►
Description
A presentation at the Hyperledger Telecom Special Interest Group on December 2, 2021. The discussion was about a project that builds an MVP for automating the drafting and negotiation processes for Telecom roaming agreements using Hyperledger Fabric. In the first phase the project addressed how to use NLP to extract main features from a number of telecom agreements drafts. Specifically, creating a model that represents the text of the roaming agreement into standard clause text, variables, text variation and custom text. The second phase addressed building the chaincode and the backend and frontend pieces of the MVP to show a simple demo of how this can work.
More about the project is at: https://wiki.hyperledger.org/display/INTERN/The+Use+of+NLP+and+DLT+to+Enable+the+Digitalization+of+Telecom+Roaming+Agreements
And if you're interested in how blockchain is being used in telecom, feel free to join the Hyperledger Telecom Special Interest Group at: https://wiki.hyperledger.org/display/TCSIG/
----
Hyperledger – Open Source Blockchain Technologies
Hyperledger is an open source community focused on developing a suite of stable frameworks, tools and libraries for enterprise-grade blockchain deployments.
It serves as a neutral home for various distributed ledger frameworks including Hyperledger Besu, Fabric, Sawtooth, Indy, as well as tools like Hyperledger Avalon, Cactus and libraries like Hyperledger Aries, Ursa. Learn more about Hyperledger projects: https://www.hyperledger.org/use
Case Studies: https://www.hyperledger.org/learn/case-studies
Training & Certification: https://www.hyperledger.org/learn/training
Tutorials: https://www.hyperledger.org/use/tutorials
Webinars: https://www.hyperledger.org/learn/webinars
Events: https://www.hyperledger.org/events
Vendor Directory: https://www.hyperledger.org/use/vendor-directory
Subscribe to the Hyperledger Newsletter: https://www.hyperledger.org/newsletter
Follow-us on Twitter @Hyperledger
Learn about Hyperledger Membership: https://www.hyperledger.org/about/join
#Hyperledger #Blockchain
C
B
I'll
just
give
a
quick,
quick
overview
of
the
project
and
the
idea
concept
behind
it
and
then
I'll.
Let
everyone
introduce
let
everyone
here
introduce
themselves.
So
the
idea
behind
the
project
is
that
telecom
roaming
agreements
are
digital,
sorry,
legal
agreements.
Usually
there
are
bilateral
agreements
and
usually
it's
it's
lots
of
legal
terms,
but
there's
also
a
technical
side
of
it
as
well.
A
commercial
and
the
concept
here
is.
B
We
are
trying
to
tackle
this
at
the
early
stage
of
building
the
terri
common
roaming
agreements,
which
is
at
the
stage
of
drafting
and
negotiation,
and
we
came
across
the
concept
from
other
from
other
domains
actually
in
financial
markets.
There's
these
over
the
counter
derivatives
agreements,
which
is
look
to
us
something
similar
and
there's
already
something
similar
in
the
financial
market
that
that
handles
something
in
in
the
domain
of
how
to
draft
and
negotiate
kind
of
master
agreement.
B
So
the
concept
is
to
use
both
natural
language,
processing
and
dlt.
In
this
case,
it's
cyber
ledger
fabric.
The
natural
language
processing
is
to
analyze
the
text
of
different
legal
agreement
documents
and
try
to
classify
the
text
into
different
different
parts
and
santiago
will
will
go
through
this
and
then
the
the
dlt
is
to
manage
the
automation
and
ensure
the
trust
and
the
auditing
and
the
traceability.
B
Myself
is
my
myself
ahmed
swayer.
I
work
in
blockchain
industry
for
the
last
three
to
four
years.
I
built
blockchain
solutions
and
I
asked
both
nuridian
and
muhammad
to
join
me
because
they
they
can
bring
the
other.
The
other
side
of
the
project,
which
is
the
natural
language
processing
which
is
their
their
speciality
and
sentiaco,
is
the
mentee
in
this
in
this
mentorship
project,
and
he
was
the
one
performing
all
this.
B
D
Hello,
everyone
and
welcome
to
this
meeting.
My
name
is
nuridin
sadawi,
I'm
from
libya
based
in
london
uk
I'm
a
senior
lecturer
at
the
university
of
london
and
the
instructor
with
o'reilly
and
pearson.
I
specialize
in
machine
deep
learning
and
in
those
areas.
So
that's
why
I
was
invited
to
co-supervise
this
project.
I'd
like
to
thank
ahmed
and
hyperledger
and
linux
foundation
for
facilitating
this.
It's
been
a
really
good
project.
Santiago
has
done
a
fantastic
job,
as
he
will
present
to
us
today.
I'm
grateful
also
to
muhammad
for
his
contribution.
D
E
Hi
everyone-
this
is
mohamed
eduard,
I'm
a
software
engineer
at
qatar
center
for
artificial
intelligence
like
I'm,
like
mainly
in
work
in
applied
machine
learning
and
deep
learning
under
like
spectrum
of
applications,
and
I'm
involved
in
many
projects
with
the
local
and
international
entities
like
we
have
like
collaboration
projects
with
boeing
in
predictive
maintenance
in
detecting
small
objects,
and
also
we
have
a
cooperation
project
with
mit
about
urban
computing
and
with
qatar
airways
in
local
entities.
