Translation: Barbara Guzik Correction: Agnieszka Cabanes
We are all different.
We have a different height, different color of hair, eyes,
Various political and religious.
However, nowadays,
despite all these differences, one thing unites us,
or mobile phone.
Thanks to advances in technology
each of us can afford such a phone,
and at home waiting for us or laptop computer.
Also, more and more parents give their children
tablet or laptop,
so that played our game, oglądnęły movie,
or that in some other way to keep them entertained.
With the development of technology
It is still bound by the term artificial intelligence,
which is usually associated with something difficult, complicated,
the reach of the common man
and present only in universities or research centers.
One such example can be a computer AlphaGo
produced by Google.
It has been designed to play him in a game,
which is known as one of the hardest games
known to mankind.
The achievement of this computer is won last year with Lee SEDOL.
It was the first victory computer
over a man with such a high level,
because it is considered the best player of the last decade.
This success shows
how artificial intelligence is developed
and what can give the possibility in the future.
However, we can not seem to matter,
but this artificial intelligence
It is already among us, and we use it every day.
Everyone certainly often irritated,
auto-correction when he mistook one word in his speech.
And this correction here
It is already a part of artificial intelligence,
because the program has to analyze characters,
that the user enters,
and then compare them with the words in a dictionary
and see which fit best in a given context.
Another example would be the next word suggestions by the entire keyboard,
and when we have some of your favorite expressions, sayings
or simply grammatical structures,
Keyboard is able to prompt when you type a single word.
I could discuss some other examples,
such as the completion of queries in Google
or displaying posts on Facebook.
But now I do not want to talk about examples, only about
as the artificial intelligence works.
Because, as I said, everyone carries a phone,
But for most, this artificial intelligence
It is something magical, inaccessible,
and still we use it.
So let’s know how it works.
How many people really know how this artificial intelligence?
We can try to estimate
with the amount of software developers in the world.
It is estimated that there are about 20 million,
and seven billion people on earth
This gives us approximately a quarter of a percent.
So as the TEDx is organized for 100 people,
This really should not be here any programmer.
(Laugh)
And yet, everyone carries a phone,
so you want to know how the artificial intelligence works.
And now, what exactly is artificial intelligence
and how it differs from the standard computer programs?
Standard programs are like the one fixed output
and, for example, we want to sort a sequence of numbers,
This is only one correct answer.
Artificial Intelligence deal with problems
which have no definite correct answer,
such as, for example, prompting the following words on the keyboard.
The program can never be 100% sure,
What word would you want to write next.
Yet the difference is,
that artificial intelligence must first learn
which should give answers,
and this process is called machine learning.
There are two main types of machine learning –
learning with the teacher and without teacher.
Learning from a teacher is like learning from a teacher,
that is, the teacher gives examples
and the responses to them.
Old student on the basis of these examples
Knead some general knowledge,
then able to apply this knowledge to the general examples,
even he had not seen before.
For example, let’s indulge in a meadow with animals:
horses, cows and fish.
Uczylibyśmy program in this way,
podawalibyśmy that he coordinates the animal
and a label which is per animal
and then the program after learning –
if we gave him a point, which had not seen before,
For example, a point on the left,
This program will be able to say with fairly high degree of certainty,
that this point will be the horse.
Learning without a teacher is different in that,
we do not know the correct answer,
we only points.
After all, we can distinguish quite well with these points three different groups.
And just learning without a teacher tries to distinguish different groups of points
on the basis of their location.
Let us now examine the term artificial intelligence.
Artificial intelligence tells us that there is some real intelligence,
that this artificial tries to imitate.
The real intelligence is, of course, animals, and we – the people.
So it is a natural idea,
to create a program that will mimic humans.
At the same idea I came up about two years ago,
when I came back from the pool with a friend and we talked about programming.
However, it soon becomes clear that the previous two techniques
not enough to do this program.
Why?
Because in those techniques we teach computer
handing him a fairly large amount of data.
A man knows quite a lot of things.
He can read, write,
speak, understand speech,
walk, use tools.
Each of these things would learn separately
and it is virtually impossible to do physically,
gain this amount of data.
When we look at the man,
This is not everything we learn by observing others
or what someone else tells us.
Take, for example, a child who learns to walk.
It can follow adults and watch as they walk,
However, it is not able to see all the muscle movements –
as they maintain the balance.
They can only try on the basis of its own failures
finally learn how to walk properly.
So the idea is to make a robot that learns itself.
But I still did not know how to do it, so I started to learn.
I signed up for the online course,
one, two, three, about artificial intelligence.
In any course, in turn, I saw once again the introduction
and listed the leading types of machine learning,
ie learning with the teacher, without a teacher.
And at some point I was surprised, because it mentioned a third type of learning,
I had never heard of, or learning with a critic.
It turned out that it is precisely this type of learning, by which I meant,
that is, a way to make the program learned from his own experience.
This course was not based on this type of learning,
He not even talked about this type of learning,
so I had to look for further information alone.
I found the manual
“Reinforcement Learning. An Introduction”
or “Learning the critic. Introduction”
and began to read it.
It turned out that although it was a manual,
read it pretty cool,
because this type of learning, as I said, is based on how a person learns.
So on the basis of their own experiences.
So that learning how to program a computer,
I learned also about how my brain works,
I function as myself
and also other people.
Particularly surprising to me it was
in that the mathematical formulas and algorithms.
that were in the book,
actually I am seen how they work in my life.
And now, as a man working properly?
Imagine that we are in a room where there are two doors: the right and left.
Behind the door is another room,
also with two doors
and so on it goes on forever.
We wonder what happens if you go through that door,
so we decide to first go through the right.
move on
and the ceiling falling us money.
We are pleased to collect the money.
We look, if every time this happens,
when you go through the right door, and actually it happens.
However, we are beginning to intrigue the left door.
What happens if you do not go through?
Can we get more money?
We try to go through it.
And boom! We get something on the head.
We are not satisfied.
Head hurts us.
But maybe it was just this one time,
maybe next time we’ll get the money, however.
We’ll try again.
We pass.
Boom! Again we get the head.
We are already very discouraged to that door
and we decide to go just right door.
And we hope that we will get more and more money.
This is how the working man,
or in any case this action is selected,
for which it expects to get the greatest reward.
Someone might be with me now, however, disagree,
because it is not uncommon that we are helping other people,
often themselves on the losing.
Why are we doing this?
Well, the body which is responsible for the action we take,
of course, it is the brain …
and the brain does not physically feel the value of money.
The brain feels as if his reward or punishment
by how he feels: is cheerful or sad.
And this is by using enzymes that secrete in him.
Therefore when we spend money when they get rid of in some way,
it accompanies us in this grief.
However, when we help others at the same time,
we are happy
and the happiness surpasses sadness.
That’s why we help other people.
In this way also it operates the aforementioned computer AlphaGo,
who gets the winning prize game,
that is, as if passing through the right door,
After losing the game and gets a penalty,
that is, as if passed through the left door.
What ever your computer may be reward or punishment?
The computer is a machine.
The machine counts the number.
So reward or punishment is simply a positive or negative number.
Fascinated by these algorithms, that I met,
I decided to make their own implementation of these programs.
I took the job, which previously built – Shoebill,
whose task was to ride the line,
and reprogrammed it this way, he taught himself to ride on this line.
Here we see – at the beginning of straight drives,
He can not do anything.
It will soon be a learning process.
When the blue LED lights up, it means that he gets the prize,
as if passing through the right door.
When the red LED,
it is as if passed through the left door,
which gets a penalty.
Yes … So they learn.
I will soon be able to overcome the track …
which previously did not know.
And at this point he only performs best stocks,
do not learn.
Here we see a robot sensor.
The sensor sends a signal to the microprocessor,
when it is over the line.
The robot is programmed in this way,
when the central sensors that are above the line,
it gets a high reward.
When the line is located under the side sensors,
then the robot gets a lesser reward.
When the robot completely deviate from the line,
it starts to get a penalty.
It is also programmed in the middle of the job.
I do not need outside to give him rewards or penalties.
So you could compare it to a kind of nervous.
Also previously, then the robot taught
can be noted that when pulled off the line,
This lifted it and improved it so
that was close to the line of the back.
But I did not give it directly to the back line.
This can be compared to the kind of upbringing,
like children keep in check
not giving a good answer directly,
but, well, we keep them in check,
so that they can quickly find themselves on the correct answer.
Thus we return to the beginning,
because raising these children
more and more use of modern technology,
which is used in artificial intelligence.
So we give them tablets, laptops, phones.
So you should know,
how it works, and raise children,
because, as it turns out,
This, the education of children, it must first be taught
and behaved in a certain way.
You should also know that,
because the AI already has a significant impact on our lives.
It will be even bigger with the introduction on the market
artificial technology and augmented reality,
as well as a new project Elon Musk, Neuralink,
which plans to build implants
which will be further expanded the possibilities of the human brain.
So I encourage you to learn of how the artificial intelligence,
because first, we can learn,
how the things we use every day,
and secondly, we also learn, as I did,
how we really act alone
and why we do some of the things this way and not otherwise.
When the project is successful Elon Musk,
then we will know how to operate the two intelligences,
which will be in our head.
Thank you.
(Bravo)
We are all different.
We have a different height, different color of hair, eyes,
Various political and religious.
However, nowadays,
despite all these differences, one thing unites us,
or mobile phone.
Thanks to advances in technology
each of us can afford such a phone,
and at home waiting for us or laptop computer.
Also, more and more parents give their children
tablet or laptop,
so that played our game, oglądnęły movie,
or that in some other way to keep them entertained.
With the development of technology
It is still bound by the term artificial intelligence,
which is usually associated with something difficult, complicated,
the reach of the common man
and present only in universities or research centers.
One such example can be a computer AlphaGo
produced by Google.
It has been designed to play him in a game,
which is known as one of the hardest games
known to mankind.
The achievement of this computer is won last year with Lee SEDOL.
It was the first victory computer
over a man with such a high level,
because it is considered the best player of the last decade.
This success shows
how artificial intelligence is developed
and what can give the possibility in the future.
However, we can not seem to matter,
but this artificial intelligence
It is already among us, and we use it every day.
Everyone certainly often irritated,
auto-correction when he mistook one word in his speech.
And this correction here
It is already a part of artificial intelligence,
because the program has to analyze characters,
that the user enters,
and then compare them with the words in a dictionary
and see which fit best in a given context.
Another example would be the next word suggestions by the entire keyboard,
and when we have some of your favorite expressions, sayings
or simply grammatical structures,
Keyboard is able to prompt when you type a single word.
I could discuss some other examples,
such as the completion of queries in Google
or displaying posts on Facebook.
But now I do not want to talk about examples, only about
as the artificial intelligence works.
Because, as I said, everyone carries a phone,
But for most, this artificial intelligence
It is something magical, inaccessible,
and still we use it.
So let’s know how it works.
How many people really know how this artificial intelligence?
We can try to estimate
with the amount of software developers in the world.
It is estimated that there are about 20 million,
and seven billion people on earth
This gives us approximately a quarter of a percent.
So as the TEDx is organized for 100 people,
This really should not be here any programmer.
(Laugh)
And yet, everyone carries a phone,
so you want to know how the artificial intelligence works.
And now, what exactly is artificial intelligence
and how it differs from the standard computer programs?
Standard programs are like the one fixed output
and, for example, we want to sort a sequence of numbers,
This is only one correct answer.
Artificial Intelligence deal with problems
which have no definite correct answer,
such as, for example, prompting the following words on the keyboard.
The program can never be 100% sure,
What word would you want to write next.
Yet the difference is,
that artificial intelligence must first learn
which should give answers,
and this process is called machine learning.
There are two main types of machine learning –
learning with the teacher and without teacher.
Learning from a teacher is like learning from a teacher,
that is, the teacher gives examples
and the responses to them.
Old student on the basis of these examples
Knead some general knowledge,
then able to apply this knowledge to the general examples,
even he had not seen before.
For example, let’s indulge in a meadow with animals:
horses, cows and fish.
Uczylibyśmy program in this way,
podawalibyśmy that he coordinates the animal
and a label which is per animal
and then the program after learning –
if we gave him a point, which had not seen before,
For example, a point on the left,
This program will be able to say with fairly high degree of certainty,
that this point will be the horse.
Learning without a teacher is different in that,
we do not know the correct answer,
we only points.
After all, we can distinguish quite well with these points three different groups.
And just learning without a teacher tries to distinguish different groups of points
on the basis of their location.
Let us now examine the term artificial intelligence.
Artificial intelligence tells us that there is some real intelligence,
that this artificial tries to imitate.
The real intelligence is, of course, animals, and we – the people.
So it is a natural idea,
to create a program that will mimic humans.
At the same idea I came up about two years ago,
when I came back from the pool with a friend and we talked about programming.
However, it soon becomes clear that the previous two techniques
not enough to do this program.
Why?
Because in those techniques we teach computer
handing him a fairly large amount of data.
A man knows quite a lot of things.
He can read, write,
speak, understand speech,
walk, use tools.
Each of these things would learn separately
and it is virtually impossible to do physically,
gain this amount of data.
When we look at the man,
This is not everything we learn by observing others
or what someone else tells us.
Take, for example, a child who learns to walk.
It can follow adults and watch as they walk,
However, it is not able to see all the muscle movements –
as they maintain the balance.
They can only try on the basis of its own failures
finally learn how to walk properly.
So the idea is to make a robot that learns itself.
But I still did not know how to do it, so I started to learn.
I signed up for the online course,
one, two, three, about artificial intelligence.
In any course, in turn, I saw once again the introduction
and listed the leading types of machine learning,
ie learning with the teacher, without a teacher.
And at some point I was surprised, because it mentioned a third type of learning,
I had never heard of, or learning with a critic.
It turned out that it is precisely this type of learning, by which I meant,
that is, a way to make the program learned from his own experience.
This course was not based on this type of learning,
He not even talked about this type of learning,
so I had to look for further information alone.
I found the manual
“Reinforcement Learning. An Introduction”
or “Learning the critic. Introduction”
and began to read it.
It turned out that although it was a manual,
read it pretty cool,
because this type of learning, as I said, is based on how a person learns.
So on the basis of their own experiences.
So that learning how to program a computer,
I learned also about how my brain works,
I function as myself
and also other people.
Particularly surprising to me it was
in that the mathematical formulas and algorithms.
that were in the book,
actually I am seen how they work in my life.
And now, as a man working properly?
Imagine that we are in a room where there are two doors: the right and left.
Behind the door is another room,
also with two doors
and so on it goes on forever.
We wonder what happens if you go through that door,
so we decide to first go through the right.
move on
and the ceiling falling us money.
We are pleased to collect the money.
We look, if every time this happens,
when you go through the right door, and actually it happens.
However, we are beginning to intrigue the left door.
What happens if you do not go through?
Can we get more money?
We try to go through it.
And boom! We get something on the head.
We are not satisfied.
Head hurts us.
But maybe it was just this one time,
maybe next time we’ll get the money, however.
We’ll try again.
We pass.
Boom! Again we get the head.
We are already very discouraged to that door
and we decide to go just right door.
And we hope that we will get more and more money.
This is how the working man,
or in any case this action is selected,
for which it expects to get the greatest reward.
Someone might be with me now, however, disagree,
because it is not uncommon that we are helping other people,
often themselves on the losing.
Why are we doing this?
Well, the body which is responsible for the action we take,
of course, it is the brain …
and the brain does not physically feel the value of money.
The brain feels as if his reward or punishment
by how he feels: is cheerful or sad.
And this is by using enzymes that secrete in him.
Therefore when we spend money when they get rid of in some way,
it accompanies us in this grief.
However, when we help others at the same time,
we are happy
and the happiness surpasses sadness.
That’s why we help other people.
In this way also it operates the aforementioned computer AlphaGo,
who gets the winning prize game,
that is, as if passing through the right door,
After losing the game and gets a penalty,
that is, as if passed through the left door.
What ever your computer may be reward or punishment?
The computer is a machine.
The machine counts the number.
So reward or punishment is simply a positive or negative number.
Fascinated by these algorithms, that I met,
I decided to make their own implementation of these programs.
I took the job, which previously built – Shoebill,
whose task was to ride the line,
and reprogrammed it this way, he taught himself to ride on this line.
Here we see – at the beginning of straight drives,
He can not do anything.
It will soon be a learning process.
When the blue LED lights up, it means that he gets the prize,
as if passing through the right door.
When the red LED,
it is as if passed through the left door,
which gets a penalty.
Yes … So they learn.
I will soon be able to overcome the track …
which previously did not know.
And at this point he only performs best stocks,
do not learn.
Here we see a robot sensor.
The sensor sends a signal to the microprocessor,
when it is over the line.
The robot is programmed in this way,
when the central sensors that are above the line,
it gets a high reward.
When the line is located under the side sensors,
then the robot gets a lesser reward.
When the robot completely deviate from the line,
it starts to get a penalty.
It is also programmed in the middle of the job.
I do not need outside to give him rewards or penalties.
So you could compare it to a kind of nervous.
Also previously, then the robot taught
can be noted that when pulled off the line,
This lifted it and improved it so
that was close to the line of the back.
But I did not give it directly to the back line.
This can be compared to the kind of upbringing,
like children keep in check
not giving a good answer directly,
but, well, we keep them in check,
so that they can quickly find themselves on the correct answer.
Thus we return to the beginning,
because raising these children
more and more use of modern technology,
which is used in artificial intelligence.
So we give them tablets, laptops, phones.
So you should know,
how it works, and raise children,
because, as it turns out,
This, the education of children, it must first be taught
and behaved in a certain way.
You should also know that,
because the AI already has a significant impact on our lives.
It will be even bigger with the introduction on the market
artificial technology and augmented reality,
as well as a new project Elon Musk, Neuralink,
which plans to build implants
which will be further expanded the possibilities of the human brain.
So I encourage you to learn of how the artificial intelligence,
because first, we can learn,
how the things we use every day,
and secondly, we also learn, as I did,
how we really act alone
and why we do some of the things this way and not otherwise.
When the project is successful Elon Musk,
then we will know how to operate the two intelligences,
which will be in our head.
Thank you.
(Bravo)
Please follow and like us: