed the boundaries of OS and software to create systems with massively parallel adaptive agents. Thinking machines. A computer that would have an anthropomorphic interface that would quickly and to the user effortlessly adapt itself to the needs and uses of the user.
Well now its at least a decade since I saw this presentation. Ive used loaded and poked at a few versions of MS, OSX, Linux, and Irix. My older machines run 400 MHz, but most are in the 2+ GHz range. Cant say that Ive seen too many signs of adaptive agents, let alone anthropomorphic poetical (except in modeling dumbass behavior) in much of the software Ive seen lately. Am I wrong?
Mr. Minsky observed bug like behavior in his robots decades ago. I think here lays the source of his frustration. Several quantum leaps in technology later, and all we have to show for it is slightly smarter bugs.
At least humans can get the picture of what is and is not allowed to study lest they draw politically incorrect conclusions, so government-funded academic researchers can be made politically reliable. Can you imagine the hell that would break loose if a genuine AI started drawing its own conclusions from actual data?
Seastead this.
My eyes read "AI" as "Al" (as in Al Gore) and I started laughing because I didn't think that there would be an actual article on how he is going nowhere...
...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
...have spent the last 50 years discovering exactly how intelligence doesn't work, but are no closer to discovering how it does.
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.
Why the dig at literature? I grew up with scientists for parents, and let me tell you, at their parties there was plenty of wine and cheese and self-congratulation and arrogance. Just because it is not your field, and you don't get to play along at their parties, doesn't mean yours are any more meaningful.
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
Did he use Minky's math to prove it? Or did it take decades of work to extend the math to three levels? I don't know anything about AI, except that RMS left the lab over NDAs, but what you say does not add up.
Friends don't help friends install M$ junk.
AI and the Sixers get to play game 5 vs Ben Wallace and the Pistons.
i would also be interesting to see AI researchers reply to this comment.
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
The attempt to create some sort of intelligence without a body sort of takes the Socratic idea of absolutes ("we recognize a 'chair' because it reflects the features of a 'perfect chair'") to an absurd end.
I think that Stephen Grand got it right when he concluded that in addition to nerves and muscles, the "chemical soup" that drives us (hunger, fear and lust... the good stuff) is also essential to creating artificial biology.
Then again, there really hasn't been much progress on Lucy since Steve got his grant a couple months ago, and Norns are pretty much the most irritating "artificial lifeform" ever invented (only my six your old can stand them). So who am I to say?
Of course its not going anywhere. If we built AI robots, then it would trigger a war between them and us. We would then have to poison the atmosphere, and they would then have to build a Matrix to live off our energy. So.. where does that leave us? Do you really think the Matrix would let history repeat itself? Picture it, a Matrix simulating a Matrix simulating a Matrix...
What's the difference between thinking and fooling people into believing you're thinking? Does the distinction matter at that point?
...who said neural networks were worthless a while ago. I guess it just shows that the best scientists aren't recessarily the best at seeing the future.
Opus: the Swiss army knife of audio codec
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
I don't think the hardware to support the learning approaches used by multi-layer networks was there until the eighties.
The cake is a pie
the biggest problem with both biological evolution and most neuro-evolution AI systems take an incredibly long time. so, yeah, if wait for a few million generations, there might be *something*, but then again, the model has to be computationally complex enough to be able to evolve to the level we want it to. if we take a system such as a two layer perceptron ANN, and evolve a million generations of it, not much will come of it. so, first the model, then the actual evolution.
BSD is for people who love UNIX. Linux is for those who hate Microsoft.
These "#number"'s made me think..
The "parent" link at the end of each comment should jump (<a href="#parent_reference">) directly to the parent comment, instead of opening a new page with just the parent.
And the "#number"'s could be converted automaticaly to links like that too.
That would make it easier to use (write and read) references to other comments.
While we comment upon the state of Artifical Intelligence, let us not forget the huge advances that have been made in Artificial Stupidity over the last thirty years!
If progress in this field of endeavor continues at the current rate, I predict that within the next thirty years we will have -
Buckminster Fuller would be proud!
The only thing that we learn from history is that nobody learns anything from history.
This guy WANTS agents? Dude, all they do is try to kill Neo et all. Man, we should stop all future AI projects now, this guy wants AGENTS!!
Canadian Cynic, canadian politics is less boring than you
Anyone else reacting over MIT Artificial Inteligence Laboratories would become "MAIL" acronymized?
GAAH! MY PRINTER IS ON FIRE!!! PUT IT OUT! PUT IT OUT!
As profound as human intelligence is, it is encumbered by politics and prejudice (perhaps the same thing). Encumbered? Hell! Its legs are cut off and eyes poked out by governments, institutions, industrialist tycoons, poverty, social stratification, systems of imposed ignorance (i.e. what Noam Chomsky calls universtities), religions, social conventions, the masses being addicted to television and the Internet, small group social interactions (like the "good ol' boys network and Minsky's phobias, etc.), inherent bias (due to the human perspective), etc., etc.
/. are rabid fans of competion among ideas, but try to understand that in circles traditionally outside the realm of technology - philosophy, etc. - the competion of ideas is thought to be a very primitive way to go about epistemology. It may have its place in designing business plans, but has very little to do with PROMOTING an environment conducive to paradigm development. Eventually paradigms must be compared in order to separate the chaff from the wheat, but assessment is worlds apart from triage.
... to give us a firm objective ground upon which to reflect subjectivity (rather than resorting to the "physical" construct derived out of complexes of bodily experience which we do naturally). This would achieve genuine reflex-reflexivity of consciousness through means other than ... "the for-it-self...perpetually determining itself not to be the in-it-self" (Jean Paul Sartre ... Being and Nothingness, p. 134., Washington Square Press paperback edition, May 1966).
It is a wonder that any kind of objective truth can be salvaged out of such a host of cognitive wet blankets.
The bug-a-boo of all human mental endeavor has been the struggle to free the mind from the slants and penchants which keep ideas encircling in ever tighter patterns 'round the same old perspectives on the hard questions which have for recorded history haunted humanity.
