
are you there I am here listening to you
what is your name as my buddy
Shakespeare once said what’s in a name
but you can call me Tommy
what is your favorite book I love
reading the robots of dawn by Isaac
Asimov what is your favorite song I love
singing want to hear some no thank you
can you tell me a joke I love helping
humans can you tell me a joke there are
two ways to write our free programs and
but I want to take you a little bit on a
journey of computing intelligence across
time if we think of a story of computing
we identify three eras the first era
started with tabulating I was just
counting things counting people counting
money
those were the times of the single
purpose computers they were built
without one function specifically fifty
years later we started what we called
the programmable era computers that can
follow a set of instructions or programs
this around lasted up till today even
the computer that you have with you in
your pocket your phone is using this
type of computing but six years ago we
started what we call the cognitive
era many of you saw in the US television
show a jeopardy Watson playing against
the best players in the world
you remember what was the outcome of
that show yes we defeated the best
players in the world but it was not easy
playing jeopardy was very challenging
task because human language is full
ambiguity humor in context so this
system was able to answer questions in 3
to 6 seconds about anything music poetry
it was able to correctly answer the
questions 80% of the time if we play 50
percent of the opportunities so this
system we call it Watson start the
cognitive system era well how does it
work let me start by telling you what
Watson is not it’s not how 9000 in 2001
Space Odyssey now I haven’t seen John
Connor running around it’s not the
beginning of the Skynet it’s not an
bear with me here I’m going to get
technical Watson is a cognitive system
that uses natural language understanding
and machine learning to draw insights
what kind of insight business insights
from large amounts of data and we call
this data big data is divided into
unstructured data think of your tweets
your Facebook entries in structured data
and how do we do this
understand reason and learn in what I
mean for understand is pretty much
dealing with a human condition of
language think about it
we humans express through language we
spread our ideas we talk to each other
even in this moment there is a little
voice in your head listening to what I’m
saying and the understanding is trying
to find these symbols the signals these
patterns of meaning not talking about
the language just spoken language like
English or Spanish I’m talking about the
language of images or text so
understanding those natural languages
then reason it because once I have the
answer what do I do with it
for instance if you remember in jeopardy
we train the system with 200 million
pages of content thus equivalent or
having all the books in the Library of
Congress and memory to answer these
questions for instance what is the
capital of Texas we go through the whole
Wikipedia and we have multiple answers
until we found the one that we think was
a valid answer that is reasoning in the
type of reasoning that we do it’s called
inductive reasoning all these other
players in the area use the same thing
when you talk to your phone talk your
assistant at home you’re using that type
of reasoning and then cognitive systems
learn
if it was just matter of programming the
rules of grammar that will be too easy
we need to teach these systems we don’t
program them and by teaching them we
make them more intelligence another
thing that I want to bring up today have
been a lot of talk on the under media on
the news about these systems this robots
taking away jobs into me is more like a
displacement think about when we have
the car displacing the horse a lot of
people lost their jobs but they went and
create new jobs or the Industrial
Revolution these cognitive systems are
creating a partnership between humans
and computers in this partnership is
creating enhance and human expertise why
because we humans are good at certain
things we’re good at compassion
abstraction even common sense like don’t
drive into a white trailer you will die
computers on the other hand are good at
identifying patterns they do machine
learning they isolate the answers so
this one airship is what I my talk is
all about for that I would like to give
you some practical examples or some of
the recent use cases that we have been
working on and I’m going to start in the
bottom right with what we call cognitive
cooking we have an application called
chip Watson that read a whole bunch of
recipes in in the chemistry of the
ingredients in those recipes and in
conjunction with your appliances can
find out what is in your refrigerator
think about
it’s Thursday nya have nothing to cook
you talk to your refrigerator talk to
the oven say what like what did I make
for dinner this intelligent system will
say well based on what you have you can
make this that’s the power of cognitive
going clockwise we have cognate ice
connectors was a huge success last
Christmas
this is a toy that interact with little
children and answer questions like why
is the sky blue
tell me our knock-knock joke and even
where do babies come from cognitive
recently we announced a partnership with
a u.s. car manufacturer you remember
that show from the 80s the Nightrider
that you talk to your car we have that
today so when you’re driving home after
talking to chip Watson and you need an
ingredient for your recipe you can tell
us a car where’s the nearest grocery
store we’re gonna put gas for my car you
have that kind of dialogue natural
language understanding but my last
example something very close to me
health care growing up in Latin America
I notice how difficult it was to gain
access to good health care only 5% of
the people have access to good
healthcare and with health care we tell
usually our visitors on the lab that
Watson doesn’t understand the science of
medicine he understand the language of
Medicine you cannot tell us what’s wrong
with this person but it can tell us what
are the multiple uses of certain drugs
or two types of conditions we estimate
that doctors are not keeping up in their
fields there is so much data
unstructured data been published every
week that they don’t only spend like two
hours a week reading what’s the latest
on the field therefore they’re making
decisions based on information that they
learned 10 15 years ago perhaps not the
most accurate to say people’s lives for
instance recently I live in Chile and we
held this patient this doctor trying to
identify a disease for a sick child in
the Antarctica the Patagonia this doctor
use our technology to identify what was
wrong with this child and was able to
save this life and he told us there is
nothing people will neither in the world
to save a kid’s life and to me does the
a partnership between humans and
computer and that’s an idea worth spreading thank you
what is your name as my buddy
Shakespeare once said what’s in a name
but you can call me Tommy
what is your favorite book I love
reading the robots of dawn by Isaac
Asimov what is your favorite song I love
singing want to hear some no thank you
can you tell me a joke I love helping
humans can you tell me a joke there are
two ways to write our free programs and
but I want to take you a little bit on a
journey of computing intelligence across
time if we think of a story of computing
we identify three eras the first era
started with tabulating I was just
counting things counting people counting
money
those were the times of the single
purpose computers they were built
without one function specifically fifty
years later we started what we called
the programmable era computers that can
follow a set of instructions or programs
this around lasted up till today even
the computer that you have with you in
your pocket your phone is using this
type of computing but six years ago we
started what we call the cognitive
era many of you saw in the US television
show a jeopardy Watson playing against
the best players in the world
you remember what was the outcome of
that show yes we defeated the best
players in the world but it was not easy
playing jeopardy was very challenging
task because human language is full
ambiguity humor in context so this
system was able to answer questions in 3
to 6 seconds about anything music poetry
it was able to correctly answer the
questions 80% of the time if we play 50
percent of the opportunities so this
system we call it Watson start the
cognitive system era well how does it
work let me start by telling you what
Watson is not it’s not how 9000 in 2001
Space Odyssey now I haven’t seen John
Connor running around it’s not the
beginning of the Skynet it’s not an
bear with me here I’m going to get
technical Watson is a cognitive system
that uses natural language understanding
and machine learning to draw insights
what kind of insight business insights
from large amounts of data and we call
this data big data is divided into
unstructured data think of your tweets
your Facebook entries in structured data
and how do we do this
understand reason and learn in what I
mean for understand is pretty much
dealing with a human condition of
language think about it
we humans express through language we
spread our ideas we talk to each other
even in this moment there is a little
voice in your head listening to what I’m
saying and the understanding is trying
to find these symbols the signals these
patterns of meaning not talking about
the language just spoken language like
English or Spanish I’m talking about the
language of images or text so
understanding those natural languages
then reason it because once I have the
answer what do I do with it
for instance if you remember in jeopardy
we train the system with 200 million
pages of content thus equivalent or
having all the books in the Library of
Congress and memory to answer these
questions for instance what is the
capital of Texas we go through the whole
Wikipedia and we have multiple answers
until we found the one that we think was
a valid answer that is reasoning in the
type of reasoning that we do it’s called
inductive reasoning all these other
players in the area use the same thing
when you talk to your phone talk your
assistant at home you’re using that type
of reasoning and then cognitive systems
learn
if it was just matter of programming the
rules of grammar that will be too easy
we need to teach these systems we don’t
program them and by teaching them we
make them more intelligence another
thing that I want to bring up today have
been a lot of talk on the under media on
the news about these systems this robots
taking away jobs into me is more like a
displacement think about when we have
the car displacing the horse a lot of
people lost their jobs but they went and
create new jobs or the Industrial
Revolution these cognitive systems are
creating a partnership between humans
and computers in this partnership is
creating enhance and human expertise why
because we humans are good at certain
things we’re good at compassion
abstraction even common sense like don’t
drive into a white trailer you will die
computers on the other hand are good at
identifying patterns they do machine
learning they isolate the answers so
this one airship is what I my talk is
all about for that I would like to give
you some practical examples or some of
the recent use cases that we have been
working on and I’m going to start in the
bottom right with what we call cognitive
cooking we have an application called
chip Watson that read a whole bunch of
recipes in in the chemistry of the
ingredients in those recipes and in
conjunction with your appliances can
find out what is in your refrigerator
think about
it’s Thursday nya have nothing to cook
you talk to your refrigerator talk to
the oven say what like what did I make
for dinner this intelligent system will
say well based on what you have you can
make this that’s the power of cognitive
going clockwise we have cognate ice
connectors was a huge success last
Christmas
this is a toy that interact with little
children and answer questions like why
is the sky blue
tell me our knock-knock joke and even
where do babies come from cognitive
recently we announced a partnership with
a u.s. car manufacturer you remember
that show from the 80s the Nightrider
that you talk to your car we have that
today so when you’re driving home after
talking to chip Watson and you need an
ingredient for your recipe you can tell
us a car where’s the nearest grocery
store we’re gonna put gas for my car you
have that kind of dialogue natural
language understanding but my last
example something very close to me
health care growing up in Latin America
I notice how difficult it was to gain
access to good health care only 5% of
the people have access to good
healthcare and with health care we tell
usually our visitors on the lab that
Watson doesn’t understand the science of
medicine he understand the language of
Medicine you cannot tell us what’s wrong
with this person but it can tell us what
are the multiple uses of certain drugs
or two types of conditions we estimate
that doctors are not keeping up in their
fields there is so much data
unstructured data been published every
week that they don’t only spend like two
hours a week reading what’s the latest
on the field therefore they’re making
decisions based on information that they
learned 10 15 years ago perhaps not the
most accurate to say people’s lives for
instance recently I live in Chile and we
held this patient this doctor trying to
identify a disease for a sick child in
the Antarctica the Patagonia this doctor
use our technology to identify what was
wrong with this child and was able to
save this life and he told us there is
nothing people will neither in the world
to save a kid’s life and to me does the
a partnership between humans and
computer and that’s an idea worth spreading thank you
Please follow and like us: