Translation: Esra Çakmak Review: Figen Ergürbüz
(Applause)
Hello.
In March of 1997, at Istanbul Technical University,
we had a popular science conference.
I also talked about artificial intelligence there.
I will talk to you soon,
World Human Chess champion will be renewed by a computer
I claimed.
On top of that, with one of the viewers there
we had a heated debate.
This is never possible, people are playing chess
claiming that computers use something that they can never have
with an audience.
I wish I had a bet, because after two months on top of this,
World Human Chess Champion Garri Kasparov,
In a New York game,
It was badly defeated on IBM’s Deep Blue computer.
I, as an expert, lived the pride of being righteous.
But the same pride, at the beginning of this year, I could not live.
Because, in 2015, any artificial intelligence teacher, including myself,
When the famous Far East game Go computers
when the Go champion will be captured by computers
If you ask to be renovated, ohoo we got it to work at least ten years.
But, the incident happened in March of 2016.
South Korean Lee Sedol, one of the world’s best actors,
Google Deep Mind has developed a secret from us
A computer program called Alfa Go was defeated.
Artificial intelligence, like a whirlpool, starts slowly and accelerates
and even surprised by his expertise, which affects us
it is moving at a pace.
Let’s take it from the beginning.
Our story was initiated in 1769 by the gentleman:
Gottfried Wilhelm Leibniz.
A very big man, but the part about our position,
This machine is built.
This machine has four arithmetic operations in history: addition, subtraction, multiplication, division,
The first mechanical device that can accomplish all at once.
And Leibniz is not that far away,
according to what we can do with a machine capable of doing this sublime work that requires the human mind,
whether it thinks about the rest of the human mind’s activities,
we can make the machine do not judge, he thought.
But for that, we can write and display arithmetic symbols
Just like it is a slice,
as well as in a language that can show formality like mathematics
I need it.
175 years after that, that gentleman, George Boole,
The English mathematician has actually come up with this language.
Now thanks to George Boole, concepts such as right and wrong
expressing with numbers such as one and zero
and the operators that establish the relationship between the cues, such as,
in a mathematical language,
everything started Boole.
On this basis Boole laid down, in 1910
Bertrand Russell is a very famous Englishman,
If the logic of mathematics is married, let’s look at the other side of the work;
nothing in mathematics, no wrong can not be proved,
Let’s show your mathematics is solid in this sense, “he built an infrastructure.
A genius mathematician named Kurt Gödel who used this infrastructure, in 1931,
He has a bomb in the world of mathematics,
because he discovered that it was a kind of hole in mathematics,
some truths can never be proved by the mathematical system,
not because of the incompetence of mathematicians,
revealing that the nature of mathematics is from its own lack.
What does this have to do with your position?
I have that interest.
Alan Turing, my favorite scientist, is impressed by Gödel’s work,
In 1936, some questions were never made to any computer or any physical laws
can not be answered by the mechanism that works according to
proved the proof, and in the article which this prophecy put forward,
as a kind of by-product, we now call the Turing machine,
described a kind of machine.
This machine is a great machine,
because it is a machine that can imitate any other system.
Turing did not leave it on paper, not in World War I,
After some of the heroism of World War II,
In 1946, how did such machines work in the real world
he gave us an example of what we could build.
Among the people who made the first electronic computer built in England
Turing also arrived, not so much,
the potential of these machines and in 1950, these machines really
since they can imitate all cognitive activity,
Towards the year 2000, we can say that they are officially intelligent,
they will be able to do all the work that requires intelligence and when they are called intelligent
people have not forgotten it, “he said.
In 1956, the Dartmouth conference in America
scientific meeting was held.
We are now in 2016.
The Dartmouth conference is the beginning of artificial intelligence.
So we’re 50 years of artificial intelligence.
During this time, things started slowly
gradually brought us to today.
I’m telling artificial intelligence techniques by taking a stand-in.
At the beginning, how computers are doing clever jobs,
teaching ” by teaching ”.
So for example chess is a good example.
Programmers who program chess-playing computers,
first they knew very well how to play the chess,
how to win, a very detailed stupid way to tell, step by step
They were preparing the instructions and loading them on the computer, so they programmed it.
And since the computer can do it any faster than anybody else
from any person, from the promising world champion,
it was becoming better chess.
By describing this, the teacher’s ability to understand your cognitive activity very well
you need to be able to tell it to be able to program.
But for some activities, we could not use this system.
For example, in an image we recognize the appearance of a friend.
Because we people can do it very comfortably,
we do not know how we do it.
But Allah has found a way to get this done with computers,
in the way I teach teaching by showing.
My ability to recognize my uncle’s face in a picture
if you want to show computer,
the image of thousands of color points forming that picture
I give it to the computer and in my pictures it is in my uncle yes,
Thousands, thousands of positive and negative pictures
I am showing example computer.
My computer, my own human brain
in the representation of millions of parametres prepared in analogy,
making small changes every time,
this recognition is gradually reducing the error you have made in the business,
After a while, my uncle really did
Even in the new pictures I have never shown,
There is a program that we can recognize as we do not exist.
Now, this work has become feasible by computers.
But the one that excited me the most,
my “self-learning” technique,
our computers are out there, for example, this Go champion
As Alfa Go did,
they learn with a technique that goes beyond teaching by showing.
How to beat the Go champion in Alfa Go
or by showing, because no-one champion of Go
He did not know how to beat it.
Alfa Go,
a hundred and fifty thousand people have seen the Go game
After learning how people are doing this job,
playing thousands of matches, tens of thousands of times
whichever of his copies he won in a match,
parameters to make it look like it
he came to the superhuman level, creating his own experience.
Artificial intelligence, he took his head.
And indeed artificial intelligence makes us happy.
I mean today, using artificial intelligence, that is, using a mobile phone,
people live ten times better than people who do not have this feature.
In the traffic they are getting easier on the road, shorter time.
Even in some countries,
they do not have to drive a car themselves.
They can negotiate with people who do not know another language.
There are ways to make a whole new friend
and with people who are interested in their interests,
even if it’s on the other side of the world
they can make nice friendships and of course I can not give all the examples here,
because there are so many good things that really do not fit into a TED talk
earning us artificial intelligence.
But there’s one thing that scares some people,
I do not know if it ‘s dark.
A company in South Korea,
completely self-employed, that is, people do not
a person who shoots and shoots at a remote person
intelligent automatic weapon.
In America, when war plane pilots have to learn this,
one of the simulation programs, including all people, including experts,
there is a war flight program flying higher than all people.
Are we missing the rope tip?
The following example, which lived in March of 2016,
Maybe give us an idea.
Microsoft company,
He invented a Twitter character called Tay,
This teenager was built on the idea of simulating a girl.
Tay’s feature, he will throw tweets, people will answer him
and has the learning feature,
in this way people will learn more about the language features.
As we can see in this news, he said that he would talk more naturally,
thanks to the learning feature.
From this news, exactly 24 hours after the opening of Tay,
The news came out.
Tay learned a little too much, some malicious people
they had spoken so easily with him,
Tay in 24 hours; Hitler’s fan, genocide is a very good thing,
he turned into a hateful character who said things about the wretched sex.
And Microsoft immediately closed the account at that time.
So we can teach computers to evil.
Unfortunately.
Our computers are becoming increasingly intelligent,
I just tried to tell you, it’s a good thing.
This is because we do not even think of it until a few years ago,
but as much as our children will not even think about giving up
it provides vital benefits.
But on the other hand, if they are too smart, if they get too much control,
do they make us evil,
there are those who fear how bad people will be in their hands.
Some people say that, to fight this possibility,
we prohibit some cognitive activities from being done by computers.
So computers with this kind of capability, programs
prohibition of production.
As far as I can see, for obvious reasons in history,
if a property gives a competitive advantage to a person who owns a device,
such prohibitions never work.
As people who obey the law, you are behind those who do not obey the law.
That’s why I do not believe it’s going to work.
Another idea,
as you know,
then let’s get my computers a kind of morality.
Isaac Asimov, American science fiction writer Isaac Asimov,
There is a famous invention called Robotic Laws.
If you read, you know, the robots in Asimov’s novels
they are furnished with some moral code at the factory;
so you will never harm people, you will protect yourself,
but if you are talking about harming a person,
like you will not protect yourself.
These bases are so buried in their brains that they can never get out of it.
And Asimov’s novels, robots with such a moral system
It is filled with interesting events that happen to you.
I think something like this, or our encoding,
or using the learning systems I just talked about
we teach our computers by showing good moral behavior
it seems like a good solution.
Again, not at the beginning of this year, not even in the beginning
I think two months ago,
I published an article on the following example, which I talked about a few.
This means that Google’s famous translation system
I am interested in the translation between Turkish and English.
And from Turkish to English ” I am a politician, ”
a sentence to be translated,
” I’m not a politician, ” he says since decades.
I would like to be able to tell you the technical reasons for this.
But that was funny.
From English to Turkish, ” Davutoglu attaches importance to climate change, ”
persistently
” Obama attaches importance to climate change, ” he said,
We laughed a lot.
But now if you open your phone and look at it,
you will see that you do not.
Because a month ago Google launched a Neural Machine Translation System
he went to a new system
and such specimens can now be turned into perfect shapes.
Even in the beginning of hocans in universities teaching Turkish
a new problem is about to open, because the technical translator is doing very well
and now a piece of work written by your student or a thesis, hop
You will not be able to find it in an English language and translate it.
Likewise, you can automatically send people the following fraudulent e-mails
the foreigners who send it are now Turkish,
will be able to use it much better than the old.
We should see our computers like our children.
We can not give up.
It is our responsibility to show them the right way.
Thus, against bad people and bad computers
but so we can fight together.
Thank you.
(Applause)
(Applause)
Hello.
In March of 1997, at Istanbul Technical University,
we had a popular science conference.
I also talked about artificial intelligence there.
I will talk to you soon,
World Human Chess champion will be renewed by a computer
I claimed.
On top of that, with one of the viewers there
we had a heated debate.
This is never possible, people are playing chess
claiming that computers use something that they can never have
with an audience.
I wish I had a bet, because after two months on top of this,
World Human Chess Champion Garri Kasparov,
In a New York game,
It was badly defeated on IBM’s Deep Blue computer.
I, as an expert, lived the pride of being righteous.
But the same pride, at the beginning of this year, I could not live.
Because, in 2015, any artificial intelligence teacher, including myself,
When the famous Far East game Go computers
when the Go champion will be captured by computers
If you ask to be renovated, ohoo we got it to work at least ten years.
But, the incident happened in March of 2016.
South Korean Lee Sedol, one of the world’s best actors,
Google Deep Mind has developed a secret from us
A computer program called Alfa Go was defeated.
Artificial intelligence, like a whirlpool, starts slowly and accelerates
and even surprised by his expertise, which affects us
it is moving at a pace.
Let’s take it from the beginning.
Our story was initiated in 1769 by the gentleman:
Gottfried Wilhelm Leibniz.
A very big man, but the part about our position,
This machine is built.
This machine has four arithmetic operations in history: addition, subtraction, multiplication, division,
The first mechanical device that can accomplish all at once.
And Leibniz is not that far away,
according to what we can do with a machine capable of doing this sublime work that requires the human mind,
whether it thinks about the rest of the human mind’s activities,
we can make the machine do not judge, he thought.
But for that, we can write and display arithmetic symbols
Just like it is a slice,
as well as in a language that can show formality like mathematics
I need it.
175 years after that, that gentleman, George Boole,
The English mathematician has actually come up with this language.
Now thanks to George Boole, concepts such as right and wrong
expressing with numbers such as one and zero
and the operators that establish the relationship between the cues, such as,
in a mathematical language,
everything started Boole.
On this basis Boole laid down, in 1910
Bertrand Russell is a very famous Englishman,
If the logic of mathematics is married, let’s look at the other side of the work;
nothing in mathematics, no wrong can not be proved,
Let’s show your mathematics is solid in this sense, “he built an infrastructure.
A genius mathematician named Kurt Gödel who used this infrastructure, in 1931,
He has a bomb in the world of mathematics,
because he discovered that it was a kind of hole in mathematics,
some truths can never be proved by the mathematical system,
not because of the incompetence of mathematicians,
revealing that the nature of mathematics is from its own lack.
What does this have to do with your position?
I have that interest.
Alan Turing, my favorite scientist, is impressed by Gödel’s work,
In 1936, some questions were never made to any computer or any physical laws
can not be answered by the mechanism that works according to
proved the proof, and in the article which this prophecy put forward,
as a kind of by-product, we now call the Turing machine,
described a kind of machine.
This machine is a great machine,
because it is a machine that can imitate any other system.
Turing did not leave it on paper, not in World War I,
After some of the heroism of World War II,
In 1946, how did such machines work in the real world
he gave us an example of what we could build.
Among the people who made the first electronic computer built in England
Turing also arrived, not so much,
the potential of these machines and in 1950, these machines really
since they can imitate all cognitive activity,
Towards the year 2000, we can say that they are officially intelligent,
they will be able to do all the work that requires intelligence and when they are called intelligent
people have not forgotten it, “he said.
In 1956, the Dartmouth conference in America
scientific meeting was held.
We are now in 2016.
The Dartmouth conference is the beginning of artificial intelligence.
So we’re 50 years of artificial intelligence.
During this time, things started slowly
gradually brought us to today.
I’m telling artificial intelligence techniques by taking a stand-in.
At the beginning, how computers are doing clever jobs,
teaching ” by teaching ”.
So for example chess is a good example.
Programmers who program chess-playing computers,
first they knew very well how to play the chess,
how to win, a very detailed stupid way to tell, step by step
They were preparing the instructions and loading them on the computer, so they programmed it.
And since the computer can do it any faster than anybody else
from any person, from the promising world champion,
it was becoming better chess.
By describing this, the teacher’s ability to understand your cognitive activity very well
you need to be able to tell it to be able to program.
But for some activities, we could not use this system.
For example, in an image we recognize the appearance of a friend.
Because we people can do it very comfortably,
we do not know how we do it.
But Allah has found a way to get this done with computers,
in the way I teach teaching by showing.
My ability to recognize my uncle’s face in a picture
if you want to show computer,
the image of thousands of color points forming that picture
I give it to the computer and in my pictures it is in my uncle yes,
Thousands, thousands of positive and negative pictures
I am showing example computer.
My computer, my own human brain
in the representation of millions of parametres prepared in analogy,
making small changes every time,
this recognition is gradually reducing the error you have made in the business,
After a while, my uncle really did
Even in the new pictures I have never shown,
There is a program that we can recognize as we do not exist.
Now, this work has become feasible by computers.
But the one that excited me the most,
my “self-learning” technique,
our computers are out there, for example, this Go champion
As Alfa Go did,
they learn with a technique that goes beyond teaching by showing.
How to beat the Go champion in Alfa Go
or by showing, because no-one champion of Go
He did not know how to beat it.
Alfa Go,
a hundred and fifty thousand people have seen the Go game
After learning how people are doing this job,
playing thousands of matches, tens of thousands of times
whichever of his copies he won in a match,
parameters to make it look like it
he came to the superhuman level, creating his own experience.
Artificial intelligence, he took his head.
And indeed artificial intelligence makes us happy.
I mean today, using artificial intelligence, that is, using a mobile phone,
people live ten times better than people who do not have this feature.
In the traffic they are getting easier on the road, shorter time.
Even in some countries,
they do not have to drive a car themselves.
They can negotiate with people who do not know another language.
There are ways to make a whole new friend
and with people who are interested in their interests,
even if it’s on the other side of the world
they can make nice friendships and of course I can not give all the examples here,
because there are so many good things that really do not fit into a TED talk
earning us artificial intelligence.
But there’s one thing that scares some people,
I do not know if it ‘s dark.
A company in South Korea,
completely self-employed, that is, people do not
a person who shoots and shoots at a remote person
intelligent automatic weapon.
In America, when war plane pilots have to learn this,
one of the simulation programs, including all people, including experts,
there is a war flight program flying higher than all people.
Are we missing the rope tip?
The following example, which lived in March of 2016,
Maybe give us an idea.
Microsoft company,
He invented a Twitter character called Tay,
This teenager was built on the idea of simulating a girl.
Tay’s feature, he will throw tweets, people will answer him
and has the learning feature,
in this way people will learn more about the language features.
As we can see in this news, he said that he would talk more naturally,
thanks to the learning feature.
From this news, exactly 24 hours after the opening of Tay,
The news came out.
Tay learned a little too much, some malicious people
they had spoken so easily with him,
Tay in 24 hours; Hitler’s fan, genocide is a very good thing,
he turned into a hateful character who said things about the wretched sex.
And Microsoft immediately closed the account at that time.
So we can teach computers to evil.
Unfortunately.
Our computers are becoming increasingly intelligent,
I just tried to tell you, it’s a good thing.
This is because we do not even think of it until a few years ago,
but as much as our children will not even think about giving up
it provides vital benefits.
But on the other hand, if they are too smart, if they get too much control,
do they make us evil,
there are those who fear how bad people will be in their hands.
Some people say that, to fight this possibility,
we prohibit some cognitive activities from being done by computers.
So computers with this kind of capability, programs
prohibition of production.
As far as I can see, for obvious reasons in history,
if a property gives a competitive advantage to a person who owns a device,
such prohibitions never work.
As people who obey the law, you are behind those who do not obey the law.
That’s why I do not believe it’s going to work.
Another idea,
as you know,
then let’s get my computers a kind of morality.
Isaac Asimov, American science fiction writer Isaac Asimov,
There is a famous invention called Robotic Laws.
If you read, you know, the robots in Asimov’s novels
they are furnished with some moral code at the factory;
so you will never harm people, you will protect yourself,
but if you are talking about harming a person,
like you will not protect yourself.
These bases are so buried in their brains that they can never get out of it.
And Asimov’s novels, robots with such a moral system
It is filled with interesting events that happen to you.
I think something like this, or our encoding,
or using the learning systems I just talked about
we teach our computers by showing good moral behavior
it seems like a good solution.
Again, not at the beginning of this year, not even in the beginning
I think two months ago,
I published an article on the following example, which I talked about a few.
This means that Google’s famous translation system
I am interested in the translation between Turkish and English.
And from Turkish to English ” I am a politician, ”
a sentence to be translated,
” I’m not a politician, ” he says since decades.
I would like to be able to tell you the technical reasons for this.
But that was funny.
From English to Turkish, ” Davutoglu attaches importance to climate change, ”
persistently
” Obama attaches importance to climate change, ” he said,
We laughed a lot.
But now if you open your phone and look at it,
you will see that you do not.
Because a month ago Google launched a Neural Machine Translation System
he went to a new system
and such specimens can now be turned into perfect shapes.
Even in the beginning of hocans in universities teaching Turkish
a new problem is about to open, because the technical translator is doing very well
and now a piece of work written by your student or a thesis, hop
You will not be able to find it in an English language and translate it.
Likewise, you can automatically send people the following fraudulent e-mails
the foreigners who send it are now Turkish,
will be able to use it much better than the old.
We should see our computers like our children.
We can not give up.
It is our responsibility to show them the right way.
Thus, against bad people and bad computers
but so we can fight together.
Thank you.
(Applause)
Please follow and like us: