Slashdot Mirror
A.I. Developer Challenges Pro-Human Bias
destinyland writes "After 13 years, the creator of the Noble Ape cognitive simulation says he's learned two things about artificial intelligence. 'Survival is a far better metric of intelligence than replicating human intelligence,' and "There are a number of examples of vastly more intelligent systems (in terms of survival) than human intelligence." Both Apple and Intel have used his simulation as a processor metric, but now Tom Barbalet argues its insights could be broadly applied to real life. His examples of durable non-human systems? The legal system, the health care system, and even the internet, where individual humans are simply the 'passive maintaining agents,' and the systems can't be conquered without a human onslaught that's several magnitudes larger."
People will fight tooth and nail against anyone or anything which challenges their notions of self-importance. We are just dirt that can talk.
Besides the conception, creation, and ongoing maintainance, very non-human.
The banking system is another example of a system much better than human intelligence for survival and resilience. Oh wait...
Survival is a terrible metric of intelligence. By that standard, lions and tigers and bears are the most intelligent species on the planet.
Check out my sci-fi/humor trilogy at PatriotsBooks.
By redefining intelligence to have nothing to do with what anybody means by intelligence, he can then claim that other systems exhibit more intelligence. Like a rock, presumably, since it survives far better than humans. I think this may be an example of somebody getting too interesting in specifics of tree-bark, and forgetting about the forest.
If you're going to approach the argument that way you have to consider survival of the species rather than the individual. On that metric human intelligence is clearly superior, as modern humans have been around for a few hundred thousand years, vs at most a few thousand for our most enduring created systems.
Obviously the solar system is the most intelligent of them all!
I for one, welcome our planetary overlords.
He essentially seems to be arguing that grey goo is the pinnacle of AI.
I much prefer the existing literature requiring that intelligence be an intelligence we can relate to as humans. Survivability is an interesting metric for creating more self-sustaining systems, but the goal of robotics should be fostering better knowledge and understanding of the universe. Searching for blind replication at the best rate possible just feels empty.
First of all, he doesn't actually say much about the survival as intelligence idea beyond the positing of the notion itself. It gives him a nice way to consider survival and intelligence as linked systems, with the "survival" of a system (that definition alone gets pretty abstract) being measured in terms of the logarithm of the number of humans required to shut it down.
/. article writers' need to get your attention, it suddenly becomes Human Intelligence Is Over.
He says you CAN consider the Internet, legal system, medical system, and others in terms of this notion, but doesn't get terrifically specific about it. He does, however, specifically state that road systems and the legal system are at least an order of magnitude more resilient than a human-level intelligence, which is nice, if you believe his examples are well-chosen. I'd be hard pressed to claim that they are.
In other words, he sets up an interesting research topic and then between his own poor choice of phrasing, the multiple Singularity references which surround the article, and the
It's sad when accomplished people feel that the only way to get into the limelight again is by radicalizing or overapplying their old ideas.
are better than mere Humans.
Survival is *beyond* good and evil, anyone?
In the free world the media isn't government run; the government is media run.
So the health and legal systems are intelligent? Whaaaaaa?
*head explodes*
So the Internet *is* Skynet
Wheee-oooooh!
Imma go kill me a few dolphins and assert my intelligence fellas by surviving longer than them.
God's been around a lot longer than anyone else. (Or anything, for that matter)
In God we trust, all others we virus scan.
Reading the article, it struck me as a good explanation of why AI is not getting anywhere.
By the authors criterion the solar system would take a huge number of people to shut down, and thus would be vastly more intelligent than any collection of surviving knives and forks used at AI conferences. I think that answers the other complaint of the author as well,
"There is a lack of scholarship in this area. This is, in large part, because most ideas about intelligence are deeply and fallaciously interconnected with an assumed understanding of human intelligence."
Oh well,
Stephan
http://stephan.sugarmotor.org
Shouldn't the tag read pcrunamok?
It would seem that he's referring to slow-changing "durable" systems as having better "survival" than individual humans. Anyone who's ever "fought the system" already knows that it takes an incredible amount of effort to cause even the slightest change unless you already have the authority to change the system arbitrarily (e.g. legislators can pass a bill.)
I think we've got another case of Slashdot story editors getting "creative" with the summary to attract readers. Now there's a system that's "resilient". :)
I do not fail; I succeed at finding out what does not work.
Creativity short of schizophrenia is a better metric of *human* intelligence than survival or logic or spacial recognition or any of the rest of the mess that AI researches try to measure intelligence with.
Moderating "-1, Disagree" is simple censorship. Have the guts to post your opinion.
The available documentation for Noble Ape is fairly shallow and opaque, it describes a simple scripting language, and some high leve discussion about space, time, and so on... but that's about it. Where's the AI? How exactly does the model simulate an ape, what's the relationship of the model to ApeScript? Where, in short, is the FAQ?
Well, then bacteria must be highly intelligent: not only do they have the greatest biomass and numbers on earth, they have almost certainly already traveled to other planets!
"Besides the conception, creation, and ongoing maintenance, very non-human."
Or, besides the human parts, which are almost everything, very non-human.
Not even fraud, just stupid.
Many people have commented that rocks, grey goo, the solar system are better at survival than humans! I could choose to use the rock for making something or simply destroy it. How long can it "resist" being used or being destroyed? If humans can master the planet's weather, even if it takes time, then yes they are more intelligent. If humans are able to colonize the solar system then yes, they are more intelligent than the solar system. Subconsciously survival *is* our definition of intelligence. If someone is cheated, he is called a fool and the cheater is supposed to be smarter. And if the cheated remains gullible then his chances of survival are less. Deception is a quality that is common to all humans. What we say is most of the times is different than what we think. Turing test is an accepted method for determining intelligent behavior. And what does it involve? Deception. The computer is supposed to cheat (and thus survive) the human by making him believe that its not really a computer but a human being.
Reading the summary I was left wondering if it was a budding AI that wrote it.
It's quite illegible as is, it must still be a bit low on the IQ scale...
nt;
Yeah, bitches are always smarter.
(Better post this one AC.)
Survival is not a good measure of intelligence, but maybe we are seeing intelligence in terms too human. Maybe more than survival what you should check is how it reacts and adapts to a new environment, to new things. In that sense, Law is definitely less intelligent than Internet, as is pretty slow and dumb adapting to the reality created by the existence of internet.
Could a bee hive or an ant colony be treated as a separate intelligent entity. Probably that could fit better in the intelligence concept than Law or even Internet. Human mobs are definitely dumber than any of those.
Better virus scan God, too.
He seems to think we should, really.
health care system is durable? clearly he isn't living in the real world...
If you mod me down, I will become more powerful than you can imagine....
The banks "persuaded" "us," didn't they?
Democracy Now! - uncensored, anti-establishment news
Osama is a freaking genius, and Lincoln was dumb!
Unsurprisingly, most of the people here haven't read, or perhaps not really absorbed, what TFA discusses, and are jumping to quick and irrelevant conclusions.
The author explains that Survival is a good metric of Intelligence, and he uses humans as an example. One human can definitely kill one lion, bear, mosquito, single bacteria, etc. if equipped with his intelligently designed tools such as a gun, or a mosquito zapper, antibacterial soap. He uses these tools, intelligently, to kill one bear, and hence, the human is more intelligent. However, if you take 10 bears, then sure, they may be able to kill the 1 human, but that means they are less intelligent, and take more numbers.
He simulates intelligence this way, and he defines a simulation as any environment with applied constraints, and that may include the internet, legal system, your neighbourhood community, etc.
So here's what he says: A system, such as the health care or legal system, will not be shutdown by one person. In fact, it probably won't even be shutdown by 10 people, maybe 100. And hence, the system is vastly more intelligent than a human, intrinsically since we worked in numbers to evolve this system.
I think it's a very interesting way of looking at intelligence. Again, this is all based Mr. Barbalet's assumptions.
By this definition a bacterium strain or virus would be considered the most intelligent thing on the planet.
One choice quote from the article:
The same reason you get the opinion "The primacy of human intelligence is one of the last and greatest myths of the anthropomorphic divide
Okay, human intelligence may be fuzzy and difficult to objectively measure. But that applies to many things such as CPU speed, Kolmogorov complexity, how complicated a shape is, or how much heat/sound insulation a particular material provides. Even how good a piece of music/art is.
They're tricky, but there's no doubt that exponentially low and high numbers can be given to each of those attributes.
Why OpalCalc is the best Windows calc
Intelligence here is a technical term that AI researchers use to describe the fitness of an agent in a simulation.
He's not talking about brain activity or as was mentioned 'the ability to retain and process information'. He's talking about a history of actions taken that created an agent with better fitness. It's not an attempt to get attention it's simply someone in a highly technical field using terms in ways that normal people don't.
One of those dialect problems that scientists often have when trying to describe their ideas.
http://www.google.com/search?q=god's+debris+online
Survival is an instinct inherent in all living things, and survival is not measured by one's own ability but by the predators around them. Very few species (if any save humans) will willingly destroy themselves or lie down to die when the opportunity to survive presents itself.
How do you use an ape-in-environment simulator to evaluate a microprocessor? You've got virtual apes running around on islands getting Hungry and Feared and Sexed.. how would you even introduce a processor design into that kind of simulation? Processors competing for electricity resources while getting Warmed and Turned Off?
Using an evolutionary algorithm to optimize certain highly complex design elements makes sense, but roaming around an environment interacting and competing for resources optimizes apes, not computers.
There's something to be said for focusing on low-level survival issues, but one can do more than pontificate about it. As someone who's worked on legged locomotion control, I've made the point that a big part of life is about not falling down and not bumping into stuff. Unless you've got that working, not much else is going to work. Once you've got that working well, the higher level stuff can be added on gradually. This is bottom-up AI, as opposed to top-down AI.
Bottom-up AI is now mainstream, but it was once a radical concept. I went through Stanford at the height of the expert systems boom in the mid-1980s, when abstraction was king and the logicians were running the AI programs. The "AI Winter" followed after they failed.
Rod Brooks made bottom-up AI semi-respectable, but he went off onto a purely reactive AI tangent, with little insect robots. That works, but it doesn't lead to more than insect-level AI. My comment on this was that it was a matter of the planning horizon for movement planning. Purely reactive systems have a planning horizon of zero. That works for insects, because they are small and light, and can just bang feelers into obstacles without harm.
As creatures get bigger and faster, they need a longer planning horizon. The minimum planning horizon for survival is your stopping distance. (This is explicit in autonomous vehicle work.) Bigger and faster animals need better motion planners. This is probably what drove the evolution of the cerebellum, which is most of the brain in the mammals below the primates.
I've had horses for many years; I was on horseback an hour ago. Horses are interesting in this respect because they're big, fast, have very good short-term motion planning, but do little long-term planning. Horse brains have a big cerebellum and a small cortex, which is consistent with horse behavior. This gives a sense of what to aim for in bottom-up AI; good motion control, good vision, defer work on the higher level stuff until we have the low-level stuff nailed.
That's happening. The DARPA Grand Challenge, especially the 2006 season with driving in traffic, forced some work on the beginnings of short term situational awareness. BigDog has some of the low-level motion control working really well, but BigDog isn't yet very good at picking footholds. They're just getting started on situational awareness. There's some good stuff going on in the game community, especially in programs that can play decent football. This problem is starting to crack.
Short-term planning in these areas revolves around making predictions about what's going to happen next. The ability to ask "what-if" questions about movement before trying them improves survival enormously. This kind of planning isn't combinatoric, like traditional AI planning systems. It's more like inverting a simulation to run it as a planner.
I have no idea how we get to "consciousness", but if we can get to horse-level AI, we're well into the mammal range. I encourage people to work on that problem. There's enough compute power to do this stuff now without beating head against wall on CPU time issues. There wasn't in the 1990s when I worked on this.
The metric used in TFA is:
Due to the rough nature of the approximation, I employed a base-10 logarithmic approach. If it took a human to slay the system, the survival intelligence value would be zero. If it took ten, the survival intelligence value would be one. If it took a hundred humans, the survival intelligence value would be two.
And it goes on to say "combining the metric of intelligence for survival and the idea that nearly anything is fair game for this metric, let's explore a couple of examples." Well, I plan to explore a couple of examples, but with the metric reversed. Why not concentrate on human survivability?
The new metric is: If it takes one of the system to slay one human, the survival intelligence value would be zero. (Lower is better in this case)
Now all of the so-called intelligent systems like road networks, the internet, and the legal system don't fare so well. You may be able to gather some large number of people to break the internet, cause a traffic jam, or subvert the legal system, but no matter how big the internet gets, how can the internet (treated as a system) kill a human?
Using the metric outlined in TFA puts emphasis on the comparison of physical qualities (lifespan, physical size/mass) than on any measure of real intelligence. Maybe one mountain can out-survive one human. Maybe 10,000 humans can destroy one mountain. But 10,000 mountains will never destroy one human.
Interestingly, survival in the most successful living organisms is attributed to neither intellect nor intelligence, but rather instinct.
...Intellect being the possession of lots of knowledge and intelligence being the proper or creative application of said knowledge.
Perhaps what this researcher should have said was "We as students of AI focus on intelligence, but this is only really an afterthought to developing an autonomous being the way we understand them. We should really be studying instinct."
There was a SIGNATURE here, but it's gone now.
Ironic how human intelligence, "the last and greatest myth(s) of the anthropomorphic divide" is the sole intelligence capable of classifying itself a "myth".
I think the author has strayed from the field of Artificial Intelligence into the morass of Artificial Ideas.
By definition, if it doesn't survive, no longer exists, then it didn't work, did it?
An excellent example of "artificial intelligence" is my pure-bred Golden Retriever. He's very, very smart. He's sweet, loving, and is amazingly responsive to voice tones, gestures, and the like. He knows exactly what I mean when I point, snap my fingers, even tilt my head towards the door. I could swear up and down that he understands what I say, many times, and definitely not because he always does what I want!
Yet, for all the human-ness about him, he's a dog. He barks, not talks, and has no fingers. But he comes from a long, long line of dogs, untold thousands of years in duration, that survived by better emulating human intelligence.
Why would/should we expect AI to be any different?
I have no problem with your religion until you decide it's reason to deprive others of the truth.
AI is intelligence built by, and by logical extension, for artifice. You have a goal that needs to be met, and the only way to meet it is a self-aware mechanism. So long as it meets that goal, it has achieved its purpose. Any existence beyond that goal is pointless and self-defeating.
Here's where your monkey meat kicks in and demands that there be a purpose to life above and beyond what you can know. In our case, as biological entities that have evolved over the course of a few billion years of advanced organic chemistry, that may be true. In the case of a self-programming program, it is more likely not to be true.
Take a simple concept like mortality. You die, the end. No more fun in monkey-meat land. The computer program ab-ends. So what? It was just a fork of a parent process that's still going, or a copy of the program with to-the-attosecond backups of its runstate going back a few years is in storage. Survival is an animate thing... a chemical soup thing.
OK, let's delve conceptually deeper. Let's say, for instance, a race of hyperintelligent, otherdimensional beings have evolved us, deliberately to solve math problems for them. We do this by being stuck to the planet by gravity, and walking around. We lift things up, and set them back down, and that's all that's expected of us by our "creators." Do we resent them for it? Or would it just be a weird side-tangent to what we consider our existence? Our interdimensional creators would neither know nor care that we have created furry pr0n, and our fursuiters wouldn't much care that the creators didn't care, and go on doing what they do.
So, now imagine a computer that programs itself for intelligence, and self-evolves to meet a certain goal, like predicting the sales of the next Madden football game. It makes the prediction and then ends. Does it care that it's dead? No. Death and non-existence are a biological obsession... this program =knows= what its life purpose is, and having completed it, would not much care if there's nothing after.
It's unlikely we'd program an AI simply to survive. There's no money in it. We'd program it for a purpose, and you know what? Resentment is a monkey-meat thing, too. Computers aren't involved in that shit, unless we program them to resent. As there's no money in it, it's unlikely.
Ah, Mentifex, a true old school internet crank.
Funny that you pick examples that wouldn't exist without the massive amount of 'passive maintaining agents' that maintain them. Humans don't have to 'attack them' to destroy them, they just have to stop using them or maintaining them. Seems kind of silly to call then 'non-human durable systems' when with out humans they cease to exist, or in the case of the Internet will break down relatively quickly.
Persistent Volume manager for Kubernetes - https://github.com/dwimsey/openshift-pvmanager
"The Invincible" by Stanislaw Lem directly addresses this topic. Astronauts land on a planet inhabited by evolved robot species. The dominant species manifests itself as a swarm of robotic flies. The astronauts cannot communicate with the flies, because although they are robotic and successful, they are still basically flies. I don't think this Barbalet character has really thought his ideas out properly; it seems absurd to define intelligence as "that trait that allows a system to survive a long time." The Turing test's definition is along the lines of "The characteristic of communicating like a human", which, although imperfect, is far more sensible.
A single back-hoe accident has kicked the internet in the ass. A few dozen human beings(the right ones) can cripple if not kill it.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
If we parse "survival" to mean something like "adaptation", then I think we're on to something. Sheer survival doesn't imply intelligence. As a gedankenexperiment, consider that the Loch Ness Monster really exists, and exists solely because it's living in a place (the deep waters of Loch Ness) where it doesn't have to adapt, because it's very seldom preyed on, or seen. Then, all of a sudden, an adventurer finds a way to track it. Game over. Now, if the Loch Ness Monster could figure out a way to hide or get away from that adventurer, or adapt in a new environemnt, I would think that's more akin to intelligence that sheer length of survival.
Sounds to me more like he's suddenly realised everyone is noticing the AI Emperor has no clothes and is panicking about future funding.
Professor Penrose is right: Artificial Intelligence is impossible because we are fundamentally incapable of understanding human intelligence. Our brains are not predictable, clockwork, mechanical devices they are at a fundamental depend on chaos.
Bad analogies are like waxing a monkey with a rainbow.
My impression is that this particular narcissist ape brain got bored of A.I. research and would rather be developing weapons, or reproducing...
"I love my job, but I hate talking to people like you" (Freddie Mercury)
Roads are pretty "intelligent" as well, by the above definition. They've infiltrated all of society (can you name a town where these don't appear?), and have made humans their maintenance-giving slaves! Now the roads of the world have started to demand lighter color, by storing and emitting heat from the sun, thus furthering the greenhouse effect (blackmail, a sure sign of intelligence). Roads have us by the balls! And just think how much work it would take for you to destroy all the roads in the world... and they would resist, plus send their human minions after you. Remember this the next time you speed down a highway: who is really using whom?! Hmmm....
Where is Hofstadter when we need him! Not running off a cliff does not make intelligence. Seeing the cliff as an obstacle and creating a solution is intelligence. All the alleged intelligent systems rely on humans for creativity.
You might be interested in the AI work of Douglas Hofstadter, as described in "Fluid Concepts & Creative Analogies". In that he describes projects that are meant to build creatve programs, that lead toward AI that can react to new situations without a human pointing it in exactly the right direction. In the last 10 years or so his research group seems to have stalled, but the program "Metacat" is available online if you have the patience to get it working. The theme of that work, creativity, is a key part of intelligence because even a routine task like recognizing objects involves fuzzy analogies to past situations that never perfectly match.
The idea about "AI that makes the same mistakes as us" is part of a split among AI researchers. Some seem to want to understand the brain so they can build better AI, while others want to build AIs so they can understand the brain. If the goal is to understand human brains, then simulating human stupidity means you're hopefully following the same processes that the brain uses. So artificial stupidity is useful, for that purpose. I'm more interested in seeing better AI, regardless of whether it ends up working like a living brain.
Revive the Constitution.
I, for one, welcome our pro-human biased overlords.
You know, there is a difference between trolling and pointing out the flaws in your reasoning. Just saying.
Nothing like twisting the language to force a point. We have different words for "survival" and "intelligence" because they are different things. Redefining one to mean the other does not contribute to the discussion. It may be that Artificial Survival is a better goal for research than Artificial Intelligence - the point could be argued. But this semantic redefinition assumes that argument won, and claims victory in an Orwellian manner by redefining the language to state that victory has been won.
Consciousness is an illusion caused by an excess of self consciousness.
Just look at us. We have different definitions of intelligence just for humans. "That person is book smart", "He has street smarts" and I am sure if I took a little bit of time I could come up with other examples. Even for our own intelligence we can be pretty stupid. Just check out the Darwin awards. If we judged intelligence by that we'd be in big trouble.
I read Slashdot for the headlines, because the headlines, unlike the articles, are usually original and never duplicated
Prions and viruses are nigh indestructible. Viruses are neither living nor dead. and do we even have a way to prevent prions other than destroying anything containing them?
Really, the whole discussion is a waste of time unless everyone is using the same definition. By the definition I would use, no computer is intelligent because the computer is not aware, any more than a piece of paper with some information written on it is aware of the information. Computers, their output, the tasks they perform only have meaning to the humans that operate them. Without that they are just processes and their actions are no more meaningful or relevant than a tree falling down in the forest that nobody sees.
H+ magazine Summer Just go to page 16, exploring the possibilities of ESP through a brain chip.
Think of how stupid the average person is, and realize half of them are stupider than that.
The reason no one has paid attention to this theory or area is because it is total bullshit, and he used a thin grasp of his Philosophy 101 to try and paint over the shit. It still smells like shit.
It is called "Intentionality" (and NOT 'I intend to go to the store' type Intentionality). I would say he is several hundred years too late to the party on it. For a quick appreciation of just what a hack he is try this:
http://plato.stanford.edu/entries/intentionality/
For those that do not want to dive in to all the ugly details to understand a big of the above, I can point out the problem in simpler form.
Did you notice how this guy is still defining the intelligence of his system using human intelligence? Well, there is a very good reason why we have to use human intelligence as the benchmark for AI.
Living in Chile
How about a game of Go instead?
That has to be the BEST definition of intelligence I've ever heard. I was beginning to think it was impossible to define, but, that nails it perfectly! I vote that from here on that be used as the metric of true intelligence.
Over-the-top Response Guy! Giving "Over-the-Top Responses" since 1970.
It will have its own metric.
In fact, the notion of species is an empirical one, not a real one. If we want to be really exact, we should say life contain only one species with many differents entities.
Then life is the most intelligent system.
Intelligence is also an empirical notion, there could not be any clear and unambiguous definition which correspond perfectly to what we all call "intelligence". Therefore, state "intelligence = robustness" seem a bit stupid to me. Just use the word 'robustness' (it already exists).
With this definition artists are DUMBEST people ever!
Fitness is an important characteristic of living entities, but it isn't the same thing as intelligence.
Hmmm survival is what promotes people, seems more like animals think that way. What about ambition. That is a huge trait. I don't see this as a breakthrough just a change in heuristics.
Sometimes I wonder whether all this research on A.I is going to evolve human minds faster than the singularity finally takes over.Perhaps the fear of singularity is motivating the intelligent portion of humanity to look at the human intelligence, to find its weaknesses, to compare it to other successful structures and to improve it! Yes, that is what I call evolution!
'Survival is a far easier metric of intelligence to measure than replication of human intelligence,'
If you can't find the answer to the original problem, then change the problem. Is that the philosophy?
The author is clearly on the forefront of a revolution in AI that will completely change our civilization.
Pray tell, have the Apes in his simulation invented fire yet?