
[Music]
when was the last time you needed to go
to a bank walk to a bank to check your
check your bank statement when was the
last time that you entered your car and
you used a paper map for navigation when
was the last time you had a question you
needed to walk to a library open an
encyclopedia to find the answer if you
don’t remember this the answer to these
questions this is because you are using
technology we are all using technology
to save time to do things easier
technology is changing our lives there
has been lots of technological
revolutions which have changed lives of
many many people billions of people and
the most recent one is the AI revolution
and I’m going to talk today about the AI
revolution and the Google assistant the
project I am working on so Google
assistant is amber excited about AI
revolution because assistant is one of
the cutting-edge implementations of what
we will see in this AI revolution so the
next 10 years is going to be will be
going to be living in a world which is
AI first this is the AI revolution we’re
talking about but before going into that
this revolution is very similar in
nature to many other revolutions which
happened in the past 100 years these are
all big technological revolutions which
changed lives of billions and billions
of people it made lots of companies to
disappear it made some new companies to
appear and if you think it’s saved lots
of time for it all of these save lots of
times for us
for example plane before before plane
people needed to spend six weeks to go
from Europe to us on ship if you have
money to do that so you’re basically
wasting six weeks instead of ten hours
just on the ship to go to go from this
point to the other point and so on
there’s so many of such such time-saving
scenarios happening with these
revolutions
if you move fast forward 20 years ago
web and PC came this is from what moment
we actually can sit at home and answer
find the answer to questions we can see
our bank account at home we can check
the weather we don’t need to wait for
the next news to hear the weather and we
can buy things so this revolution
changed our life really really in a big
way just think about the moment which
web and internet was not there it’s only
20 years ago now there was another
revolution which happened 10 years after
and that’s the mobile revolution that’s
when everything that’s a big huge clunky
PC could do and more you can do it with
these Mobile’s which is in our pocket
and it’s with us everywhere it’s moving
it has it’s even more powerful because
it has a microphone it has a camera and
it’s more personal you can take pictures
you can share photos and you can call
and everything else so it’s really
becoming in a more powerful version of
the previous revolution and now if you
think the mobile revolution is in that
scale bigger than the previous
revolution of the web and the reason is
that because the numbers are everything
is really one order of magnitude bigger
five billion people will have phones
connected by 2020 this includes lots of
countries in Africa also India and every
everywhere else this is really much
bigger than the scale of the previous
revolutions while this revolution is
still happening the next revolution is
going to make mobile revolution even
bigger and that’s the eye first
revolution I was talking about this is
from this is the moment where the phones
are becoming smarter smartphones now
finally become smart they start to
understand they start to understand
language they can see things understand
what we are seeing you can they can
understand the context of what I’m
talking to them about and they can
understand at what moment it should help
me and how so this is going to be very
interesting era where you can talk to
phones but not also to phones to all
different types of devices Google if you
have like TV car or watch and many many
more with the types of devices will be
devices that you can talk to them in a
consistent way and get help and
assistance from them so assistant is a
conversation between you and your Google
or between you and your assistant which
is helping you to get things done
throughout throughout your day I want to
insist on the word conversation here
because conversation is the most
universal interface human beings have
between themselves we all can talk to
each other from the moment we work it to
the moment with your older we can always
try to have express ourselves using this
conversation independent of age race
literacy so if you solve the problem of
conversation with machine if you can
talk to machines then I can talk to
everything I can talk to that chair I
can talk to this phone I can talk to
that device there and this is the same
experience when we’re talking and in the
conversation with all of this so I think
the best way to actually show where we
are moving is to show that with live
demos and that’s why I have this device
here which is going to project my phone
and so in the arrest of in the rest of
my talk now I’m going to show you some
of the different capabilities that the
assistant can do today and some of the
capabilities which is coming in the
future so let’s get started my password
you can see it you could get it to I
guess so the first thing I can I want to
show you is to show you how the
assistant should be answering questions
we’re talking about the future assistant
an ultimate assistant so let’s see can
you please tell me how is the weather
going to be in Lugano tomorrow so first
please note that everything I’m showing
you these are live demos they might fail
and you might have fun it’s risky for me
second note that the way I asked this
question I asked a natural language
question like can you please tell me
before when we were entering the search
box that on your PC you enter
tomorrow Lugano but now you use this
long sentence which is like how the
humans talk so this is this means the
natural language understanding and now
we’re reaching the point that we can
understand through machine learning and
artificial intelligence that these mean
the same thing but these are a task
which might come day to day now when we
talk about the assistant for billions of
people we’re thinking about lots of
different types of questions not only
day-to-day questions so we have to work
on really making the range of the
answers that assistant is capable of
saying to be hundreds of millions of
different questions
now I’m going to show you one example of
a specific question to show you an idea
of what’s possible to do show me a list
of Rights in Europa Park okay so this is
the right right and this is showing that
this is a live demo show me a list of
Rights in Europe a park Europa parks
rides include Bluefire Silver Star Road
and timber coaster and others what is
the height restriction for blue fire you
must be at least 1.31 meters tall to
ride blue fire so we’re answering even
very specific questions about a
particular ride in a particular
attraction park and this is capabilities
of today and it’s going to grow and grow
now let me show you another example of
when you try to ask a question which is
more complex and this is like a question
which is a longer sentence I want to
show you this kind of state of art of
natural language understanding that
we’re building what is the name of the
movie where Tom Cruise acts in it and he
plays pool and while playing pool he
dances
here are some information about the
color of money
former pool hustler Felsen decides he
wants to return to the game by taking a
pupil so as you see this is really like
getting closer and closer that you don’t
need to think how to ask your questions
you should just say it that way and it
should work now let’s show an example of
how this voice interface would work with
Street View for example again natural
language and voice recognition together
hey I would like to be on the top of
eiffel tower can you please bring me
there now so we don’t go totally there
but at least we can have the Street View
experience so so now think about think
about when this voice interface is
connected to all different types of
applications and services and then you
can just jump directly but just talking
to the right points that you want okay
so this is the first part of my demos
which is about answering things getting
things done but these are about
answering things about the world and I
want to now show you some examples of
personal aspect of the assistant because
this is a personal assistant it’s my
assistant and it should also know me and
do things for me for example my
assistant should be able to do how is
the traffic from Zurich Airport to my
home so it should take 43 minutes so my
assistant knows where my home is no sir
my work is my calendar and this is all
again thanks to machine learning it’s
automatic it’s not that I explicitly
said this email in my gmail is a flight
or that’s my home address now you can
also in the future and already from some
now start to teach your assistant for
example how is my team doing Barcelona
second group so you start to know me
during this talk
the the way that the assistant knows my
what my favorite team is what is my
favorite team is that explicitly
actually I said to their sis and I like
Barcelona I said like few months back
that think about how how much next level
of smartness is that that you can just
teach things let me show you one of the
upcoming types of such changes when the
weather is more than 25 degrees I can
swim in the lake of Zurich so I just
taught something to my assistant on this
stage and now I’m going to check whether
it understood it or not can I swim in
the lake observe this weekend
[Applause]
okay now the assistant is not only about
talking and conversation it’s also about
seeing things humans see things and if
you want to have natural conversations
you expect that your assistant can
understand what’s in the pictures now
I’m going to show you how my assistant
by applying machine learning and
artificial intelligence on pictures can
try to understand what’s in the picture
and then combine that with voice and
show you what that experience would look
like show me pictures of me and my
nephew drinking tea together so I never
attacked any picture about drinking tea
or about me this picture that I’m
showing to you now this is all automatic
this is all machine learning trying to
understand that this is what’s what is
happening in that picture and who are
the people in that picture so let me
show you another one show me my pictures
of mountains so these are different
mountains picture that I have let’s take
on a one particular one anyone knows
which one is this okay young forever
that’s a good answer but let’s ask
assistant what is the name of this
mountain and how tall is it so again my
assistant understands what’s in the
picture and I can talk to him or it or
her about what’s on the picture so let’s
now move to other part of my demos which
is about context humans when do we talk
to each other we always use context to
understand each other and so that the
conversation is more natural now we’re
adding context to machine conversations
also to make the as the conversations
become much more natural let me start by
an example like show me pictures of
Thomas without any context perhaps
that’s the bestest someone can do now if
I say
Bionic team roster the roster for FC
Bayern Munich includes Robert
Lewandowski James Rodriguez Arjen
Robinson others so now you see that
there is someone here in this team
called Thomas Muller so if that’s an
exact same question show me pictures of
Thomas so now we really understand what
the meaning of Thomas here is Thomas
Muller
this means contextual smartness now let
me show you a context in conversation
where is the Empire State Building the
address for Empire State Building is
Empire State Building 355th Avenue so if
I want to see pictures I want to see
pictures and then the assistant will
understand that the meaning of that
sentence is pictures of the Empire State
Building
you can continue that such type of
conversation how tall is it who built it
starik corporation when construction for
the Empire State Building started in
show me Italian restaurants around there
we are in Lugano
I found a few places call the first one
so note that I’m having a conversation
reader machine this is not a human sorry
and I said only one time in Peyer State
Building maybe if you don’t call
[Applause]
I called them many times from Europe at
the night time
so but this is this is a machine where
you can continuously have a conversation
and not needing to repeat yourself at
each step the sentence is understood
thanks to the context the context is
helping that the conversations become
more natural so for my last demo last
example I want to you notice that speech
recognition failed one time and this is
actually a real estate of art it has
improved a lot but we also have been
working on making a speech recognition
improved in a noisy environment for that
I need your help so we’ll play again
we’re something together which means
that I will ask at some moment to you to
make as much noise as you can you can
scream shout and whatever use your
creativity and I will try to ask my
assistant for example what is when is my
team playing next Barcelona and try to
show you that still should normally work
inside setup your goal should be to make
me say fail please mate and we can also
cut my mic when I’m doing that so please
please make as much noise as you can
[Applause]
so how can you fail so we can go back to
the slides and so that I wrap up my talk
so what you have what you have seen in
these demos is kind of the state of art
art of where we are we’re so excited of
what can be done there is much much more
because the technologies are getting
ready to be put together and many types
of assistive scenarios can be helping us
throughout the day there are lots of
different types of technologies the
speech recognition image recognition
context personalization and many more
which is coming together to make these
demos possible now the AI revolution is
not only about the assistant the
assistant is only one of those one of
the parts of the AI revolution we have
the health AI in health in agriculture
smart cities smart cars there is lot
more happening in AI the world is really
becoming AI first but but even when you
only think about the age of assistance
and what assistant is we could able to
do I’m super excited about this because
I kind of feel that going back to the
initial initial part of the talk we are
also in the business of saving time
think about without having this how much
time you need to do to spend for doing
the other types of examples which I
showed you how much time you needed to
spend to find those photos my photos
with my nephew now this makes me even
even more excited because saving time is
something big
if you save time from people’s life
people have more time to spend on things
which really matter thank you
you
when was the last time you needed to go
to a bank walk to a bank to check your
check your bank statement when was the
last time that you entered your car and
you used a paper map for navigation when
was the last time you had a question you
needed to walk to a library open an
encyclopedia to find the answer if you
don’t remember this the answer to these
questions this is because you are using
technology we are all using technology
to save time to do things easier
technology is changing our lives there
has been lots of technological
revolutions which have changed lives of
many many people billions of people and
the most recent one is the AI revolution
and I’m going to talk today about the AI
revolution and the Google assistant the
project I am working on so Google
assistant is amber excited about AI
revolution because assistant is one of
the cutting-edge implementations of what
we will see in this AI revolution so the
next 10 years is going to be will be
going to be living in a world which is
AI first this is the AI revolution we’re
talking about but before going into that
this revolution is very similar in
nature to many other revolutions which
happened in the past 100 years these are
all big technological revolutions which
changed lives of billions and billions
of people it made lots of companies to
disappear it made some new companies to
appear and if you think it’s saved lots
of time for it all of these save lots of
times for us
for example plane before before plane
people needed to spend six weeks to go
from Europe to us on ship if you have
money to do that so you’re basically
wasting six weeks instead of ten hours
just on the ship to go to go from this
point to the other point and so on
there’s so many of such such time-saving
scenarios happening with these
revolutions
if you move fast forward 20 years ago
web and PC came this is from what moment
we actually can sit at home and answer
find the answer to questions we can see
our bank account at home we can check
the weather we don’t need to wait for
the next news to hear the weather and we
can buy things so this revolution
changed our life really really in a big
way just think about the moment which
web and internet was not there it’s only
20 years ago now there was another
revolution which happened 10 years after
and that’s the mobile revolution that’s
when everything that’s a big huge clunky
PC could do and more you can do it with
these Mobile’s which is in our pocket
and it’s with us everywhere it’s moving
it has it’s even more powerful because
it has a microphone it has a camera and
it’s more personal you can take pictures
you can share photos and you can call
and everything else so it’s really
becoming in a more powerful version of
the previous revolution and now if you
think the mobile revolution is in that
scale bigger than the previous
revolution of the web and the reason is
that because the numbers are everything
is really one order of magnitude bigger
five billion people will have phones
connected by 2020 this includes lots of
countries in Africa also India and every
everywhere else this is really much
bigger than the scale of the previous
revolutions while this revolution is
still happening the next revolution is
going to make mobile revolution even
bigger and that’s the eye first
revolution I was talking about this is
from this is the moment where the phones
are becoming smarter smartphones now
finally become smart they start to
understand they start to understand
language they can see things understand
what we are seeing you can they can
understand the context of what I’m
talking to them about and they can
understand at what moment it should help
me and how so this is going to be very
interesting era where you can talk to
phones but not also to phones to all
different types of devices Google if you
have like TV car or watch and many many
more with the types of devices will be
devices that you can talk to them in a
consistent way and get help and
assistance from them so assistant is a
conversation between you and your Google
or between you and your assistant which
is helping you to get things done
throughout throughout your day I want to
insist on the word conversation here
because conversation is the most
universal interface human beings have
between themselves we all can talk to
each other from the moment we work it to
the moment with your older we can always
try to have express ourselves using this
conversation independent of age race
literacy so if you solve the problem of
conversation with machine if you can
talk to machines then I can talk to
everything I can talk to that chair I
can talk to this phone I can talk to
that device there and this is the same
experience when we’re talking and in the
conversation with all of this so I think
the best way to actually show where we
are moving is to show that with live
demos and that’s why I have this device
here which is going to project my phone
and so in the arrest of in the rest of
my talk now I’m going to show you some
of the different capabilities that the
assistant can do today and some of the
capabilities which is coming in the
future so let’s get started my password
you can see it you could get it to I
guess so the first thing I can I want to
show you is to show you how the
assistant should be answering questions
we’re talking about the future assistant
an ultimate assistant so let’s see can
you please tell me how is the weather
going to be in Lugano tomorrow so first
please note that everything I’m showing
you these are live demos they might fail
and you might have fun it’s risky for me
second note that the way I asked this
question I asked a natural language
question like can you please tell me
before when we were entering the search
box that on your PC you enter
tomorrow Lugano but now you use this
long sentence which is like how the
humans talk so this is this means the
natural language understanding and now
we’re reaching the point that we can
understand through machine learning and
artificial intelligence that these mean
the same thing but these are a task
which might come day to day now when we
talk about the assistant for billions of
people we’re thinking about lots of
different types of questions not only
day-to-day questions so we have to work
on really making the range of the
answers that assistant is capable of
saying to be hundreds of millions of
different questions
now I’m going to show you one example of
a specific question to show you an idea
of what’s possible to do show me a list
of Rights in Europa Park okay so this is
the right right and this is showing that
this is a live demo show me a list of
Rights in Europe a park Europa parks
rides include Bluefire Silver Star Road
and timber coaster and others what is
the height restriction for blue fire you
must be at least 1.31 meters tall to
ride blue fire so we’re answering even
very specific questions about a
particular ride in a particular
attraction park and this is capabilities
of today and it’s going to grow and grow
now let me show you another example of
when you try to ask a question which is
more complex and this is like a question
which is a longer sentence I want to
show you this kind of state of art of
natural language understanding that
we’re building what is the name of the
movie where Tom Cruise acts in it and he
plays pool and while playing pool he
dances
here are some information about the
color of money
former pool hustler Felsen decides he
wants to return to the game by taking a
pupil so as you see this is really like
getting closer and closer that you don’t
need to think how to ask your questions
you should just say it that way and it
should work now let’s show an example of
how this voice interface would work with
Street View for example again natural
language and voice recognition together
hey I would like to be on the top of
eiffel tower can you please bring me
there now so we don’t go totally there
but at least we can have the Street View
experience so so now think about think
about when this voice interface is
connected to all different types of
applications and services and then you
can just jump directly but just talking
to the right points that you want okay
so this is the first part of my demos
which is about answering things getting
things done but these are about
answering things about the world and I
want to now show you some examples of
personal aspect of the assistant because
this is a personal assistant it’s my
assistant and it should also know me and
do things for me for example my
assistant should be able to do how is
the traffic from Zurich Airport to my
home so it should take 43 minutes so my
assistant knows where my home is no sir
my work is my calendar and this is all
again thanks to machine learning it’s
automatic it’s not that I explicitly
said this email in my gmail is a flight
or that’s my home address now you can
also in the future and already from some
now start to teach your assistant for
example how is my team doing Barcelona
second group so you start to know me
during this talk
the the way that the assistant knows my
what my favorite team is what is my
favorite team is that explicitly
actually I said to their sis and I like
Barcelona I said like few months back
that think about how how much next level
of smartness is that that you can just
teach things let me show you one of the
upcoming types of such changes when the
weather is more than 25 degrees I can
swim in the lake of Zurich so I just
taught something to my assistant on this
stage and now I’m going to check whether
it understood it or not can I swim in
the lake observe this weekend
[Applause]
okay now the assistant is not only about
talking and conversation it’s also about
seeing things humans see things and if
you want to have natural conversations
you expect that your assistant can
understand what’s in the pictures now
I’m going to show you how my assistant
by applying machine learning and
artificial intelligence on pictures can
try to understand what’s in the picture
and then combine that with voice and
show you what that experience would look
like show me pictures of me and my
nephew drinking tea together so I never
attacked any picture about drinking tea
or about me this picture that I’m
showing to you now this is all automatic
this is all machine learning trying to
understand that this is what’s what is
happening in that picture and who are
the people in that picture so let me
show you another one show me my pictures
of mountains so these are different
mountains picture that I have let’s take
on a one particular one anyone knows
which one is this okay young forever
that’s a good answer but let’s ask
assistant what is the name of this
mountain and how tall is it so again my
assistant understands what’s in the
picture and I can talk to him or it or
her about what’s on the picture so let’s
now move to other part of my demos which
is about context humans when do we talk
to each other we always use context to
understand each other and so that the
conversation is more natural now we’re
adding context to machine conversations
also to make the as the conversations
become much more natural let me start by
an example like show me pictures of
Thomas without any context perhaps
that’s the bestest someone can do now if
I say
Bionic team roster the roster for FC
Bayern Munich includes Robert
Lewandowski James Rodriguez Arjen
Robinson others so now you see that
there is someone here in this team
called Thomas Muller so if that’s an
exact same question show me pictures of
Thomas so now we really understand what
the meaning of Thomas here is Thomas
Muller
this means contextual smartness now let
me show you a context in conversation
where is the Empire State Building the
address for Empire State Building is
Empire State Building 355th Avenue so if
I want to see pictures I want to see
pictures and then the assistant will
understand that the meaning of that
sentence is pictures of the Empire State
Building
you can continue that such type of
conversation how tall is it who built it
starik corporation when construction for
the Empire State Building started in
show me Italian restaurants around there
we are in Lugano
I found a few places call the first one
so note that I’m having a conversation
reader machine this is not a human sorry
and I said only one time in Peyer State
Building maybe if you don’t call
[Applause]
I called them many times from Europe at
the night time
so but this is this is a machine where
you can continuously have a conversation
and not needing to repeat yourself at
each step the sentence is understood
thanks to the context the context is
helping that the conversations become
more natural so for my last demo last
example I want to you notice that speech
recognition failed one time and this is
actually a real estate of art it has
improved a lot but we also have been
working on making a speech recognition
improved in a noisy environment for that
I need your help so we’ll play again
we’re something together which means
that I will ask at some moment to you to
make as much noise as you can you can
scream shout and whatever use your
creativity and I will try to ask my
assistant for example what is when is my
team playing next Barcelona and try to
show you that still should normally work
inside setup your goal should be to make
me say fail please mate and we can also
cut my mic when I’m doing that so please
please make as much noise as you can
[Applause]
so how can you fail so we can go back to
the slides and so that I wrap up my talk
so what you have what you have seen in
these demos is kind of the state of art
art of where we are we’re so excited of
what can be done there is much much more
because the technologies are getting
ready to be put together and many types
of assistive scenarios can be helping us
throughout the day there are lots of
different types of technologies the
speech recognition image recognition
context personalization and many more
which is coming together to make these
demos possible now the AI revolution is
not only about the assistant the
assistant is only one of those one of
the parts of the AI revolution we have
the health AI in health in agriculture
smart cities smart cars there is lot
more happening in AI the world is really
becoming AI first but but even when you
only think about the age of assistance
and what assistant is we could able to
do I’m super excited about this because
I kind of feel that going back to the
initial initial part of the talk we are
also in the business of saving time
think about without having this how much
time you need to do to spend for doing
the other types of examples which I
showed you how much time you needed to
spend to find those photos my photos
with my nephew now this makes me even
even more excited because saving time is
something big
if you save time from people’s life
people have more time to spend on things
which really matter thank you
you
Please follow and like us: