Press "Enter" to skip to content

Machine learning, automation and the implications for everyone | ALESIS NOVIK | TEDxISM


hello everyone I’m very excited to be

hello everyone I’m very excited to be here today and talk to you all about

here today and talk to you all about

here today and talk to you all about machine learning and what it means to us

machine learning and what it means to us

machine learning and what it means to us machine learning is everywhere now

machine learning is everywhere now

machine learning is everywhere now you can’t take a newspaper without

you can’t take a newspaper without

you can’t take a newspaper without reading how artificial intelligence is

reading how artificial intelligence is

reading how artificial intelligence is making our lives simpler or more

making our lives simpler or more

making our lives simpler or more efficient yet some are still anxious

efficient yet some are still anxious

efficient yet some are still anxious about what automation holds for the

about what automation holds for the

about what automation holds for the future however I’m not here to talk

future however I’m not here to talk

future however I’m not here to talk about the social implications of this

about the social implications of this

about the social implications of this technology I am a machine learning

technology I am a machine learning

technology I am a machine learning person and every day I see the impact it

person and every day I see the impact it

person and every day I see the impact it has and then told valued holds for the

has and then told valued holds for the

has and then told valued holds for the future but I’m jumping ahead let me

future but I’m jumping ahead let me

future but I’m jumping ahead let me introduce myself I am Alice’s co-founder

introduce myself I am Alice’s co-founder

introduce myself I am Alice’s co-founder and chief technology officer of a m–

and chief technology officer of a m–

and chief technology officer of a m– brain a biometrics company from London

brain a biometrics company from London

brain a biometrics company from London before starting up the company I went to

before starting up the company I went to

before starting up the company I went to university right here in Vilnius where I

university right here in Vilnius where I

university right here in Vilnius where I studied computer science after having

studied computer science after having

studied computer science after having called the machine learning and

called the machine learning and

called the machine learning and artificial intelligence bug I decided to

artificial intelligence bug I decided to

artificial intelligence bug I decided to further my studies in the University of

further my studies in the University of

further my studies in the University of Edinburgh almost did my PhD but decided

Edinburgh almost did my PhD but decided

Edinburgh almost did my PhD but decided it would be more fun to try it in

it would be more fun to try it in

it would be more fun to try it in practice now I’m here today to talk

practice now I’m here today to talk

practice now I’m here today to talk about machine learning and how it’s

about machine learning and how it’s

about machine learning and how it’s everywhere and this is not gonna be a

everywhere and this is not gonna be a

everywhere and this is not gonna be a lecture so don’t worry all I hope for is

lecture so don’t worry all I hope for is

lecture so don’t worry all I hope for is that by the end of my presentation you

that by the end of my presentation you

that by the end of my presentation you will all have the same passion as I do

will all have the same passion as I do

will all have the same passion as I do about questioning the expected even when

about questioning the expected even when

about questioning the expected even when you think it is true now let’s get

you think it is true now let’s get

you think it is true now let’s get started does anyone here own an

started does anyone here own an

started does anyone here own an autonomous car no well but I bet you use

autonomous car no well but I bet you use

autonomous car no well but I bet you use recommendations when compiling your

recommendations when compiling your

recommendations when compiling your music playlists or when shopping online

music playlists or when shopping online

music playlists or when shopping online if you use Gmail you can see that it can

if you use Gmail you can see that it can

if you use Gmail you can see that it can now predict your responses so you can

now predict your responses so you can

now predict your responses so you can reply with a single push of a button

reply with a single push of a button

reply with a single push of a button it can also recommend you some more

it can also recommend you some more

it can also recommend you some more interesting and extreme sports

interesting and extreme sports

interesting and extreme sports activities as well this is machine

activities as well this is machine

activities as well this is machine learning and it is everywhere however

learning and it is everywhere however

learning and it is everywhere however today I want to talk to you about some

today I want to talk to you about some

today I want to talk to you about some more interesting and creative

more interesting and creative

more interesting and creative applications of machine learning to do

applications of machine learning to do

applications of machine learning to do this I first need to mention deep

this I first need to mention deep

this I first need to mention deep learning this is gonna be a little bit

learning this is gonna be a little bit

learning this is gonna be a little bit technical but trust me you need to know

technical but trust me you need to know

technical but trust me you need to know this because this technology is

this because this technology is

this because this technology is underlying most of the things used today

underlying most of the things used today

underlying most of the things used today at a m– brain everything we build and

at a m– brain everything we build and

at a m– brain everything we build and use is based on deep learning it is one

use is based on deep learning it is one

use is based on deep learning it is one of the most ubiquitous technologies out

of the most ubiquitous technologies out

of the most ubiquitous technologies out there today without actually ever being

there today without actually ever being

there today without actually ever being seen like a famous celebrity everyone

seen like a famous celebrity everyone

seen like a famous celebrity everyone has probably at least heard of deep

has probably at least heard of deep

has probably at least heard of deep learning

learning

learning however unlike that person that somehow

however unlike that person that somehow

however unlike that person that somehow is always on TV no one really sees deep

is always on TV no one really sees deep

is always on TV no one really sees deep learning I won’t go into too deep of a

learning I won’t go into too deep of a

learning I won’t go into too deep of a technical and explanation of deep

technical and explanation of deep

technical and explanation of deep learning and its methods but suffice it

learning and its methods but suffice it

learning and its methods but suffice it to say that they’re considered universal

to say that they’re considered universal

to say that they’re considered universal approximator what this means is that it

approximator what this means is that it

approximator what this means is that it can they can approximate almost any

can they can approximate almost any

can they can approximate almost any function to put it into simple terms a

function to put it into simple terms a

function to put it into simple terms a large enough deep neural network can be

large enough deep neural network can be

large enough deep neural network can be trained to do a wide variety of tasks as

trained to do a wide variety of tasks as

trained to do a wide variety of tasks as long as you have a data and a way to

long as you have a data and a way to

long as you have a data and a way to quantify the objective and something

quantify the objective and something

quantify the objective and something like this in a nutshell now I’ve started

like this in a nutshell now I’ve started

like this in a nutshell now I’ve started talking about what machine learning we

talking about what machine learning we

talking about what machine learning we are experiencing today with predictive

are experiencing today with predictive

are experiencing today with predictive text an autonomous driving

text an autonomous driving

text an autonomous driving but what if machine learning could be

but what if machine learning could be

but what if machine learning could be applied to more creative industries such

applied to more creative industries such

applied to more creative industries such as art music or writing well it can this

as art music or writing well it can this

as art music or writing well it can this would be a rather short and really

would be a rather short and really

would be a rather short and really rather pointless presentation if it

rather pointless presentation if it

rather pointless presentation if it couldn’t with the latest advancements in

couldn’t with the latest advancements in

couldn’t with the latest advancements in deep learning we can now create the

deep learning we can now create the

deep learning we can now create the creative

creative

creative and I want to show you three examples of

and I want to show you three examples of

and I want to show you three examples of exactly what I mean let’s start with

exactly what I mean let’s start with

exactly what I mean let’s start with text generation now you all are probably

text generation now you all are probably

text generation now you all are probably using this already in some capacity with

using this already in some capacity with

using this already in some capacity with your emails or your messages but there’s

your emails or your messages but there’s

your emails or your messages but there’s a whole world beyond that let’s take

a whole world beyond that let’s take

a whole world beyond that let’s take some text that has been generated by a

some text that has been generated by a

some text that has been generated by a bot that’s been trained by one of the

bot that’s been trained by one of the

bot that’s been trained by one of the Great’s Shakespeare and I won’t make you

Great’s Shakespeare and I won’t make you

Great’s Shakespeare and I won’t make you read this because we have computers

read this because we have computers

read this because we have computers these days and you might think that they

these days and you might think that they

these days and you might think that they still sound that’s on the idea arose

still sound that’s on the idea arose

still sound that’s on the idea arose from the ground something like that

from the ground something like that

from the ground something like that while in reality with the advancements

while in reality with the advancements

while in reality with the advancements of deep learning they sound more like

of deep learning they sound more like

of deep learning they sound more like this King Lear

this King Lear

this King Lear I’ve heard make thee and but what are

I’ve heard make thee and but what are

I’ve heard make thee and but what are the driving words and give you against

the driving words and give you against

the driving words and give you against it against her shoot live save you to

it against her shoot live save you to

it against her shoot live save you to the win for when my sick in strings and

the win for when my sick in strings and

the win for when my sick in strings and remedy has some haze on my gentleman

remedy has some haze on my gentleman

remedy has some haze on my gentleman Ross Mary and by his good in prison with

Ross Mary and by his good in prison with

Ross Mary and by his good in prison with the victory with fire house house mired

the victory with fire house house mired

the victory with fire house house mired in her swiftly somewhat with a lisp and

in her swiftly somewhat with a lisp and

in her swiftly somewhat with a lisp and Shakespeare is not the only master we

Shakespeare is not the only master we

Shakespeare is not the only master we can learn from we can also learn from

can learn from we can also learn from

can learn from we can also learn from the great such as Dostoyevsky and

the great such as Dostoyevsky and

the great such as Dostoyevsky and Tolstoy to the song these ideas arose

Tolstoy to the song these ideas arose

Tolstoy to the song these ideas arose from the ground after the brash one was

from the ground after the brash one was

from the ground after the brash one was the French of feelings even given for

the French of feelings even given for

the French of feelings even given for him to report to himself to her account

him to report to himself to her account

him to report to himself to her account authority she would in desperate

authority she would in desperate

authority she would in desperate protects grew here about the Saints

protects grew here about the Saints

protects grew here about the Saints spirits very Brailled she did not suffer

spirits very Brailled she did not suffer

spirits very Brailled she did not suffer and that there was nothing to say but

and that there was nothing to say but

and that there was nothing to say but elastic first without which a quick he

elastic first without which a quick he

elastic first without which a quick he had came into small ensk

had came into small ensk

had came into small ensk now while we can see that there are some

now while we can see that there are some

now while we can see that there are some questionable issues with this you can

questionable issues with this you can

questionable issues with this you can definitely feel the flavor of the

definitely feel the flavor of the

definitely feel the flavor of the author’s and obviously the voice is also

author’s and obviously the voice is also

author’s and obviously the voice is also has advanced significantly now what you

has advanced significantly now what you

has advanced significantly now what you really tend to think about is how this

really tend to think about is how this

really tend to think about is how this technology can help current writers with

technology can help current writers with

technology can help current writers with tasks such as getting over writer’s

tasks such as getting over writer’s

tasks such as getting over writer’s block and words are not everything what

block and words are not everything what

block and words are not everything what about art sure creating complex pieces

about art sure creating complex pieces

about art sure creating complex pieces of art from scratch is still a very

of art from scratch is still a very

of art from scratch is still a very complicated project but what if

complicated project but what if

complicated project but what if you take a photo and turn it into a

you take a photo and turn it into a

you take a photo and turn it into a Monet or a Picasso well with deep

Monet or a Picasso well with deep

Monet or a Picasso well with deep learning we can let’s take this for

learning we can let’s take this for

learning we can let’s take this for example a beautiful painting by Vincent

example a beautiful painting by Vincent

example a beautiful painting by Vincent van Gogh called the starry night and

van Gogh called the starry night and

van Gogh called the starry night and also let’s take this beautiful photo of

also let’s take this beautiful photo of

also let’s take this beautiful photo of the Vilnius castle tower at night

the Vilnius castle tower at night

the Vilnius castle tower at night now with deep learning we can combine

now with deep learning we can combine

now with deep learning we can combine this and I present to you the Vilnius

this and I present to you the Vilnius

this and I present to you the Vilnius castle tower at starry night now you can

castle tower at starry night now you can

castle tower at starry night now you can see how these tools can help designers

see how these tools can help designers

see how these tools can help designers and artists and really simplify their

and artists and really simplify their

and artists and really simplify their work and know talk about creative

work and know talk about creative

work and know talk about creative industries can be finished without music

industries can be finished without music

industries can be finished without music we all know that music comes from the

we all know that music comes from the

we all know that music comes from the soul but with all respect to all the

soul but with all respect to all the

soul but with all respect to all the musicians in the house deep learning can

musicians in the house deep learning can

musicians in the house deep learning can also be taught to play music let’s

also be taught to play music let’s

also be taught to play music let’s listen to a classical piece by Bach

beautiful now let’s listen by something

beautiful now let’s listen by something that was composed by a deep neural

that was composed by a deep neural

that was composed by a deep neural network that learned under Bach

now this all was done with a simple

now this all was done with a simple machine and you can see how this

machine and you can see how this

machine and you can see how this technology will transform the music

technology will transform the music

technology will transform the music industry by not only helping composers

industry by not only helping composers

industry by not only helping composers but just fundamentally writing the whole

but just fundamentally writing the whole

but just fundamentally writing the whole scores it will change the industry

scores it will change the industry

scores it will change the industry probably more than auto-tune has now as

probably more than auto-tune has now as

probably more than auto-tune has now as I mentioned all of this was actually

I mentioned all of this was actually

I mentioned all of this was actually built and create created on a simple

built and create created on a simple

built and create created on a simple home desktop with relatively simple data

home desktop with relatively simple data

home desktop with relatively simple data and a clear objective the examples are

and a clear objective the examples are

and a clear objective the examples are simple but its implications are

simple but its implications are

simple but its implications are incredible

incredible

incredible just as computers have changed the way

just as computers have changed the way

just as computers have changed the way we live our lives fundamentally so is

we live our lives fundamentally so is

we live our lives fundamentally so is machine learning nap changing the way we

machine learning nap changing the way we

machine learning nap changing the way we interact with those computers and with

interact with those computers and with

interact with those computers and with each other I hope you enjoyed this

each other I hope you enjoyed this

each other I hope you enjoyed this presentation and there everything else

presentation and there everything else

presentation and there everything else left to say is thanks see you and I hope

left to say is thanks see you and I hope

left to say is thanks see you and I hope you enjoyed it

you enjoyed it

you enjoyed it [Applause]

”

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *