ever wonder what is actually connecting
us
what can we do in order to engage into
something for example if you want to be
in relationship with somebody or if you
want to grab a good opportunity so what
we should do the number one action of a
suggests is we have some quality our
information we have to tell people
something that we want like if you miss
somebody
how can you make him do the simple
approach is just tell them we miss them
and if you are too sure you can actually
send over some text but here now I
wanted to point out the difference
between sing get out and also write it
out so imagine if one day there’s a
stranger approached you and politely say
hello you might think that he or she is
asking for help or ask you for a favor
but what if you hurt hey hello so now
you should know that you are the one who
might need some help so this is the idea
I wanted to tell you tax doesn’t really
help in conveying the emotion but we can
easily motion by listening to it however
not everyone is that fortunate and also
healthy as marvelous some of them might
be suffering from different diseases
which make them lost their ability to
speak it may due to accident the brain
injury the vocal cord damage and so on
being unable to speak must slow them
down in engaging themself into the
society and hence making them feel
isolated
another amazing usage of our voice is it
actually tells who we are without
exactly and physically looking at you I
can still recognize you by listening to
your points our voice is actually
hearing our identity our voice is our
trademark but the zip indirectly imply
that
losing our boys is equal to losing our
identity so here I want to show you one
of the video clip which I depicted from
a movie maybe spend years ago let’s have
a look
[Music]
we dis slowly read it pleasure
Busby’s company is what I most enjoy but
attack on Missy Anne sees chair when she
called him a horrible boy
in a lot of fun let’s bring into the
kitchens across the
common schemes to reveal
[Music]
copy the guards eyes way don’t you late
for that
[Music]
[Music]
everything is not alright for semester
still compiling
[Music]
[Music]
[Music]
Suzanne’s a movie
[Music]
okay so now can you actually get what
the agent is actually doing first it
forced this divorce the guy in order to
speak some words in order to capture the
voice characteristic and features and
then the agent can use the recording to
quickly train a talking machine in order
to talk using his voice
tenders back this technology is look
incredibly advanced and impossible but
now a day it has been realized
under the effort of Center of speech
technology research from the University
of Edinburgh which I firmly attached to
so the name of this technology is called
hidden Markov models each delicious
system this is a system which is able to
speak whatever language in order you
give the correct input when people talk
about speech synthesis system they will
think that this is a playback from our
recording however this is not a playback
play by having a limitation which is it
is only able to speak the recorded video
and will be coded database but with this
system it is actually using our recorded
voice characteristic in order to
construct an ill and novel words so
let’s take a look of how does it work
this is our voice speech waveform even
though it looks like a super messy and
complicated with combination but if we
take a look using a magnifier or
microscope we can actually see the
constant styling units we have a minimum
time frame so this constant starting
unit is called funny
so this is California so this folio is
actually the basic basics language unit
in every language so taking English as
an example our basic unit conceded to
this morning is actually our alphabet so
one good thing about the alphabet is
each alphabet only consists of a
constant starting unit only so let’s
take a look with this implementation we
can easily blend them into the hidden
Markov model that I have mentioned
earlier but what is the hidden Markov
model hidden Markov model is actually a
finite state machine which is able to
change to generate a situation discrete
time of the nation throughout the time
so when the each time unit has passed it
will change the state according to the
state transition probability so meaning
that the probability to stay we did in
the state or chunk into the next state
can be actually trained by us by giving
the current input so how does this edit
Markov model helps in speech synthesis
we can actually imagine each model is
actually a baby
but why baby because we don’t teach may
be difficult things we only teach it
baby still awaits so that is the reason
in each model we only train with only
one alphabet so let’s take the
perspective in an overall so if you want
to synthesize the words that we want to
say firstly we have to construct the
words using the alphabet and then these
alphabets will hence being match with
the corresponding hidden Markov models
and this hidden Markov models will
release the constants starting the study
coefficient and when we put them all
together in a sequence and make that
speech together this sequential study is
actually made up our speech so now we
understand how does this work but our
next problem is does it exactly sound
like what we wish for our student voice
should have a
Kuras the Esper so let’s take a pop quiz
I prepared puppies for all of you can
anybody exactly pronounce these tourists
correctly without any problem
so the first word is called prepare but
the second word is called preparation
what’s the difference the difference is
the first e is started differently when
they were put in a different world
so we as the human we can easily
distinguish the difference
but how can our target machine to speak
the difference so in order to maintain
accuracy we have incorporated artificial
intelligence in order to classify
different pronunciation our division
intelligence requires our classifiers in
order to tell the difference and also
make the decision so we have to find out
a way in order to see what is the
difference between the words in a
simplest thing we can actually make used
the words combination in order to make
the classifier so look at the words look
at the letter around a bit they can
easily become our classifiers so
therefore with this implementation our
artificial intelligence is able to
improve our student experience of the
intelligence so this speech to the SIS
system can actually do what we can
affect the researchers never stopped
after develop this program the strongly
believe that our voice is actually
carrying more than our stuff so here I
wanted to introduce to you a big
voluntary project which is called speak
unique this is a multidisciplinary done
by the University of Umbra and also the
Ian McDonald Centre for motor neuron
disease research what they do is
basically to create a personalized
digital voice for communication it
simply and effectively the motivation
behind this project is with our speech
become in
we tends to use the augmentative and
alternative communication in order to
get our message across this approach can
range from a simple strategies such as
our gestures or any voice output
communication aids in order to project
the synthetic speech however many users
complain that this kind of synthetic
speech doesn’t really reflect their
identity and they were forced to choose
only from a limited section of voices
which only maybe one or two is come from
the same country it may never be
occurred to you previously but our
speech is actually as unique as our work
as our eyes or even our fingerprint our
voice is the basic component in order to
build our image and our identity so what
they are actively doing is the voice
banking so voice banking is a project
which is they will collect a lot of
different voice samples around the club
because they want to use the voice in
order to make a collection of data
collection and in order to have this
project voice banking can actually have
two ways
so first if you are first seeing you are
going to lose your voice in future where
I hope it will not happen so you can
actually make a voice banking in order
to let them customize a voice back a
talking machine using your own voice
however the second approach it can be we
can actually merge different voices in
order to create to customize your
personalized voice in order to speak
something about your voice so therefore
here I mean this idea let me show you a
successful case about how this Speak
unique project successfully have a
prison in order to regain the box
I’m a student at Edinburgh University
where I studied wrong so now can you see
the hardship for those who are unable to
speak in order to articulate to begin
this is not easy so we as the healthy
people we have a healthy voice should be
always grateful of what we have we
should always watch our words because we
are not being even a voice to hurt
people but if they want to do more you
can actually sign up to become a voice
donor in order to contribute to the
speech unit project your voice could
actually have so now why are we trying
that hard in order to help you to regain
your voice we have a lot of different
approaches in order to get our message
around losing our voice does not mean
that is the end of the world right
however the reason behind is even don’t
know today we have a lot of different
approach to deliver your message but
sometimes we still wish to feel the love
even though our eyes are closed and
that’s all I want to share and thank you
for letting your ears thank you
[Applause]
us
what can we do in order to engage into
something for example if you want to be
in relationship with somebody or if you
want to grab a good opportunity so what
we should do the number one action of a
suggests is we have some quality our
information we have to tell people
something that we want like if you miss
somebody
how can you make him do the simple
approach is just tell them we miss them
and if you are too sure you can actually
send over some text but here now I
wanted to point out the difference
between sing get out and also write it
out so imagine if one day there’s a
stranger approached you and politely say
hello you might think that he or she is
asking for help or ask you for a favor
but what if you hurt hey hello so now
you should know that you are the one who
might need some help so this is the idea
I wanted to tell you tax doesn’t really
help in conveying the emotion but we can
easily motion by listening to it however
not everyone is that fortunate and also
healthy as marvelous some of them might
be suffering from different diseases
which make them lost their ability to
speak it may due to accident the brain
injury the vocal cord damage and so on
being unable to speak must slow them
down in engaging themself into the
society and hence making them feel
isolated
another amazing usage of our voice is it
actually tells who we are without
exactly and physically looking at you I
can still recognize you by listening to
your points our voice is actually
hearing our identity our voice is our
trademark but the zip indirectly imply
that
losing our boys is equal to losing our
identity so here I want to show you one
of the video clip which I depicted from
a movie maybe spend years ago let’s have
a look
[Music]
we dis slowly read it pleasure
Busby’s company is what I most enjoy but
attack on Missy Anne sees chair when she
called him a horrible boy
in a lot of fun let’s bring into the
kitchens across the
common schemes to reveal
[Music]
copy the guards eyes way don’t you late
for that
[Music]
[Music]
everything is not alright for semester
still compiling
[Music]
[Music]
[Music]
Suzanne’s a movie
[Music]
okay so now can you actually get what
the agent is actually doing first it
forced this divorce the guy in order to
speak some words in order to capture the
voice characteristic and features and
then the agent can use the recording to
quickly train a talking machine in order
to talk using his voice
tenders back this technology is look
incredibly advanced and impossible but
now a day it has been realized
under the effort of Center of speech
technology research from the University
of Edinburgh which I firmly attached to
so the name of this technology is called
hidden Markov models each delicious
system this is a system which is able to
speak whatever language in order you
give the correct input when people talk
about speech synthesis system they will
think that this is a playback from our
recording however this is not a playback
play by having a limitation which is it
is only able to speak the recorded video
and will be coded database but with this
system it is actually using our recorded
voice characteristic in order to
construct an ill and novel words so
let’s take a look of how does it work
this is our voice speech waveform even
though it looks like a super messy and
complicated with combination but if we
take a look using a magnifier or
microscope we can actually see the
constant styling units we have a minimum
time frame so this constant starting
unit is called funny
so this is California so this folio is
actually the basic basics language unit
in every language so taking English as
an example our basic unit conceded to
this morning is actually our alphabet so
one good thing about the alphabet is
each alphabet only consists of a
constant starting unit only so let’s
take a look with this implementation we
can easily blend them into the hidden
Markov model that I have mentioned
earlier but what is the hidden Markov
model hidden Markov model is actually a
finite state machine which is able to
change to generate a situation discrete
time of the nation throughout the time
so when the each time unit has passed it
will change the state according to the
state transition probability so meaning
that the probability to stay we did in
the state or chunk into the next state
can be actually trained by us by giving
the current input so how does this edit
Markov model helps in speech synthesis
we can actually imagine each model is
actually a baby
but why baby because we don’t teach may
be difficult things we only teach it
baby still awaits so that is the reason
in each model we only train with only
one alphabet so let’s take the
perspective in an overall so if you want
to synthesize the words that we want to
say firstly we have to construct the
words using the alphabet and then these
alphabets will hence being match with
the corresponding hidden Markov models
and this hidden Markov models will
release the constants starting the study
coefficient and when we put them all
together in a sequence and make that
speech together this sequential study is
actually made up our speech so now we
understand how does this work but our
next problem is does it exactly sound
like what we wish for our student voice
should have a
Kuras the Esper so let’s take a pop quiz
I prepared puppies for all of you can
anybody exactly pronounce these tourists
correctly without any problem
so the first word is called prepare but
the second word is called preparation
what’s the difference the difference is
the first e is started differently when
they were put in a different world
so we as the human we can easily
distinguish the difference
but how can our target machine to speak
the difference so in order to maintain
accuracy we have incorporated artificial
intelligence in order to classify
different pronunciation our division
intelligence requires our classifiers in
order to tell the difference and also
make the decision so we have to find out
a way in order to see what is the
difference between the words in a
simplest thing we can actually make used
the words combination in order to make
the classifier so look at the words look
at the letter around a bit they can
easily become our classifiers so
therefore with this implementation our
artificial intelligence is able to
improve our student experience of the
intelligence so this speech to the SIS
system can actually do what we can
affect the researchers never stopped
after develop this program the strongly
believe that our voice is actually
carrying more than our stuff so here I
wanted to introduce to you a big
voluntary project which is called speak
unique this is a multidisciplinary done
by the University of Umbra and also the
Ian McDonald Centre for motor neuron
disease research what they do is
basically to create a personalized
digital voice for communication it
simply and effectively the motivation
behind this project is with our speech
become in
we tends to use the augmentative and
alternative communication in order to
get our message across this approach can
range from a simple strategies such as
our gestures or any voice output
communication aids in order to project
the synthetic speech however many users
complain that this kind of synthetic
speech doesn’t really reflect their
identity and they were forced to choose
only from a limited section of voices
which only maybe one or two is come from
the same country it may never be
occurred to you previously but our
speech is actually as unique as our work
as our eyes or even our fingerprint our
voice is the basic component in order to
build our image and our identity so what
they are actively doing is the voice
banking so voice banking is a project
which is they will collect a lot of
different voice samples around the club
because they want to use the voice in
order to make a collection of data
collection and in order to have this
project voice banking can actually have
two ways
so first if you are first seeing you are
going to lose your voice in future where
I hope it will not happen so you can
actually make a voice banking in order
to let them customize a voice back a
talking machine using your own voice
however the second approach it can be we
can actually merge different voices in
order to create to customize your
personalized voice in order to speak
something about your voice so therefore
here I mean this idea let me show you a
successful case about how this Speak
unique project successfully have a
prison in order to regain the box
I’m a student at Edinburgh University
where I studied wrong so now can you see
the hardship for those who are unable to
speak in order to articulate to begin
this is not easy so we as the healthy
people we have a healthy voice should be
always grateful of what we have we
should always watch our words because we
are not being even a voice to hurt
people but if they want to do more you
can actually sign up to become a voice
donor in order to contribute to the
speech unit project your voice could
actually have so now why are we trying
that hard in order to help you to regain
your voice we have a lot of different
approaches in order to get our message
around losing our voice does not mean
that is the end of the world right
however the reason behind is even don’t
know today we have a lot of different
approach to deliver your message but
sometimes we still wish to feel the love
even though our eyes are closed and
that’s all I want to share and thank you
for letting your ears thank you
[Applause]
Please follow and like us: