Total Posts:16|Showing Posts:1-16
Jump to topic:

can human brain overtake chess engine?

Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/25/2016 4:58:07 AM
Posted: 5 months ago
Scientifically, is that possible? let us say modern chess engine, talk about chess engines of june 2016.

I read about The chess computer that beat Garry Kasparov 19 years ago was able to calculate 200 million positions per second. Now, almopst 20 years passed and technology advanced, I do not know about current chess engines. if you know, share.

can a human brain achieve a degree in which he can calculate as good, precise, fast as Komodo or stockfish? Theoretically or scientifically, is it possible?
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer
Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/25/2016 5:01:11 AM
Posted: 5 months ago
some of us may say "it is not, it is obvious" but our brains should not be as underrated, underestimated as that.

There are some people who speaks 40-50 languages. once A person from Brazil was claimed to speak 120 languages fluently. Interpreter of Europarliament speaks 32 languages.
There are some people who are as good calculator as a calculator machine. Considering all of these, even though it seems impossible I think a human brain can achieve chess engine level in chess.

but opinions are not enough, do any of us know scientific answer? Is there a border, limit that our brain can not pass?
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer
Cobalt
Posts: 991
Add as Friend
Challenge to a Debate
Send a Message
6/25/2016 5:29:50 AM
Posted: 5 months ago
At 6/25/2016 4:58:07 AM, Artur wrote:

Our brains have a limited amount of space to store information.

A chess algorithm works by checking all legal solutions some number of turns ahead. It then uses various standards to come up with a "score" and the best scored move will be played. The human brain currently can't a) store that much information with accurate recall and b) isn't fast enough to iterate through all possible moves in a reasonable amount of time, even if you're only looking a few moves ahead.

Will the human brain ever be capable of this? Probably not. Will the human race evolve into a species with a brain capable of doing this? If we live long enough, I imagine so.
keithprosser
Posts: 2,045
Add as Friend
Challenge to a Debate
Send a Message
6/25/2016 10:11:35 AM
Posted: 5 months ago
Will the human brain ever be capable of this? Probably not. Will the human race evolve into a species with a brain capable of doing this? If we live long enough, I imagine so

Now I would have imagined not. If the question had been about physical strength, do we expect to evolve into a race of muscle men? Unlikely - we use machines to do our hard physical work these days, and will continue to do so in an increasing way. Increasing weediness seems more likely, Similarly, there is no need for us to evolve super-brains if we have computers to do our thinking for us.

Evolution occurs as a reaction to external conditions, not progression towards any imaginary goal. If future conditions are such that we need super-computers in our brains just to survive then I don't think that is a direction we should go in.

Given that brains use inherently slow effects such as ions diffusing across synapses to work, brains are limited. Creating a super-fast brain out of neurones would be like trying to make a concorde using steam engines.

Enhanced brain power in the future - if that is desired - will most likely come about by prosthetic implants inside the skull, not by normal evolution by natural selection.
PetersSmith
Posts: 5,848
Add as Friend
Challenge to a Debate
Send a Message
6/25/2016 5:13:14 PM
Posted: 5 months ago
At 6/25/2016 4:58:07 AM, Artur wrote:
Scientifically, is that possible? let us say modern chess engine, talk about chess engines of june 2016.

I read about The chess computer that beat Garry Kasparov 19 years ago was able to calculate 200 million positions per second. Now, almopst 20 years passed and technology advanced, I do not know about current chess engines. if you know, share.

can a human brain achieve a degree in which he can calculate as good, precise, fast as Komodo or stockfish? Theoretically or scientifically, is it possible?

There is technically a "right way" to play because technically their are a finite number of things you can do to win. If the computer can process all ways no matter what way the human opponent moves, then I think the human brain can't beat a chess engine like that because it's all math.
Empress of DDO (also Poll and Forum "Maintenance" Moderator)

"The two most important days in your life is the day you were born, and the day you find out why."
~Mark Twain

"Wow"
-Doge

"Don't believe everything you read on the internet just because there's a picture with a quote next to it."
~Abraham Lincoln

Guide to the Polls Section: http://www.debate.org...
RuvDraba
Posts: 6,033
Add as Friend
Challenge to a Debate
Send a Message
6/25/2016 8:29:49 PM
Posted: 5 months ago
At 6/25/2016 5:01:11 AM, Artur wrote:
but opinions are not enough, do any of us know scientific answer? Is there a border, limit that our brain can not pass?

Neuroscience still doesn't know precisely how the brain's neurology produces thought, although we are able to create neurological simulations that do quite clever jobs. [https://cs.stanford.edu...] But the neurology has physical limits. For example, we know approximately how many neurons and synapses a human brain has (100 billion, and 100-1,000 trilion, respectively [https://en.wikipedia.org...]), and how fast they fire (up to 50 times per second.) So on the reasonable supposition that thought is produced by synapse-firing, we know how much activity that can produce, and that's an upper bound on how fast we can think, including all the brain-functions needed to keep us alive, aware of our environments and maintain our relationship and sense of place.

But how efficient our thought is depends on the methods we use. It's already possible to create a chess program that, in unbounded time, is 100% effective in its play (meaning it will always win or draw whenever that is possible.) What holds it back is time-efficiency: depending on the competition, you have to play fast enough to finish in (say) 2.5 hours. So if there are more efficient methods for finding the best next move, and if humans learn them but a program does not, then humans might outplay the program.

On the other hand, the people researching those methods tend to be computer scientists, so programs are finding new methods faster than human players are. :)
Rukado
Posts: 527
Add as Friend
Challenge to a Debate
Send a Message
6/25/2016 9:06:47 PM
Posted: 5 months ago
At 6/25/2016 5:01:11 AM, Artur wrote:
some of us may say "it is not, it is obvious" but our brains should not be as underrated, underestimated as that.

There are some people who speaks 40-50 languages. once A person from Brazil was claimed to speak 120 languages fluently. Interpreter of Europarliament speaks 32 languages.

Those are lies and exaggerations. I think the average human has no potential to speak more than three languages fluently, and that's only if living in a place where those three languages are used regularly. A professional language interpreter with a good brain might manage a dozen languages of about three language families.

Most adult immigrants to America will never speak English fluently, as a second language.

Human brains are very limited.

No human has any chance against a top chess computer, anymore.
dee-em
Posts: 6,481
Add as Friend
Challenge to a Debate
Send a Message
6/26/2016 12:36:15 AM
Posted: 5 months ago
At 6/25/2016 8:29:49 PM, RuvDraba wrote:
At 6/25/2016 5:01:11 AM, Artur wrote:
but opinions are not enough, do any of us know scientific answer? Is there a border, limit that our brain can not pass?

Neuroscience still doesn't know precisely how the brain's neurology produces thought, although we are able to create neurological simulations that do quite clever jobs. [https://cs.stanford.edu...] But the neurology has physical limits. For example, we know approximately how many neurons and synapses a human brain has (100 billion, and 100-1,000 trilion, respectively [https://en.wikipedia.org...]), and how fast they fire (up to 50 times per second.) So on the reasonable supposition that thought is produced by synapse-firing, we know how much activity that can produce, and that's an upper bound on how fast we can think, including all the brain-functions needed to keep us alive, aware of our environments and maintain our relationship and sense of place.

But how efficient our thought is depends on the methods we use. It's already possible to create a chess program that, in unbounded time, is 100% effective in its play (meaning it will always win or draw whenever that is possible.) What holds it back is time-efficiency: depending on the competition, you have to play fast enough to finish in (say) 2.5 hours. So if there are more efficient methods for finding the best next move, and if humans learn them but a program does not, then humans might outplay the program.

On the other hand, the people researching those methods tend to be computer scientists, so programs are finding new methods faster than human players are. :)

The sheer processing speed of computers has to factored in too. CPU chips are only going to get faster. Unfortunately the human brain is not. Even with current algorithms and strategies, chess-playing computer systems will improve just by putting them on the next generation of hardware.
Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/26/2016 5:30:58 AM
Posted: 5 months ago
At 6/25/2016 9:06:47 PM, Rukado wrote:
At 6/25/2016 5:01:11 AM, Artur wrote:
some of us may say "it is not, it is obvious" but our brains should not be as underrated, underestimated as that.

There are some people who speaks 40-50 languages. once A person from Brazil was claimed to speak 120 languages fluently. Interpreter of Europarliament speaks 32 languages.

Those are lies and exaggerations.
why do you think so?
I think the average human has no potential to speak more than three languages fluently, and that's only if living in a place where those three languages are used regularly.
then, I think you have to change your opinion. I myself am can more than that and I know many people, face many people in everyday life that speak more 5-6 languages fluently.
A professional language interpreter with a good brain might manage a dozen languages of about three language families.

Most adult immigrants to America will never speak English fluently, as a second language.

Human brains are very limited.
so, how far can we reach if our brains are limited? Where does that limit lie?
No human has any chance against a top chess computer, anymore.
at the moment, I too think so. even with pawn or move odds a person no chance against medium chess computers I think.
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer
Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/26/2016 5:39:19 AM
Posted: 5 months ago
At 6/25/2016 8:29:49 PM, RuvDraba wrote:
At 6/25/2016 5:01:11 AM, Artur wrote:
but opinions are not enough, do any of us know scientific answer? Is there a border, limit that our brain can not pass?

Neuroscience still doesn't know precisely how the brain's neurology produces thought, although we are able to create neurological simulations that do quite clever jobs. [https://cs.stanford.edu...] But the neurology has physical limits. For example, we know approximately how many neurons and synapses a human brain has (100 billion, and 100-1,000 trilion, respectively [https://en.wikipedia.org...]), and how fast they fire (up to 50 times per second.) So on the reasonable supposition that thought is produced by synapse-firing, we know how much activity that can produce, and that's an upper bound on how fast we can think, including all the brain-functions needed to keep us alive, aware of our environments and maintain our relationship and sense of place.

But how efficient our thought is depends on the methods we use. It's already possible to create a chess program that, in unbounded time, is 100% effective in its play (meaning it will always win or draw whenever that is possible.) What holds it back is time-efficiency: depending on the competition, you have to play fast enough to finish in (say) 2.5 hours. So if there are more efficient methods for finding the best next move, and if humans learn them but a program does not, then humans might outplay the program.

On the other hand, the people researching those methods tend to be computer scientists, so programs are finding new methods faster than human players are. :)
actually, the answer I am looking for is not whether we stand a chance against them.

do not get focused on chess only, I just used it as an example because in my opinion they are the most advanced artificial type in 'artificial area.'
what I wanna know is is it possible to be as precise, fast as them in mathematical, calculation problems? Let it be chess, or math or science. or let me give another example:

imagina a list that includes 1000 people and their data. if it is saved on computer, once you check for people who were born, say, on 16th of June, computer will immediately say how many people out of them were born on 16th of June. and imagine a person who can say it at once after reading, analyzing all people's data once. He reads that list, then you ask him "how many of them were from Japan" and he gives the precise, correct answer just in a second.

this too is an example.
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer
Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/26/2016 5:45:06 AM
Posted: 5 months ago
At 6/25/2016 5:29:50 AM, Cobalt wrote:
At 6/25/2016 4:58:07 AM, Artur wrote:

Our brains have a limited amount of space to store information.
so, what is the approximate amout of space our brain can store? cant we enlarge it?
A chess algorithm works by checking all legal solutions some number of turns ahead. It then uses various standards to come up with a "score" and the best scored move will be played. The human brain currently can't a) store that much information with accurate recall and b) isn't fast enough to iterate through all possible moves in a reasonable amount of time, even if you're only looking a few moves ahead.
may be, chess is not the only goal I am talking about.
My question is about being as fast, deeply, precise calculator as them. Let it be in chess, let it be in mathematical calculations or let it be memorizing and checking looooooong lists after reading them of course.
Will the human brain ever be capable of this? Probably not. Will the human race evolve into a species with a brain capable of doing this? If we live long enough, I imagine so.
believe me, if we go on using technology, computer for our thinking, mental problems then the evolution will lead us to a specie that has lower level of brain than our current one.
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer
Rukado
Posts: 527
Add as Friend
Challenge to a Debate
Send a Message
6/26/2016 5:47:29 AM
Posted: 5 months ago
At 6/26/2016 5:30:58 AM, Artur wrote:
then, I think you have to change your opinion. I myself am can more than that and I know many people, face many people in everyday life that speak more 5-6 languages fluently.

Maybe that depends on what your definition of fluency is (it's more than functional literacy). And, I only said the average person has no potential beyond three languages. Half the population is smarter than average.

I studied language enough to know there's no one who is fluent in dozens of languages. I've seen those supposed hyper-polyglots crash and burn when someone tests their knowledge of languages.
Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/26/2016 7:14:20 AM
Posted: 5 months ago
At 6/26/2016 5:47:29 AM, Rukado wrote:
At 6/26/2016 5:30:58 AM, Artur wrote:
then, I think you have to change your opinion. I myself am can more than that and I know many people, face many people in everyday life that speak more 5-6 languages fluently.

Maybe that depends on what your definition of fluency is (it's more than functional literacy). And, I only said the average person has no potential beyond three languages. Half the population is smarter than average.
oh, sorry then. I misunderstood
I studied language enough to know there's no one who is fluent in dozens of languages. I've seen those supposed hyper-polyglots crash and burn when someone tests their knowledge of languages.
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer
Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/26/2016 7:20:47 AM
Posted: 5 months ago
At 6/25/2016 5:29:50 AM, Cobalt wrote:
At 6/25/2016 4:58:07 AM, Artur wrote:

Our brains have a limited amount of space to store information.

A chess algorithm works by checking all legal solutions some number of turns ahead. It then uses various standards to come up with a "score" and the best scored move will be played. The human brain currently can't a) store that much information with accurate recall and b) isn't fast enough to iterate through all possible moves in a reasonable amount of time, even if you're only looking a few moves ahead.

Will the human brain ever be capable of this? Probably not. Will the human race evolve into a species with a brain capable of doing this? If we live long enough, I imagine so.
ny the way, I know a man whose eyes work like a computer screenshot, once he sees something, he can copy it by drawing.

we even tried him with very very absurd, complex, randomly drawn lines like a draught of children, he did the same copy of it. if you show him an A4 paper written till the end, once he sees it he can repeat it without reading, to see is enough for him. but his memory does not work for other things like remembering poems, homeworks e.t.c

he is a photographer and painter. considering all of them, I think if human tries, his/her brain can be in a level of chess computer.
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer
Stronn
Posts: 318
Add as Friend
Challenge to a Debate
Send a Message
6/27/2016 10:17:06 AM
Posted: 5 months ago
At 6/25/2016 4:58:07 AM, Artur wrote:
Scientifically, is that possible? let us say modern chess engine, talk about chess engines of june 2016.

I read about The chess computer that beat Garry Kasparov 19 years ago was able to calculate 200 million positions per second. Now, almopst 20 years passed and technology advanced, I do not know about current chess engines. if you know, share.

can a human brain achieve a degree in which he can calculate as good, precise, fast as Komodo or stockfish? Theoretically or scientifically, is it possible?

Modern chess engines can usually beat the best human players, and they will only get better.

Since chess has a finite number of possible configurations, in principle a computer could play a perfect game by simply storing the optimal move for every possible state. But with at least 10^120 possible configurations (called the Shannon number) (https://en.wikipedia.org...), computers will have nowhere near the memory or speed required in the foreseeable future. They can and do, however, get pre-programmed with all standard openings and variations on those openings, up to 10 or so moves deep into a game. If a human makes a mistake during those openings, a computer can capitalized on it with near-optimally.

Modern chess engines also use a database lookup for the endgame. As of 2012, all endgames with 7 or fewer total pieces have been solved. So if a game reaches the point where 7 or fewer pieces are left, computers can play perfectly.

The middle game is where a human can still occasionally beat a computer. A typical human thinks 2 or 3 moves ahead. Gary Kasparov said once that he typically thinks 4 or 5 moves ahead, but sometimes as many as 12 or 14 depending on the position. Computers "look ahead" by brute force, analyzing millions of potential future positions a second. While it is hard to put an exact number on it, analyzing all positions 7 or 8 moves ahead is not uncommon for computers.

As a side note, it must either be that 1) white can always win or 2) black can always draw (or win). Which one is true is as yet an open question. Empirically, though, it seems unlikely that 1) is the case, since most grandmaster games end in draws.
Artur
Posts: 723
Add as Friend
Challenge to a Debate
Send a Message
6/28/2016 8:58:06 AM
Posted: 5 months ago
At 6/27/2016 10:17:06 AM, Stronn wrote:
At 6/25/2016 4:58:07 AM, Artur wrote:
Scientifically, is that possible? let us say modern chess engine, talk about chess engines of june 2016.

I read about The chess computer that beat Garry Kasparov 19 years ago was able to calculate 200 million positions per second. Now, almopst 20 years passed and technology advanced, I do not know about current chess engines. if you know, share.

can a human brain achieve a degree in which he can calculate as good, precise, fast as Komodo or stockfish? Theoretically or scientifically, is it possible?

Modern chess engines can usually beat the best human players, and they will only get better.

Since chess has a finite number of possible configurations, in principle a computer could play a perfect game by simply storing the optimal move for every possible state. But with at least 10^120 possible configurations (called the Shannon number) (https://en.wikipedia.org...), computers will have nowhere near the memory or speed required in the foreseeable future. They can and do, however, get pre-programmed with all standard openings and variations on those openings, up to 10 or so moves deep into a game. If a human makes a mistake during those openings, a computer can capitalized on it with near-optimally.

Modern chess engines also use a database lookup for the endgame. As of 2012, all endgames with 7 or fewer total pieces have been solved. So if a game reaches the point where 7 or fewer pieces are left, computers can play perfectly.

The middle game is where a human can still occasionally beat a computer. A typical human thinks 2 or 3 moves ahead. Gary Kasparov said once that he typically thinks 4 or 5 moves ahead, but sometimes as many as 12 or 14 depending on the position. Computers "look ahead" by brute force, analyzing millions of potential future positions a second. While it is hard to put an exact number on it, analyzing all positions 7 or 8 moves ahead is not uncommon for computers.

As a side note, it must either be that 1) white can always win or 2) black can always draw (or win). Which one is true is as yet an open question. Empirically, though, it seems unlikely that 1) is the case, since most grandmaster games end in draws.
I am familiar with chess history, theory and computer chess development, current plight.
I know all what you wrote here. My topic is not about that. I am looking for scientific answer.

is there a law, theory in science that prevents our brain from reaching that highness? is it scientifically possible for our brains to be as good, precise, fast calculator as modern chess engines? if no, how far can we reach? what is the last point our brain can climb?
"I'm not as soft or as generous a person as I would be if the world hadn't changed me" Bobby Fischer