Sign in to follow this  
lomateron

human intelligence

Recommended Posts

lomateron    491

Does anyone here thinks human intelligence is overrated and that it's just a matter of months until someone finds the right algorithm and just with a intel i7 and some GBytes of memory we can surpass human intelligence after running the algorithm for some months?

 

prove me wrong

Share this post


Link to post
Share on other sites
LennyLen    5715


prove me wrong

 

How can we? You haven't told us which side of the matter you're on.

 

But I seriously doubt that we will have human-comparable intelligence programmed on today's hardware in a matter of months.

Share this post


Link to post
Share on other sites
krinosx    606

Does anyone here thinks human intelligence is overrated and that it's just a matter of months until someone finds the right algorithm and just with a intel i7 and some GBytes of memory we can surpass human intelligence after running the algorithm for some months?

 

prove me wrong

 

 

And... why don't you prove you are right?

 

How deep is your knowledge of A.I.? 

 

Where are your arguments to prove you are right?? 

 

Its not so simple to deal with Machine Learning and its not so simple to machine and actual public A.I. techniques to simulate human behavior even human intelligence....

 

There is a little thing called 'Context Dependency' that is not so simple to teach a computer to deal with...

 

 

So... we are awaiting for your arguments proving your 'thesis'....

Edited by krinosx

Share this post


Link to post
Share on other sites
Tutorial Doctor    2573

Nope. Human intelligence is far superior than what any human will ever be able to replicate. I just realized that when thinking I can program artificial intelligence that behaves like humans. 

 

I realized it when I was trying to do a task at work. I had to reorganize a shelf, but a lot of factors had to be noticed first. I had to infer some things, and I had to suppose. 

 

The thing is, I could readjust my logic on the fly when new, unexpected things occurred. My sensory mechanisms interact with my logical operator to create a new plan. 

 

So, what one would have to do is make a program that can debug itself without any help from a human. And then one would have to make that same program able to edit it's own program to make itself more efficient. 

 

Not happening on even the latest processor 100 years from now. 

 

As I mentioned in another post, try to find a computer that can calculate 1,000,000! 

 

One more thing, that same program would have to be able to write a program to solve any problem you gave it, when your instructions are written in broken english with arabic as your first language. 

 

Computers try to be perfect, but it is the imperfection of humans that gives us an advantage. 

Edited by Tutorial Doctor

Share this post


Link to post
Share on other sites
alvaro    21246

I don't find anything compelling in Tutorial Doctor's arguments. My own point of view is that computers are already better than humans at many tasks: arithmetic, finding primitives of functions (a.k.a. indefinite integrals), playing many board games like chess or checkers, making investment decisions, playing Jeopardy, etc. The list will just expand over time until there is nothing a human can do better than a computer. I don't know if "replicating" human intelligence is relevant at all, once computers are better at everything. In any case, it won't be a matter of months.

 

Ideas about code that can debug itself, or modify itself or similar things were very prominent in early attempts at AI (that's why Lisp was popular in the field). Those attempts failed miserably, probably because that line of thinking is completely misguided. Your introspection about how you reorganize a shelf is just not very insightful, that's all. You certainly didn't have to perform brain surgery on yourself to be able to finish the task, which is the analogous situation to code that debugs itself.

 

The central problem of AI is how to make decisions. A solution to this problem has existed for at least 50 years: The best decision is one that maximizes the expected value of a utility function. The devil is in the details, of course.

 

There have been several areas where AI has been successful recently: machine translation, visual object classification, self-driving cars... Most of those areas have seen enormous advances thanks to the availability of large amounts of data and the ability to process it using machine learning.

 

[EDIT: By the way, 1,000,000! only has about 5.5 million digits. I am sure it's not that hard to compute, but I don't know what point you were trying to make: Do you know of any human that can compute it?]

Edited by Álvaro

Share this post


Link to post
Share on other sites

Some of the most advanced AI created by humankind is able to drive an automated motor vehicle according to human commands coming from a different planet at the amazing speed of almost 90 meters per hour, all without causing the 6-wheeled vehicle to topple or crash into larger obstacles. That's more than twice the maximum cruise speed of Helix aspersa.

 

The same AI is further able to abstain from moving for nearly one month without human intervention in case of a connection loss, keeping the vehicle reliably in one place (and in one piece).

 

So yeah, AI is impressive, but some humans will argue that they can do comparably. It may still take a few weeks before AI considerably outperforms the human.

 

By the way, 1,000,000! only has about 5.5 million digits. I am sure it's not that hard to compute, but I don't know what point you were trying to make: Do you know of any human that can compute it?

That's actually a good example for showing intelligence, though. A computer will certainly beat a human (or at least most humans) at computing numbers or at playing chess (which, too, is only computing numbers and maximizing a utility function).

 

A truly intelligent computer playing chess would offer a tough challenge, but would let the human win (and it would figure this out on its own!) because humans who keep losing against a computer program are likely to delete the program or smash the computer to pieces.

 

On the other hand, being told to compute 1,000,000!, a computer (assuming the human who built the program thought of including an arbitrary precision library!) would likely start working, and eventually, after a long time, find a solution or abort with an error.

A human would tell you right away: Wait, what? Fuck you, this is totally pointless! It's a really big number, OK? Do it yourself if you really want to know the exact value.

 

It's the difference between what you can do and what you will do because you immediately realize that it doesn't make any sense (or the insight that any 10k-digit number or any 1-million-digit number is exactly identical to any 5-million-digit number for all practical purposes).

 

What you will do or won't do doesn't even necessarily have to be logical at first sight (or at second sight, or at all) to be "intelligent".

Edited by samoth

Share this post


Link to post
Share on other sites
Tutorial Doctor    2573

Computers are not better than humans. Faster, yes, but they cannot, by themselves solve complex problems unknown to them. They cannot infer, nor adjust their code for completely new situations.

As I said, a computer would have to be able to upgrade it's information without help of a human, and it would have to be able to debug itself if it were to even be close to human intelligence.

It would have to be 100% autonomous and have 100% self automaton to start.

I am surprised anyone would even begin to compare computers to human intelligence.

Edited by Tutorial Doctor

Share this post


Link to post
Share on other sites
Tutorial Doctor    2573

As for chess, I beat the hardest level on Chessmaster simply by playing an imperfect game. All they do is load all this information about all the different patterns of opening game, middle game, and end game, and have the computer sort through them to play the most optimal move possible. But I haven't that knowledge, and I am not restrained to playing the most optimal move. So I played a move that wasn't expected, that wasn't "supposed" to be played. The game froze for a sec, and made a terrible move. I then took the computer's queen. Then it froze indefinitely. 

 

Humans don't have glitches like that. And if we do get hung on something, we can go research it and use that information to upgrade our knowledge. 

 

I'd also say that a human could calculate 1,000,000! before they can build a computer to do be able to do it without imploding. 

 

Now, 1000! has around half a million digits. 

Share this post


Link to post
Share on other sites
Nathan2222_old    395

As for chess, I beat the hardest level on Chessmaster simply by playing an imperfect game. All they do is load all this information about all the different patterns of opening game, middle game, and end game, and have the computer sort through them to play the most optimal move possible. But I haven't that knowledge, and I am not restrained to playing the most optimal move. So I played a move that wasn't expected, that wasn't "supposed" to be played. The game froze for a sec, and made a terrible move. I then took the computer's queen. Then it froze indefinitely. 
 
Humans don't have glitches like that. And if we do get hung on something, we can go research it and use that information to upgrade our knowledge. 
 
I'd also say that a human could calculate 1,000,000! before they can build a computer to do be able to do it without imploding. 
 
Now, 1000! has around half a million digits.

And besides, a computer without a human is really a stupid, unintelligent piece of hardware.

What would your pc, smartphone be without a human programming it to do what it does? Nothing but a bunch of materials joined together with no function.
It's a human that makes it intelligent.

Share this post


Link to post
Share on other sites

Agree. If I wanted to calculate 1000000! I wouldn't ask a mathematician to start sharpening his pencil, I'd write a library and I'd still beat the human.

 

EDIT: Even if the library was inefficient.

Edited by Paradigm Shifter

Share this post


Link to post
Share on other sites
Nypyren    12062
It seems to me that humans and computers currently are good at what the other one's bad at:

Vision/Hearing (humans MUCH better than computers in all aspects):
- Humans have a massive parallel signal processing network which recognizes an astounding amount of different kinds of features in visual/audio input VERY quickly.
- Computers have a tough time with this because we don't have hardware that's as good as human eyes/ears yet, parallelism that's massive enough to process the signal quickly, or the wide variety of different types of feature recognizers that humans have.

Control (robots are better in controlled situations, humans are better in novel situations):
- Robots are better for speed and precision.
- Humans are better for adaptation (for instance, if a servo/muscle stop functioning, robots often cannot perform their task anymore, whereas a human can probably figure out an alternate way to perform the task)

Natural languages (neither human nor computer is very good at this):
- NLP systems have thousands of individual systems all working together to handle typos, solve ambiguities, reason about likely meanings, and learn new meanings on the fly. Humans have lots of trouble getting their intended meaning across to other people. Computers have trouble dealing with imperfect grammar, meanings, ambiguity, error correction, and learning.

Logic and reasoning:
- Computers quickly follow the rules they're given, and have problems when the rules they've been provided aren't sufficient. Humans are slow at following rules but can adapt to cases they haven't seen before. Computers won't make mistakes. Humans make lots of mistakes.

Arithmetic:
- Computers are much better in all aspects.

System Interop:
- AI research has typically been approached one isolated piece at a time (vision, NLP, control, logic/reasoning, planning). Lots of complex AI problems need to be solved by having the different systems help each other out, but generally AI researchers focus only on their individual problems without seeking to interop with other fields. For example, when attempting to handle NLP input, you won't get very far without a logic/reasoning system to help resolve ambiguities and likely meanings. When handling vision, recognizing characters from a language requires dealing with different orientations/mirroring/perspective changes, stylistic variations, and reasoning about what a heavily corrupted glyph probably is based on the other glyphs around it. Edited by Nypyren

Share this post


Link to post
Share on other sites
Nathan2222_old    395

Agree. If I wanted to calculate 1000000! I wouldn't ask a mathematician to start sharpening his pencil, I'd write a library and I'd still beat the human.
 
EDIT: Even if the library was inefficient.

But you'ld have to write it. Try telling your computer to write the program without your help then try telling a person to write it.

Share this post


Link to post
Share on other sites

(long comparison of different fields)

 

Nice writeup. In summary, I think one must admit that while computers do have some fields where they clearly beat the human (in some fields in a stellar manner!), they still have large deficiencies at what one would commonly call "intelligence". Finding the shortest route is not intelligence, nor is minimizing a graph in general.

It's about doing things correctly (or mostly correctly) that you haven't been taught and which you've not been given a set of rules for, and it's about combining tasks and abilities in a sensible manner (again, without someone telling you).

 

For example, I could tell a human "Get me the book about that wizard kid from my room, please", and a human would almost certainly come back with the correct book. Failure rate pretty much 0%.

 

A computer might find the correct route to the room (or it might not know that my room is upstairs, or it might not be programmed to walk stairs) and it might not stumble on the stairs, might not bump into a closed door, and it might successfully identify the shapes of two dozen books in the room with an error rate of only 2%. It might even do OCR to read the book titles, but it will almost certainly fail to bring the correct book back anyway.

 

If that isn't enough, imagine I'm just reading a book by Tom Clancy, and I tell my favorite human to get me my Grisham book. And imagine that I'm saying: "Can you get me oh fuck it's snowing again my Grisham book?".

She will know that I don't mean to have intercourse, and she will know that "it's snowing again" isn't the book's title, and she will know that I mean Clancy, not Grisham.

 

A computer might know that I am reading Clancy, but it would still go for a Grisham book because that is what I ask for (or, it might reject the request because my syntax is wrong and won't go at all). Or, the computer might simply answer: "Yes I can".

Edited by samoth

Share this post


Link to post
Share on other sites
Waterlimon    4398

I just scrolled through this thread and, after seeing the amount of derpage in it, I think the OP may be on to something. Perhaps computers can equal the intelligence of SOME people in a few years, but certainly not MOST people.

But computers only do what you program them to do so they cant possibly be more intelligent than the programmer.

 

And you cant program consciousness because it arises from quantum mechanics.

 

QED

 

no im not serious ;_;

Edited by Waterlimon

Share this post


Link to post
Share on other sites
Adaline    710

I'm afraid this topic is a complete nonsense.

Computers are finite-state machines, human brain is not.

Comparing them makes no sense ... sad.png

 

And it will probably end in a sterile flame war in which I don't want to participate biggrin.png

 

The power of the (human) brain is that it can optimize the decisions, like a muscle can be stronger.

The power of a computer is that it can compute faster and without the factor "error".

 

"In some mounths" .... serioulsy ? Simulating a neural population that represents a single neo-cortical column needs a super-computer ... blink.png

 

They are so different machines !

Edited by Tournicoti

Share this post


Link to post
Share on other sites

 

(1) But computers only do what you program them to do so they cant possibly be more intelligent than the programmer.

(2) And you cant program consciousness because it arises from quantum mechanics.

QED

 

I think that is a bit too strict. According to (1) a child could not possibly ever become more intelligent than its parents and its teachers at school. At least the second part of that is demonstrably wrong. A lot of children become adults that are more intelligent (and more knowledgeable) than their school teachers. One of the big differences between a computer and a child is that eventually a child will decide to do something else than what it's told. Also, it will eventually realize that not everything it is told is necessarily true.

 

Though what you are saying is somewhat correct, nevertheless. Although things like fuzzy matching exist, computers are generally very inflexible in doing any such thing, or in doing meaningful things on their own, without an a-priori well-defined set of rules for every possible case. But I think it is this ability to take a decision (not simply by following a programmed ruleset, but really taking a decision) which is what distinguishes an intelligent being.

 

About (2), I daresay that nobobdy knows where consciousness comes from (or what exactly "makes" intelligence). Nobody even knows even remotely how our brain works.

Sure, we know there are synapses and stuff, and we kind of know how they work and how they react to certain chemicals, from a bird's eye perspective. We also have a somewhat rough overview of what regions in the brain are connected to what others, and what they most probably do. Or, we think that we know that -- what we really know is that there is electrical activity in certain regions when a person does certain things, and that non-lethal defects in certain regions tend to have certain failures (but with a huge variance and a huge regeative ability in using different regions). Everything else is just conclusions and theories, but not something that we truly know.

 

None of what we know happens in the brain can't be simulated. The only thing that is still forbidding is the immense complexity of the brain.

 

However, it it was really that simple, someone would have created a very simple artificial brain with a concience a long time ago. It needed not be the fastest, most brilliant brain in the world, as long as it has a concience. Maybe an artificial brain that has half the intelligence of a very stupid dog. It would even be enough of a sensation if such a brain placed into a robot with some unknown wiring learned that being shown a red triangle means it's going to be beaten in a moment (which gives a "negative" stimulus) and if it learned to walk away without someone telling it how to do that -- solely by observation and learning that sending certain impulses will move a "thing" in its body and create a certain stimulus, which corresponds to where it is in the world. Alas, no such thing has happened.

Share this post


Link to post
Share on other sites
Nypyren    12062

Simulating a neural population that represents a neo-cortical column needs a super-computer...


Saying we have to simulate neurons to simulate intelligence is like saying that we need to simulate what every transistor is doing in order to make a SNES emulator. We can do things as differently as we want, as long as the resulting behavior does what we want it to. It does not need to be identical in every way.

The people who are making neuron simulations are doing it to better analyze and understand biological brains, not because they want to make an AI by simulating neurons. They're really just doing reverse engineering at this point. Edited by Nypyren

Share this post


Link to post
Share on other sites
Adaline    710

Saying we have to simulate neurons to simulate intelligence is like saying that we need to simulate what every transistor is doing in order to make a SNES emulator. 

How to do reverse ingeneering if we can't say nowadays if neurons are communicationg via frequence of spikes or by delaies between spikes ?  

If Nature, through Evolution, is using neuron to manifest intelligence, it should be probably for a good reason. Even if we can't understand why yet.

 

The question is : What do you mean when you are talking about intelligence ?

Is it the ability to do something, or the ability to learn something ?

Edited by Tournicoti

Share this post


Link to post
Share on other sites
Nypyren    12062

How to do reverse ingeneering if we can't say nowadays if neurons are communicationg via frequence of spikes or by delaies between spikes ?


The purpose of reverse engineering is to find out how something's implemented. You can discover novel tricks and interesting deficiencies in systems this way. Another goal of making biological models is to identify new ways of treating people with hard-to-understand disorders.
 

If Nature, through Evolution, is using neuron to manifest intelligence, it should be probably for a good reason. Even if we can't understand why yet.


The reason things exist in evolution is because they work. Not necessarily because they are the best. Evolution works if the offspring of an organism survive. Drastic mutations that completely change the architecture of any organ are unlikely to survive. Neurons haven't been replaced with semiconducting transistors because there hasn't been an evolutionary path to do so - not because either method is superior.

If you think about evolution as if it were a software design process, the method is "randomly change a few lines of code and hope it outperforms its peers" rather than "OK, this system is turning out to be inefficient - rewrite the whole thing from scratch".

Although if you consider computers to be an indirect form of human offspring (because we create them), then computer development could be considered an unusual form of evolution...

 

The question is : What do you mean when you are talking about intelligence ?
Is it the ability to do something, or the ability to learn something ?


To me, "intelligence" just means "data processing". To other people, it means something different. It's an extremely overloaded term, which makes its actual meaning in a conversation vague. Since it's vague, each person is free to interpret it based on their own understanding.

Discussions about intelligence which don't burrow down into the more specific and clearly defined topics are doomed to make hasty generalizations and use overly simplified words that don't have a concise meaning. That's why in my first post, I focused on a few specific topics which could be discussed more clearly, and broken down further if the conversation wants to focus on any of them. Edited by Nypyren

Share this post


Link to post
Share on other sites
alvaro    21246

If Nature, through Evolution, is using neuron to manifest intelligence, it should be probably for a good reason. Even if we can't understand why yet.


Exactly. It's the same with legs, and this is why a vehicle that travels on wheels will never be successful. Nature knows you need legs for that.

[/sarcasm]

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this