A_W90CQbSdE

[Music]
so hello there and welcome to another
tutorial my name is Tam Baki and this
time we're going to be going over IBM
Watson a few Services it provides and of
course today I'm with my mentor Rob High
here at again IBM interconnect 2017 as
you saw in the last video was with John
con uh talking about the TJ bot and iot
uh with the TJ bot that actually Rob I
here actually gifted to me on stage
during our ey talk uh and so uh we're
going to be of course creating many more
videos with that but for now we're
talking about Watson and doing a quick
Q&A session about some of your
frequently asked questions about Watson
but before we begin Rob would you like
to introduce yourself okay I'm Rob High
I'm the CTO for IBM Watson I've had the
privilege for the last several years to
be able to guide the technical strategy
for IBM's uh work in cognitive Computing
that is great all right so let's begin
with one of the most asked questions
that my subscribers have for you which
is really what's the difference between
Ai and cognitive uh versus cognitive
Computing or sorry traditional Computing
tradition yes what's really the
difference here what what makes
cognitive Computing so special so one of
the key things that cognitive Computing
does is it really begins to look at the
world the same way that we do you know
if you really kind of look at our world
around us it's full of signals you know
the light is a set of of uh of uh of
either particles or waves depending on
how you want to look at them that have a
certain frequency to them that give them
color give them brightness um um that's
just one thing you know our voices are
carried as as audio signals um the
things that we feel the things that we
um are able to touch the things that we
smell all these things are signals and
what's really amazing about human beings
is essentially what our brain is capable
of doing very very well is figuring at
the patterns of meaning in all those
signals and that's true whether these
are physical signals the signals that we
associate with physics and chemistry uh
and biology or even more softer signals
the signals of emotion the signals of
intent the signals of imagination all
those are signals as well and one of the
things that we've tried to do in the
cognitive Computing system is really
kind of Follow That same approach to
understanding the world that we live in
by just picking up all these signals and
then looking at the patterns of meaning
that we can find in the cognitive system
similar to the way that humans do as
well and that's just completely
different than what we've done in the
past where you know to build a program
you had to write the logic of every
mathematical turn that was going to
occur within um that particular
algorithm Perfect all right so now that
we've gone over really the difference
between cognitive and traditional
Computing and what makes cognitive so
special uh would you like to touch upon
a little bit about how cognitive is
assisting humans so one of the things
that we really focus on is not just
simply recreating the human mind uh in
part because frankly the human mind is
really amazing it's it's you know
mysterious um but yet it also is very
good at doing certain things instead we
look at is how do we augment human Minds
how do we take what human beings are
really good at and and extend it um give
it more strength give us more power by
be able to take our mental capabilities
and using a cog system to do things that
we could not naturally do on our own
perfect uh so that's what uh What uh is
IBM you know I mean we've all heard that
IBM is working with cogni Computing and
Watson is a System created by IBM uh
that can understand it can reason it can
learn it can make decisions for itself
based off of learning from these signals
and these inputs that it's being given
in order to give output that the user
would want uh and so now though uh
another question how can developers
around the world actually use Watson's
capabilities in their apps so one of the
things that we were able to do over the
years of experience with Watson is begin
to identify what were sort of the key
building blocks that were necessary to
construct a really useful cognitive
application so we took each of these
building blocks and we put them in the
cloud given them an API and made those
apis available through the cloud for
anybody to use so really building an
application is as simple as identifying
the service that you want to make use of
writing the code all these code all
these a apis are available as restful
services so if you know how to build a
Json structure as you have done in many
of your videos um you can make use of
the services there's a wide variety of
them available so all you have to do is
go to ibm.com BL miix navigate into the
Watson Zone and you're going to find
those services and be able to pick them
up and make use of them exactly so
Watson has things like speech Vision
natural language processing and you know
General reasoning and logic right that's
right and the key ones are things like
conversations right because that's a
very common use case for people today is
they want to build what some people
refer to as jaat chat Bots but I like to
think of one it's not just a chat but
really a conversation right trying to
build an application that walks people
through their conversational Bots
exactly got okay yeah there we go so in
fact if you'd like to find out more
about how you can build these
conversational Bots and use Watson's
apis there will be a link to the TM
teachers Watson playlist down below and
then you you can you know go ahead and
check out how you can use those Waton
services in your applications but now
another question is well what's the
underlying technology behind Watson what
types of algorithms do you do you use
like what what's going on behind it so
we've actually gone through a transition
over the years and when we first began
we were heavily focused on using classic
machine learning algorithms um you know
things like uh logistical regression uh
sbms um even linear regression in basan
we've more recently evolved over to
making better use of deep learning
algorithms um convolutional neuron Nets
Uh current neuron Nets or even more
exotic variations of uh of long
short-term uh memory um algorithms um
what we call
lstms um and usually in combination it's
not usually just a single algorithm it's
usually different algorithms that we're
leveraging to pick up on different kind
the signals and find those patterns
exactly uh so that's what goes behind
Watson but now another question is well
uh and this will be on my YouTube
channel soon but I've actually sometimes
you know gone ahead and created um I
guess you could say mixes of services to
create one extremely powerful service
like conversation enhanced by Tone
analyzer by say for example uh actually
combining the conversation and tone
analyzer by whenever you send a uh a uh
message to conversation it'll actually
put the tone of that message into a
context variable so that you can of
course access uh access the tone and
give personalized responses do you think
or do you believe Watson uh will I guess
you could say start to support like
mixes of these apis built in soon
mashing them up together in in both uh
structured and unstructured ways um the
answer is very likely you know there's
still choices that we have to make as we
move forward making sure that people can
Mash these things up in the style that
you describe yet at the same time keep
keep each individual Service as simple
as possible and of course that's a
balance we're always trying to measure
how complicated we're making things so
that we don't and we can keep things to
the point where people can do everything
they really want to do without being
bogged down so yeah you're going to see
more and more at least patterns of
integration where you can take something
like a conversation integrate it with
tone analyzer integrate with personality
insights integrate it with other
services like the weather to be able to
understand you know what the weather is
like outside where you are or where
you're going um what the environment is
like the place that you're at you know
if that's in your home or in your car or
in your office what else is happening in
that room that would also inform how the
conversational system should interpret
what you need to be able to anticipate
that need and be able to offer you
choices in that context that are
meaningful to you or relevant to you in
that environment so yes absolutely
you're going to see lots of combinations
of this kind that is perfect I can
imagine that being being very powerful
for a lot of developers applications now
one more thing how can developers give
feedback to the Watson team if they want
to say incorporate a new feature give
feedback for an existing service so we
have a variety of uh social channels
that we invite people to come and join
us of course we have GitHub slack um
Twitter LinkedIn channels where you can
provide us input we love to hear from
you um we want to hear what is working
for you the great case of of of success
that you've had doing fun and
interesting things uh and where things
are not working so well for you CU
obviously we always want to make these
these things
better all right so that's going to be
for this video and of course I'd like to
say a big thank you to Rob hot Mentor
here uh for actually taking out the time
from his extremely busy schedule to
record this video with me all right so
that's going to be it for this video of
course Rob High where can my subscribers
contact you um I'm on LinkedIn uh and uh
under uh H
ghr us. ibm.com or on Twitter at um R
high rhi G and uh I uh welcome you to
join me on either Channel perfect your
link your contact information will be
down in the description below so if you
like to contact Rob please feel free to
do so all right so that's going to be it
for this video if you'd like to leave
any comments suggestions questions or
feedback leave them down in the comment
section below or you can email email
them to me or drop high at tajim Manny
gmail.com or his contact information or
you can tweet them to me at tajim Manny
and of course if you really like the
video please make sure to leave a like
down below and if you believe it can
help anybody else you know develop with
Watson and his cognitive abilities
please do share the video as well that's
going to be it for this video but if you
really do like my content though and you
want to be notified whenever I release
new content and new videos please do
make sure to subscribe to the channel
and turn on notifications and that's
going to be it for this video today
thank you very much for watching today
goodbye thank you Tam