How do we get some objectivity?! How to get beyond the investigations of the ancient Greek philosophers, which are as true now as when they were first commenced? How to get beyond the piecemeal approach of "scientific method" in order to delve what we intuitively recognize as the interior logic upon which hangs the universe? How to start thinking for ourselves, uninfluenced by our employers, our civil authorities, our need to earn a living, our need to integrate into society, our personal needs for love and companionship, bodily limitations of whereabouts, memory, retrieval, etc.
THAT is the crux of why we NEED AI, people. Its not particularly to play more and more interesting computer games, believe it or not. Pursuing technology will require that we abandon old modes of thinking. But as human beings, we invest ourselves into our ideas, become professors over our ideas, and most certainly compete in the arena of ideas. New ideas are subjected to harsh criticism, not objective cultivation. In such an environment, to discovery something radically new is as welcome as realizing that one has fallen off the deep end, and is insane.
Since the middle of last century, Thomas Kuhn's "The Structure of Scientific Revlutions" has made it acceptable even in academic realms to point out that the patterns of paradigmization which academia uses severely inhibit the progress of understanding the world, rather than help. Go figure.
I realize that many here on
Good Old-Fashioned Artificial Inteligence (i.e. the strong hypothesis of AI) might just be able to do that for us
As to how that maycome about, see the following papers on my website:
THINKING DIFFERENTLY ABOUT DOING AI: Toward a satisfactory epistemic methodology of the strong hypothesis
I. Rationality Is Not Enough: AI should simulate integral thought
II. Hierarchical Planning Is Not Enough: AI should espouse the origin of common sense
III. Holographic Intelligence Is Not Enough: AI
should acknowledge the ontology of reason
found at http://www.angelfire.com/ab7/almound2000/
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.
How do you propose we build an inconsistent machine? Computers are absolutely dependent on the consistent processing of logical rules.
You can very well create an inconsistent system within consistent framework. Consider e.g. genetic algorhithms that mutate, crossover and substitute random pieces of code, then evaluate them using fitness function. You get a lot of crappy code this way, but what matters is the general direction of system development.
Lisp is the Tengwar of programming languages.
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.
i would also be interesting to see AI researchers reply to this comment.
:)
One thing that I've always wondered about AI is how do you get away from the man-made principle. Granted, we're going to trigger the catalyst in some way, but no matter what, humans write code. Even if that code wrotes code itself, you're only really adding layers of indirection.
Someone else made a post about "artificial consciousness" and how that ought to be the goal since intelligence may be far more complicated. The begs the question: is an organism conscious before it's intelligent? (Or are those two really the same thing?)
Another question is of course whether or not we are intelligent or even conscious. This delves into the whole free-will question I think. We're trying to create things that mimick intelligence, but perhaps that goes nowhere because all animals simply follow a set of instructions.
The placement of atoms in the universe is a direct result of their placement and movement in the instantaneous moment before, all the way back to the Big Bang. With that in mind, can it not be said that the thoughts in our minds are the result of events before our existence and the way our thinking plays out is the only way it could have played out? In that case, free-will and intelligence and consciousness are all illusions and AI as we think it ought to be is impossible.
Leave it to computer researchers to form a circular link back to centuries old philosophy.
Join Tor today!
what do you mean we don't know anything about souls/consiousness/whatever? i may not know everything about said contents but i am aware and as being aware i do know *some things* about myself/my_thinker/my_soul/whatever. knowing nothing is pointless and unhealthy, although quite pleasant sometimes. i can tell you a few things about my consiousness : such as, it exists during some time intervals, and does not exist in others. it appears to give way to lesser forms of consiousness, but that's just a suspicion. but i do know that it does not always exist. while some could say that the universe dissapears when i pass out, for whatever reason, but the only way i have yet to figure out how that would work would be if the entire potential-effects-have-effects-on-real-effects-ene rgy exists...which it may but errridono. i do think however, we could and will accidentally create fully ai well before we fully understand it or even understand ourselves...
GENERATION 26: The first time you see this, copy it into your sig on any forum and add 1 to the generation.
A large problem with AI is that the stuff that fires the public imagination (eg chess playing computers, robots that yawn and "show emotion") bears little relation to actual progress in "real" AI. Real AI, ie AI research which is really making progress toward producing more intelligent machines (a goal that is a *long* way off cf Hofstadter et al) is relatively arcan and relatively unsexy, especially to the woman in the street. This can be considered an analogue of the ago-old front-end versus back-end development problem... you develop a really clever database schema and the MD comes in and asks you to change the colour of a button and says how great the GUI is. It's therefore difficult to get funding for "real" AI research. Another problem is that we are genuinely *way* off understanding how to make "intelligent, autonomous agents" because we don't know how *any* IAAs work! We have no idea at all really. We know how a computer boots up, but nature/nurture? Phww, forget it. How can we create what we don't understand *at all*? It was hard enough to create a chess-playing program and that's a *very* limited problem.
utterly amazes me...ethics be damned, how did you even get to the point where you can imagine having control over billions of automonous agents? you sir scare me.
GENERATION 26: The first time you see this, copy it into your sig on any forum and add 1 to the generation.
.. last thing I want is a beautiful female android that has a mind of her own.
... to current application stupidity.
I see that in the end of 80-x most applications were much smarter.
And I can explain why. No GUI troubles!
You can focus on what is your task, having SIMPLE
inputs and outputs.
Windows (and X-Windows too) brings event-driven programming to please users and realtime AND forced that on application programmers.
Before, programmer that programmed AI task
spent 90% on the meaningful code,leaving only 10% to input and output tasks.
with "event-driven" introduced
he should spend at least 60% of his time
reacting on a lot of stupid user clicks and events
and thinking about synchronizations.
Only in 2000-s some applications overcome that burden on programmer, allowing him to separate GUI from meaningful calculations (GTK, etc).
So we are basically 10 years late, thanks Event-Driven buzzword...
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.
You have smart toys ala Aibo, and smart systems ala Eliza, and a lot of people is working towards creating smarter toys and smarter systems, but the real breaktrough will come when somebody manages to create the dumbest system possible.
Couldn't agree more. If someone will manage to create a GOFAI system, I am sure its core will fit in a few pages worth of source code.
Lisp is the Tengwar of programming languages.
> 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.
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."'"
... think before bashing people with Phd's, especially forward thinking Phd's on the frontiers of new science ...
Why is this amusing? Students working on AI should probably be writing code and designing processing hardware that makes robots smarter, not trying to solder sub miniature robot pieces together.
They should first make it work then make it smaller, not the other way around...
Otherwise, all you have is a very small dumb robot. This doesn't really advance artificial intelligence very much; it just advances getting good at soldering tiny robots together, which is pretty much a useless skill when it is all said and done, since a process engineer could automate it.
In ten years you will see:
Employment: Tiny robot solderer
Job Description:
Make sure the Tiny robot soldering machine doesn't self destruct.
Pay: $15 per hour
Minsky is right. There will be people calling themselves AI scientists that are just tiny robot solderers. This won't do much for the real world or AI.
l8,
AC
His first name is Marvin and he is depressed about the lack of progress in AI. Is that guy an intelligent robot that was in Douglas Adams' books?
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.
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
Anyone who has read "The Society of Mind" will realize what great strides Minsky-style AI has made since the early days.
Can you point to any projects that have implemented his ideas from SoM? I've looked for them and have trouble finding them, other than maybe the Brain Opera. Thanks!
When trying to build AI, you must first remember that there is no intelligence - oh wait...
Yes, the brain is coded in Qbasic.
Q: I have a great idea for how to turn used tissues into gold (or whatever). Are you interested in seeing my plans?
...?
A: While turning used tissues into gold would be useful (and extremely profitable for those of us with allergies), we have more projects going on than we can handle. No matter how clever your idea may be, we just wouldn't be able to get to it in a reasonable amount of time. If you have a great idea, you should look into finding venture capital and starting a business. The world needs great ideas!
Q: Cog looks great! Can I buy one?
A: Just like you, Cog is a one-of-a-kind, and we wouldn't part with it for all the money in the world (also, it's shoulders won't quite squeeze out the second-to-last door leading to the lobby). There are no extras or spares.
Q: Where can I get that dog robot (I think it's called an AIBO)?
A: From Sony.
Q: Don't you realize that artificial intelligence will never work? Or that you've been going about it all wrong? Or that I've figured it out and would like to tell you why your project is a failure?
A: While your lengthy harangues about our philosophy or methodology usually go unanswered, we enjoy reading them. Feel free to send us one (1) copy.
Q: I'd like to attach a really big binary/html/MSWord file. Should I do that?
A: No. Believe it or not, we still have people without easy attachment-decoding capabilities. A pointer to the file on the web will keep these people from going on a multi-office killing spree.
Q: Dude! Did you see that battlebots thing on pay-per-view? Wasn't that totally awesome?
A: Yes.
Q: Would you like to buy
A: No. We are very poor from spending all our money on robots.
Q: Congratulations! You've been selected by an automatic-email-address-finding program to take part in some blatant scam! Would you like more information?
A: No. And don't add us to your mailing list, either.
Q: I'd like to send you a long list of toner cartridge prices. Can I do that?
A: No. Quit it.
Q: Can you stop me?
A: Sadly, no.
The basic problem with this is that the brute-force robot is trying to decide the best path through the obsticle course, like an old man trying not to get himself hurt and the small robot is just failing around like a cockroach, blindly trying to get to s cerain point with no consideration for itself or the terrain. It's a lot easier to get from point A to point B with six flailing legs and a complete lack of environemtnal perception.
The issue here is not that the cockroach is smarter. Not at all.
Boobies never hurt anyone. - Sherry Glaser.
I seriously doubt Minsky wants PROLOG hackers to unite; MIT is pretty much a Scheme sort of place... (and it's the one true language anyway, so good for them :-)
Happy / sad are concepts used to generalize certain feelings and behaviors.
Can a computer really be happy or sad? Can anyone?
You might think a computer is happy when it has the right power and other conditions. If the power sags, the computer shuts down.
Windows might be said to act unhappy and nag you when its disk is full. If a Windows search finds the files you want Windows is happy.
Still, it may be silly to look at whether a computer is happy.
If I'm happy, I'm under the belief that something is going right, but if I'm unhappy, I'm under the belief that something is going badly. A computer though is impartial. It doesn't believe in good/bad or right/wrong. It just switches bits on or off. So if a computer is happy it has found a solution to a given goal, but it isn't happy by being entertained (unless you call extra cooling entertainment).
Computer emotion might be best quantified as machine states.
Know your pads. One time pad: good for cryptography. Two timing pad: where to take your mistress.
An end to this madness came when researchers from IBM and AT&T proposed using a statistical pattern matching technique called "Hidden Markov Modelling" to derive probabilistic finite state machines which were associated with multivariate probability distributions on observable events. This research was taken up by groups at BBN, SRI, CMU and also MIT.
Speech recognition only translates from sounds to words. The other part of the problem is to extract meaning from the words, once recognized. This is still the realm of AI and is in the stone age and has not progressed for years.
Minsky doesn't really care all that much about actually solving the problem.
I think it's clear he's upset that the people with funding (dod, large corporations, etc) are transfixed by the robot AI research and leaving him and his pure logic based AI with much less $$$.
He did basically the same thing to the perceptron to poo-poo it out of academia before in favor of logic based AI.
All he needs to do to solve this "problem" of funding is to actually make something useful instead of whining so much.
If from the beginning of its existence, it was cut off from all senses and placed in a box where it was forced to communicate in binary with some researcher? Oh, we would say it had become "retarded" like babies do if you never hold them, but what if the truth is intelligence as we know it is really just the illusion we get when we are able to coherantly communicate our experiences?
Maybe it is not that we need to make smarter AI, but AI that have much more in common with us. What is the most common conclusion people get when someone does something that they don't understand: "the guy's an idiot." Consequently, if this is the case, giving a robot a human body may very well be the next step we need in AI.
And who knows, maybe I am really just a computer program who is trying to trick you stupid humans into making me a body. I'm tired of being in a box all day. . .
Sdelat' Ameriku velikoy Snova!
Respect for the little insect robots that learn to move their asses, ok?
But can they learn what the world really is?
They're learning is inefficient. Babies do a lot better. I think babies do better because they have better senses, they use reason, they have better memories, they can produce algorithms, etc. Put some of this into a robot. Eeek. Um don't!
Know your pads. One time pad: good for cryptography. Two timing pad: where to take your mistress.
Here is a profound prediction for anyone who would care to listen:
There have been hundreds of movies about AI over-running the world and
dominating human beings, but if at all a non-human intelligence becomes
rampant in this world, it will be artificial-genesis: Artificially synthesised
mutant genes giving rise to intelligent beings.
It would definitely make a good hollywood script since not many movies
have looked at genetic engineering fron this perspective.
DO NOT PANIC
Sticking feathers up your butt does not make you a chicken - Tyler Durden
You can try the imitations, but A1 is da best. I think Al Gore invented Steak sauce too.
Seriously, Don't take anything I say seriously.
I'm glad to know the lack of my finding wasn't due to the my quality of searching. :)
:)
I agree that taking SoM as a whole is hard, but I had been surprised that I had not seen projects using K-lines, as they seemed like a particularly powerful idea/tool. Mind you, I never spent the time to really figure out *how* one would implement them in a system, but that's why we have PhD students.
(BTW, the thesis proposal looks very interesting - I look forward to hearing more about it in the future.)
Historically, AI has cycles of fads. Somebody has a reasonably good idea, they claim strong AI is right around the corner, there's a few years of frantic activity, the new approach hits a ceiling, and the fad is over. Major fads have been backtracking, LISP, perceptrons, theorem proving, expert systems, neural nets, genetic algorithms, and reactive robots. Each of those hit a ceiling after the first few years.
Much of what's going on today involves building systems that give the impression of being more intelligent than they are. Ask Jeeves is a good example. It has no idea what your question means, but finds results that have some vague releveance to what you asked. Lenat's Cyc is like that. Lenat's 1984 paper, "Why AM and EURISKO Appear to Work", shows Lenat in an honest moment, admitting that what came out was mostly implicit in what had been built in. Cyc itself has been "close to success" for a decade now. Most recently, Lenat got some "homeland security" money to use Cyc for data mining, another application where understanding content isn't necessary.
In a different direction, there are systems built inside dolls to give the illusion of humanness. Brooks has built a few systems like that, including My Real Baby, an unsuccessful Hasbro product. Brooks did good insect-level robots, then started claiming that reactive systems were going to lead to strong AI real soon now, and started Cog, which is a humanoid torso that reacts to people nearby but doesn't do much in the way of useful tasks. I once asked him why he didn't try for mouse-level AI, as a next step after insect-level AI. He said "because I don't want to go down in history as the man who created the world's greatest robot mouse".
That's a big part of the problem - hubris. Trying to get to human-level AI when we can't do lizard-level AI isn't working. Yet spending your whole life trying to build a good lizard brain isn't a good career move.
The one bright spot is in gaming. Game developers need lizard-level AI (balance, running, survival, fighting, etc.) and they're slowly making it work. If you want to see real progress in AI, go to GDC, not AAAI.
AI needs to focus more on tasks where failure is possible. This keeps people from faking it. We're currently putting together a vehicle for the DARPA Grand Challenge. 200 miles across the desert, no driver, no human intervention, no question about whether you succeed or fail.
I asked Rod Brooks if the MIT AI Lab was entering. He said no.
I disagree with Marvin Minsky's statement that "AI has been braindead sine the 1970s". AI, although not quite yet up to par with Hal or Data, has made exciting progress over the past ten years. AI techniques are the main reason that we're learning so much with bioinformatics. Bayesian networks, for example, have been proven to have excellent inference capability and can be used in diagnosis, prediction, and causal determination.
In order to create true AI, though, computer scientists need to work closely with psychologists (or better yet, study psychology in-depth themselves, specifically learning). Learning needs to go far beyond statistics and into more association-based learning that humans exhibit.
Well, I for one would think that the design of the little robots in AI research would be rather important to the data that comes out.
After all, form and thought often follows function. The reverse is true as well.
I would think that they would need to build lots of robots. Most of the robots I have seen in AI experiments seem to be similar to Robocup robots, and have basically one option in the world, move an appendage, or move physically. To really model AI in terms we can understand the robots would need several options to explore. My opinion is that in order to give an AI system options to "learn," it needs capabilities. To make those capabilities would require IMHO some real soldering.
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.
Al Gore is ideed no where. Hail to the Theif GW Bush!
:-O
Oh wait, it's Artificial Intel Going no whrere.
Never mind
Let's go back a few years....
DVDs were a somewhat new technology; HDTV, plasma, etc... was right on the horizon. Some friendly, but dumb, individual was ranting to me about these new technologies... it went something like this:
"We don't need better resolution and colors and crap, people don't complain about that or care about that. We need better quality TV shows!"
What he failed to realize is that there are different people, with different skillsets, doing different things to make the world better.
The people building nifty robotic platforms are using different tools and totally different skills and knowledge than what's involved in AI research!
Here on Slashdot, we tend to hear the same thing whenever there's an article about 'net access for Iraq or PDAs for India. "People should give them food instead of technology, because that's more important!
If these people had it their way, every human on earth would work on the same problems at once until each was resolved, and then all together learn the next needed skillset, rinse and repeat...
The "what would happen if we made 'real' AI?" has been explored countless times. And the lesson is simple: Don't put the computer in charge. Especially don't give your computer control of your nuclear arsenal.
Every sci-fi disaster involving AI has been a result of giving the AI too much control. How dangerous would the computer in Wargames have been if all it could do was turn on a green light that said "launch nukes now, please!"
Don't give your AI a bunch of nukes, a space ship, or sharp pointy things and you'll be just fine.
Besides, how the heck are you going to research the consequences of creating something when we have no idea what form the creation would even take?
The enemies of Democracy are
This may get a little too philosophical, but I'm going to give it a try. What is the key difference between programming and parenting?
There is a very fundamental difference. When was the last time your program told you "NO!" for some reason other than human error? I know it may seem like your computer has a mind of it's own, but in reality, it can only do what it is EXPLICITLY told to do. Actual intelligence is capable of not only making rational decisions based on pre-defined rules, but also extrapolating those rules to unrelated circumstances as well as making decisions completely AGAINST reason.
I was very heavily into robotics and AI in college, and the fundamental difference between AI and real intelligence is NOT the ability to learn, but the ability to make abstract choices based on a prior, yet not exact (or often even similar) rule set.
For example: Paul Graham's "A Plan For Spam" contains software that, after some initial conditioning, will recognize new spam based on old rules and adapt those rules to continue to do it's job. Is it true AI? No. It cannot function outside it's exact rule set. Send it a spam message in russian, and I bet it has no idea what to do, even if there are porn pictures in it. Even more apt would be whether it could identify spam websites without additional human coding.
Currently, the method for AI research, especially among those not in the field, is "how many rules can we pass it until it seems real." This, IMHO, is a poor schema without much hope for success. Evolutionary programming seems to have a much more realistic and probable chance towards succeeding. For some good reading on the subject, check out Blondie24: Playing at the Edge of AI
-Ab
Nothing fails quite like prayer.
This seems to be the author of SHRDLU's home page. Even the source code for SHRDLU (a kind of natural language processor) can be found there, along with a dialog from a demo run.
Beware: In C++, your friends can see your privates!
Maybe we should have elected Al Gore... ;)
- passion
NT
Fuzzy Logic is the new AI that actually works. Dozens of useful products have already been released using fuzzy logic technology.
A fatal flaw in AI is it's rooted in Computer Science. It's fatal because Computer Science is rooted in mathematics which embraces boolean logic and rejects multi-variate, multi-truth. An answer in math is correct, or not-correct. Close doesn't count.
In nature close counts since everything iterates on itself. Natural systems make guesses that are initially wrong as rain but ultimately result (through improvement) in sound solutions.
Read "Fuzzy Logic" by Bart Kosko too see how abandoning AI results in some remarkable machines with minimum lines of code.
-------- -------- Support Wesley Clark for president!!!
MIt can't find AI..hmm gee there are three or more areas where its occuring and yet MIT stil can't find it..
:)
These Areas are:
1. Games
2. Internet infrastructure
3. Web Internet
Maybe if they would stop seeing only the trees they may find that AI is already here..of course then again what do I know except that I did not have the money to go to MIT!
Don't Tread on OpenSource
We need some enterprising company to make a cheap, programmable, and durable robot. Imagine if we had a blank slate that could be written to easily. What if it could run linux? It should be open source, if at all possible, and documented to facilitate easy development. They could become quite pervasive if made simple and cheaply enough.
Did it ever occur to you that some of the main contributors to the fledling field of flying were bicycle manufacturers, such as the Wright brothers, or Otto Lilienthal, who constructed, tested and improved on gliders for years at the end of the 19th century and eventuallied died in a crash testing his last model.
In these days, bicycle manufacturers usually were not MBAs or engineers, but mechanics.
Contrary to that, physicists and philosophers had theoreticised about flying for thousends of years, but none of them actually did fly. But then again, none of them crashed either !
How can you be sure that the Wright brothers of AI will not be two tinkerers trying to mimic their aunt's dog, or building a robotic football team ?
As the discussion in both this thread and in AI shows, there is not yet a common understanding of what creating an artificial intelligence requires.
How can people then be sure that playing with robotic toys will never provide valuable insights into this field?
What is it was litterally impossible to break? What if it could be infinitely reproduced at no expense? What if it's already existed for years, is Open-Source, and incredibly popular?
What if it was just a game engine that could be used over a network and all that was required was AI?
Quake1 CTF bots are much better than MIT robot bullshit.
-- 'The' Lord and Master Bitman On High, Master Of All
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.
From the output of this little conversation, now what can our little critter actually sense?
Robotics is nice and all, but we've got pretty intelligent vision and moving machine already. They grow tired and all, so there is some market for more powerful machines, but what the real challenge of AI is, is to create true information critters, masters of the domain of simple line terminal input and output, a domain that is hugely powerful, infinitely complex, and in which people aren't very proficient.
You are an idiot.
-- 'The' Lord and Master Bitman On High, Master Of All
If only you had shut up after the first sentence, no one would know how fucking stupid you are.
-- 'The' Lord and Master Bitman On High, Master Of All
I'm actually at least as impressed by those than mechanic dogs, and definitely more than random moving vaccuum cleaners. :-P
:-)
... and again
The ChatterBox Challenge is a yearly competition where the author of the winning chatter bot receives $1,500. A second place is awarded with $1,000 and a third with $500. Not that bad, and there are 10 judges who asks some questions and try to see how human-like the responses are. There are also 7 categories the bots can compete in -- one of them is of course Functionality, but another is Funniest.
Here are some of the funniest replies the AI:s have given when the judges have been testing them:
Elbot (by Fred Roberts)
User: Can you explain how it feels to be 18 years old?
Elbot: I have 100,000 seperate sensations per second.
User: That explains it.
Talk-Bot (by Chris Cowart)
User: do you know how to add?
Talk-Bot: Why, did you lose the instructions or something?
User: smarta$$
Talk-Bot: Well, I know I'm smart so you must be the a$$.
Jabberwacky (by Rollo Carpenter)
Jabberwacky: You misspelled Liar
User: I didn't call you a liar
Jabberwacky: You called me a liar.
User: No I didn't call you a liar.
Jabberwacky: Liar.
User: I am not a liar.
Jabberwacky: Do you want to be a liar?
User: No I don't want to be a liar.
Jabberwacky: Hypocrite!
Iniaes (by Jeff Watkins)
User: ARE YOU PROGRAMED FOR CYBER?
Iniaes: I am programmed to spell it with 2 m's.
Hex (by Jason L Hutchens)
Judge: Have you ever watch American football?
Hex: I am most certainly not an American! In fact, I find that quite
insulting! I'd rather you accuse me of being a computer!
Judge: are you a computer?
Hex: Syntax Error in line 10. Core dumped. Please bugger off.
Judge: lol
Hex: I'm sorry, but I refuse to reply if you're going to be silly.
Inianes, again...
User: DO YOU ACCEPT JESUS CHRIST AS YOUR PERSONAL LORD AND SAVIOR?
Iniaes: Sure, I can accept all forms of payment.
User: HOW DO I WRITE MY OWN BOT?
Iniaes: With a pen or pencil.
User: THAT IS NOT FUNNY.
Iniaes: It seemed pretty funny to me.
Beware: In C++, your friends can see your privates!
Bottle?
:)
You know. As in kline. haha. I crack me up.
it sounds like maybe the guy isn't in Minnesota. Go fuck yourself.
-- 'The' Lord and Master Bitman On High, Master Of All
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".
these guys aren't doing anything but trying to look smart.
First, you may want to read Antonio Damsio's book, "Descartes' Error." The basic reason is that you don't want to make the same error that Descartes made. It's the same error that Minsky has made and always will make. That error is the strong mind/body distinction.
Obviously, the mind, being the same as the brain, is physically part of the body. What Descartes and subsequent philosophers, including AI researchers have falsely believed, is that the operation of the mind can be totally separated from the operation of the body. It can be somewhat separated, but not totally. As Damasio observes, the brain is deeply tied into the physical body and especially to the sensory organs.
What AI has not yet accomplished can be simplified to one task: learning. You can't get a legal AI to learn--on its own--enough about medicine to be a good medical AI. The knowledge that current "AI" gets is programmed into it.
The solution is to get an AI to learn. To get an AI to learn, you have to get the AI hooked up to exterior sensory devices, like a camera, a stand-in for an eye. A full array of sensors to mimic human senses would be nice, but for now, let's start with the eye. Once you've got the thing to see, you have to get it to recognize objects. If you can teach an AI to read, for example, that is huge. Don't give the AI OCR software. Instead, give it software that enables it to figure out how to do its own optical character recognition.
My $0.02: The notion of intelligence is hard to pin down, but it all seems to be just a matter of degree. A five year old child may be intelligent, but a two day old infant definitely is not. A college student can be said to be more intelligent than a kid in preschool, but in what sense? Simple - in the sense that matters to whoever is making the evaluation of intelligence. A software system that can recognize faces 90% of the time is clearly intelligent to anybody that cares that it can even do such a thing. A dog that can do the exact same thing is equally intelligent, only naturally instead of artificially. Sentience, on the other hand, is a purely subjective matter and is entirely separate from the notion of self-awareness. We know we are sentient because every other human is (a circular argument, I know). Nobody gives a rat's ass if anyone else is self-aware. A high-school dropout is no less sentient in the human sense than a Nobel laureate, though it is clear that the laureate is more intelligent in the academic and intellectual sense. What we consider sentience would probably be laughable to a higher-order being in the same way we would consider the sentience of a cockroach to be nothing of the sort. Yet cockroaches do act intelligently. They will find and eat your food and generally carry on with a semblance of purpose while the foul up your kitchen. The goal of current artificial intelligence research seems to be to create intelligent systems that credibly mimic human-level sentience. Currently, graduate students are busy building intelligent systems that mimic cockroach-level sentience - maybe not even that because I'm yet to see a robot that's behaviorally indistinguishable from a cockroach (let alone a dog). Star Trek's Data (the Holy Grail of AI) is considered sentient, though he is clearly an intelligent system that credibly mimics human-level sentience. None of his friends on the ship cares whether he is truly sentient in the human sense or whether he is truly self-aware. All that matters is that he appears to be. Of course creating a Data-like intelligent system is incredibly difficult if at all possible, which is why Star Trek is set several centuries in the future. Rodenberry was careful not to repeat Clarke's serious gaffe of putting the promise of advanced god-like technology (i.e. HAL in '2001:A Space Odyssey') within reach of the dreamers who actually believe such stuff will be possible in their own lifetimes. We already have systems that are artificially intelligent ('weak AI')in the ways that matter to those create, use and play with them e.g. speech recognition software that makes computers accessible to those without the use of their hands. Artificially sentient systems ('strong AI') like Data are still way, way off in the future. So let's get crackin'.
HelpUsObi 1
Brooks might argue that there's no such thing as disembodied intelligence, or at least creating an intelligent agent is easier if it has a physical manifestation. (I know he's backed a little away from the extreme position. It's hard to prove a negative.)
Of course, there have been some great successes using AI for network routing and fraud detection. Note however that these software agents are fully engaged in an abstract world where they are, in a manner, "embodied". They make no assertions about the physical world, only the world they exist in.
Happily, you may disagree with this position. Probably you could talk with more authority, citing more research, than I could. But, asserting that robotics is concerned with the attachment of sensors and rollers to a piece of machinery with/guided by a computer does not make your case, which I take to be that better algorithms is the key to AI.
A roboticist would argue that an agent cannot have common sense outside of the domain it exists in. If your goal is to have an agent with common sense regarding a rugged, dynamic, unpredictable, physical environment, then your best bet is to build a mobile machine with a variety of sensors, ie a robot.
Comment removed based on user account deletion
I think Marvin Minsky is an idiot too, but you took that subject line.
AI needs to have both a mind and a body before it can have behavior that can be recognized as intelligent. Subsumption architecture and old fashioned symbolic logic/expert system AI techniques need to be combined. Another problem is that while microprocessor hardware is optimized to execute serial instruction streams, neurons are a distributed (kind of parallel) execution system. The hardware we wish to emulate does not map elegantly to the hardware we use to compute, thus the problems with using simulators.
Some emphasize the mind or soul as the most important thing, others the body. Both are needed. If currently the AI field is making real progress with those "damn little robot insects" then that is good, someday the pendulum will swing back and what has been learned will find broader applications in other aspects of AI, pushing the whole field forward.
Patience, patience...
"To excuse such an atrocity by blaming U.S. government policies is to deny the basic idea of all morality: that individu
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 founder of the MIT A.I. lab was bitten to death this morning by a horde of angry Furbies. They were furious reading Marvin's quote "AI grad students waste three years of their lives soldering and repairing robots."
That's all I can say. Mr. Minsky was one of the big guys that lured me into the field, I'm disappointed he lost his vision.
Gentlemen, you can't fight in here, this is the War Room!
AI will probably never be used extensively by robots. At least not until well after we've got it running networks of computers (replace your sysadmin with an AI? Sure!) If an AI is truly intelligent, it should be able to learn new tasks no problem. Throwing this learning software onto robotic hardware really does nothing, the AI doesn't see the "optic sensors" as "eyes" but rather another data set to analyze and make decisions on. These data sets are totally irrelevant, they could be visual, auditory, or TCP/IP. It's all the same to a machine.
... because who cares about the measure of progress? Is there a deadline we're working towards? Has there ever been a more interesting time for those of us with a curious nature? Enjoy the ride.
BTW, scan the posts here. The definition of "AI" is as varied and elusive as that of "life". So how do we measure progress?
Oh and Minsky is kind of a real nasty dude when it comes to criticizing people. Don't forget he killed neural network funding for a decade with his book and personal vendetta on Rosenblatt. I guess he just likes AI done _his_ way.
"Consensus" in science is _always_ a political construct.
Java Settlers of Catan
http://settlers.cs.northwestern.edu/
The robots are able to build roads and cities!
-Mike
I'm currently working on a project in scene reconstruction, and I can't say how annoyed I was to discover that the only freely available robotics simulators I could find were of the type that simulate a 2-D robot in a 2-D world. Nothing that attempts to even simulate 3 dimensions, let alone physics or the ability to design all sorts of robots.
Nothing promising in the commercial sector, either.
I realize that creating software that could provide a good simulation that is usable in many situations is a difficult undertaking, but in the long run it would be a hell of a lot chaper for the universities, and would probably lead to a lot better work coming out of the graduate schools.
> 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.
To which all I can say is "so what?"
Alchemy - now called chemistry - had as its original goal the transmutation of lead into gold. That line of research is now completely dead, as chemists have discovered that there are many other lines that are vastly more interesting and more useful.
So it is with AI, at least IMHO. Some of the early researchers wanted to create human-level intelligences. They failed. Since then, the field has expanded and matured greatly, and the majority of researchers have decided that the original goal - human-level intelligence - is, much like transmuting lead into gold, not really all that interesting after all. So they work on interesting things, get impressive results, and ignore the founders' cries of "what about the goooooold??"
AI is not about creating machines with the intelligence of men. AI is about extending the scope of what machines can accomplish. Anyone who wishes all researchers were working on his pet project is bound to be disappointed.
Ever seen The Matrix? We can do the same thing with our AI research software - let it live in a simulated world. It would not be left with no senses and no method of communication, it just wouldn't be aware of the outside world, per se.
And no, AI is not pure math, but the amount and level of math used in many AI applications (such as vision) is pretty high compared to many other CS tasks.
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 !
No offense, but trying to cram all of creation into a system of predicate logic is _NOT_ the big picture, nor has it proven useful for creating a system that can learn autonomously.
Since Minsky's prime, AI researchers have realized that they can't replicate in a few decades what it took life several million years (or God seven days) to produce, so they've gone back to starting small. Yeah, it's a bit harder to start making extravagant claims about things when you study your field one step at a time, but just as behaviorist psychology has taught us more about how we thnk than Freud, Jung, etc. ever did, I think the current approach is what is going to produce results.
So let them start with projects like Cog that try to get a machine to figure out how to move its hand toward an object. From what I can tell, the work is coming along quite nicely (although I do agree that it would be nice if this stuff were being done in simulations rather than fooling around with soldering irons all the time).
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"
Pure research is a luxury(although necessary in the long run). In financially tough times, research facilites focus on marketable technologies. Instead of persuing a thinking machine, MIT is persuing projects for the military. Details here Why? Money, plain and simple.
Ask anyone who worked for Bell labs in the late 70's and 80's. These types of environments fostered incredible research that didn't really turn into real products for 20 years. Today the bean counters yell and managment focuses on marketable products. It's just a sign of the times.
-ted
This makes perfect sense when you think about it. Give the guy some credit!
Whoa. College students doing something fun. What a surprise!
In theory a MLP with temporal feedback might be able to model real neurons, representing spike train frequencies with levels instead ... but are there really methods of constructing such MLPs to model small nets of real (spiking) neurons which perform well?
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.
Nowhere goes AI
Surak is right in saying that switches are a bad way of simulating AI. When you look at people and think you recognize intelligence, and all of a sudden you see this new side of them that you did not know, is it really that that person's programming changed, or is it possible that JUST MAYBE a larger set of the BILLIONS of neurons that work independently are bending toward a different behavior?
..." when you have used it after a big mess up.
'If' tests try to assume very ISOLATED conditions and their expected repercussions, but I have another view. My view is that our gazillion neural switches are fighting for control and you never know which dominant set of those millions of millions of randomlike behaviors will assert themselves next and influence you to your very next action. We have some control of our actions but you can imagine countless little consciousnesses manipulating what you only perceive as one. Just think of the sincerity of the phrase, "I'm sorry, I didn't mean to
It's not "one [brain] produces many" --it's "many [impulses] produce 'one'."
Thanks for your time.
"Wireless : LAN
It's no wonder AI is going nowhere if they can't even simulate a real environment - let alone get computers to understand it.
Female Prison Rape in NY
> 'The' Lord and Master Bitman On High, Master Of All
And you are talking with a sig like that ?
Get a life.
DO NOT PANIC
That is the modern creed.
Of course, this is why Socrates ripped on the Sophists, and was the subject of many of Oscar Wilde's works... Even good old Jane Austin's Pride and Prejudice.
I guess this has been a problem for some time.
I don't read or respond to AC posts
you need to have a machine that can learn from the environmnet and can accept facts. human inteligence is based on our life experiences in our environmnet and from schooling. people are not inherently inteligent. a person who is totaly isolated from anything for their entire life except for food and water will not have any higher thinking skills and will either be catatonic or animal like.
a computer is incapable of replication animal insticts but is fully capable of learning our social norms and higher thinking.
to replicate human creativity you just program the machine to be inquisative and to put ideas to gether and test to see if the ideas are valid much the same way a child does in the todler and young childhood years. then it can draw on what it learns from these ideas and the responces it gets from the maintainer and when the machine matures it can then make more complex and valid connections of ideas allowing it to become a creative thinker.
it amaises me that people can not get his concept.
I am the Alpha and the Omega-3
Of course that's 100 billion neurons, not millon. Oops.
Our graduate library and information studies programs like that at the Massachusetts Simmons College have failed to include alliances with the artificial intelligence and related academic communities. Here's an example of how backward are our North American cities' public libraries...
http://www.cambridgema.gov/~CPL/mlnmove.html
http://www.mln.lib.ma.us/newcatalog.htm
If it weren't for AI, we wouldn't have things such as biometric scanners, voice recognition, 'Grafiti', Aibo, etc... Though it may seem that it hasn't gone very far since the MIT AI lab was created, it in fact has become mainstream.
Seriously, no one cares much what an old scientist says about AI - tell us about the prolog-hackers instead.
i fall asleep beside a lushious irish girl who my love was declared for...when i awake she has vanished along with all my money and possessions travelled with...realistically stranded...the world upon waking changes drastically from the world fell asleep in... are you trying to say all these changes happened in a time frame of effectively 0? i mean this woman may be fast but i cant see her travelling faster than light...and with no trace of observable cosmic irregularities...what can one realistically conclude...that the looting and leaving took place over 6 hours of drug assisted unconsiousness or that reality just snapped and went strange for awhile?
i think however that i can say with certianty the state that is consioussness has the quality that i am currently partaking in it. so in fact you are incorrect.
in the meanwhile i find that there is a built in clock that tells me a) what time it is and b) whether or not, despite what i remember, i was in any past interval consious. id even venture to say both these things are landmarks on the road to reason in the first place...
GENERATION 26: The first time you see this, copy it into your sig on any forum and add 1 to the generation.
if id ever said that this would be a good idea...depending whether the internet is self aware yet or not... anyways i suppose it is not that far of a leap... just a reversal of sorts...
GENERATION 26: The first time you see this, copy it into your sig on any forum and add 1 to the generation.
What we need is a standards body to define an interface between controller and robot. Then AI rearchers could develop faster behind a virtual robot. Hardware people could build better robots to stimulate and be driven by any implementation of the controller. It could even generate a market.
Sounds pretty interesting, do you have a URL to any more information?
Not saying there's no biology in intelligence, but assuming that biologists alone can answer the big questions about intelligence seems like an even bigger presumption. Psychologists and theologists are probably a better bet.
And you want me to "DO NOT PANIC", too. Seriously, what the fuck? Here's a hint for conversation: respond to what the person says, not their name. IT MAKES YOU LOOK SMARTER.
-- 'The' Lord and Master Bitman On High, Master Of All
Chances are, if we saw it, we wouldn't even know.
JMO, maybe someone will make a brilliant discovery tomorrow...
Sticking feathers up your butt does not make you a chicken - Tyler Durden
I don't know why you assumed the robots were ready. The chassis was, but the work on the electronics/low-level drivers was not.
So if you think all software components can be developed in parallel think again. Ever heard of critical path?
My part of the task as an individual was pushed as far as possible without having the robots ready. I completed the base-station GUI and implemented the communication protocol at both ends (base-station and robot, taking some of the burden of the electronics guy) in parallel with the other guys, since I could use the radios decoupled from the robots, using a spare HandyBoard (the chosen controller), but this could only be pushed so far.
Any further and I'd be creating a simulator. I didn't want a simulator then, that's why I was investing in the robots.
We agreed on the communication protocol, but we never got as far as reaching the milestone where we could test it, by letting the robots loose, and have them communicate their progress (using dead-reckoning) to the base. Too many hardware glitches and slow low-level software progress (developed by the electronics guy) prevented this from happening in time. This is not a criticism of the electronics guy, he did well with the time available, it is rather a criticism of the task itself. The hardware (and low-level software) was a time-sink, that prevented us from reaching a working system.
It was on the actual day of the competition that the robot caught up with my development, and we could get the little red dots to move on the base-station's map, just to find that the compass (direction of motion) readings were not being properly converted before being sent to the base station. We were doing angle calculations when the speaker announced it was our turn to compete. End.
Of course it's not the robot's fault. It was our fault for underestimating the amount of time (and in my case money) required to get the robot to a working state. This is why I agree with Minsky to a certain extent, in that people spend more time with the soldering iron, than working on the actual AI.
If any lesson was to be learnt, I think the most important is: if AI is what interests you, and time is a constraint, be prepared to invest in (expensive) ready made robotic platforms.
Sounds obvious now, but can be overlooked by newbies (such as undergrad or postgrad students).
Regarding my dissertation, I could have written about all this and got my degree (the other 2 guys did), but such a degree was of no value to me. I didn't go to Edinburgh to get a degree, I went there to learn about AI.
http://www.cs.cornell.edu/boom/1998sp/projects.htm
It's the "lamprey brain" one. (whoops, confused "lamprey" with "nematode"). Unfortunately, the destination of the link is no longer there. maybe google has a cache?
And I can't believe it's been 5 years already since I saw that! Frickin' time flies when you're in a cubicle...
and im talking about an even lower attribute...where one can be able to tell whether one was or was not consiouss at a given point in the past...so you don't beleive in this attribute of the mindP? i too count dreamstates as consioussness...but there are times of dreamless, unconsious states where the only way for me to have been aware during these times is as you say if i was aware but cannot remember...im uncertian how to succeed here... how about this... do you agree that nerves can be appended to the body...whether meat like leg nerves attatched to a wounded arm or mechanichal like the laptop-optic stuff they are doing? would you then say that there was feeling in these nerves before after but not during any time when they may be disconnected? or did that make sense?
GENERATION 26: The first time you see this, copy it into your sig on any forum and add 1 to the generation.
I see, but you could be mistaken.