So
like
yeah.
So
that's
why
ahmed?
E
F
I
recently
finished
my
phd.
I
am
currently
working
in
data
analysis
and
information
management
group
of
faith
that
is
a
research
center
in
basket
country
that
is
part
of
the
university
of
nabarla.
First
of
all,
I
would
like
to
thank
to
the
special
telecom
group
for
the
invitation
to
the
presented
work.
They
use
of
nlp
dlt
to
enable
the
digitalization
of
telecom
running
agreement
as
part
of
the
hyper
legend
mentorship
2021
project.
F
Running
agreement
operation
ensures
business
continuity
and
ubiquitous
service
access
to
the
end.
Customer
across
various
technologies
stack
including,
for
instance,
new
technologies
like
5g,
iot
technologies
and
iot,
and
the
robbing
agreement
address
the
technical
and
commercial
components
necessary
to
enable
the
service
to
a
running
customer
so
that
it
constitutes
an
essential
part
of
the
business
managing
issues
such
as
you
know.
F
F
F
In
this
way,
blockchain
integrated
with
other
technologies
such
as
natural
language
processing,
can
cover
these
capabilities.
You
know
an
overview
of
the
project
can
be
found
in
the
first
medium
article
that
that
we
published
as
part
of
our
project
and
the
url
is
included
in
the
in
the
footer
of
this
slide.
F
In
general
terms,
we
can
think
in
this
project
that
looks
a
very
early
phases
of
the
running
agreement
and
specifically
the
drafting
and
negotiation
process.
Our
solution
can
facilitate
the
process
of
drafting
an
agreement
and
turning
it
into
digital
format
can
help
in
establish
a
root
of
truth
of
the
agreement.
F
In
terms
of
article
proposal,
articles
change
until
the
entire
agreement
is
finalized,
a
background
level
we
want
to
analyze,
on
the
one
hand,
the
nlp
technology
and
the
amazon
comprehension
tool,
but
a
quick
concept
in
this
term
natural
language
processing
is,
you
know,
it's
a
soup
view
it's
a
social
site
of
linguist
and
computer
science
and
artificial
intelligence
concern
with
the
interaction
between
computer
and
human
language
at
the
at
the
same.
At
the
same
time,
you
know,
amazon,
comprehend
is
a
natural
language
processing
service.
F
It's
like
a
tool
that
use
machine
learning
to
uncover
variables
inside
and
connection
in
text
to
understand
what
are
variables,
variation
and
standard,
standard
cloud,
sorry
and
custom
text.
A
snapshot
of
two
documents
are
used
on
the
left
side.
You
can
see
the
a812
gcma
template
and
on
the
right
side,
you
can
see
one
sample
of
the
mobile
network
operator
proximus
running
algorithm.
In
this
way,
a
standard
class
implies
to
leave
an
article
or
sub
article,
as
found
in
the
temple.
F
Variables
represent
the
value
of
certain
attributes
that
are
found
in
a
certain
text
such
as,
for
instance,
dates
and
name
of
entities,
amounts
and
others.
In
addition,
variation
represents,
you
know
changing
variables.
For
instance,
you
know
mobile
networks,
operator
dates,
penalty,
currency
and
so
on,
and
so
on.
Concerning
the
original
text,
I
mean
the
dcma
template
in
the
article
or
article
finally,
customized
text
introduce
completely
completely
new.
You
know
article
of
soft
article
that
respond
to
particular
interests.
F
The
application
life
cycle
is
composed
by
four
stages.
The
first
one
is
enabled
the
running
agreement
principle.
The
second
one
is
the
registration
of
both
the
mobile
network
operator
and
the
robbing
agreement.
In
addition,
the
third
stage
is
the
running
women
drafting
process
and,
finally,
is
the
the
the
face
off
to
obstruct
the
the
agreement.
F
The
first
page,
the
third
phase,
starts
when
two
mobile
networks
operator
enable
the
roaming
agreement
principle
throw
out.
You
know
a
document
that
could
be
shared
with
the
ecma
as
a
central
authority.
At
this
point,
dcma
is
able
to
create
the
roaming
agreement
output
template
using
the
nlp
ending
execution.
F
The
two
main
components
of
the
project
are,
first
of
all,
the
nlp
engine
and
the
second
one.
The
shame
code,
which
are
discussed
below
the
overall
architecture
of
the
nlb
engine,
is
built
over
a
docker
infrastructure
establishing
as
input
the
rubbing
agreement
draft,
as
well
as
the
decimate
template
the
processing
layer
is,
you
know,
a
logic
identified
as
the
nlp
ending
which
include
as
output,
and
you
know
the
classification
of
the
article
in
a
standard
class
in
variation
in
customized
texts
and
variables.
F
F
Details
of
the
nlb
envy
can
be
found
in
the
second
media
article
published
as
part
of
this
project
and
again
the
url
is
included
in
the
in
the
footer
of
of
the
of
the
presentation
of
this
slide.
Sorry,
this
this
image
of
this
slide
illustrate
the
part
of
the
set
of
meta
that
can
make
up.
You
know
the
shame
called
life
cycle.
The
invocation
of
this
method
established
the
transition
between
the
different
state
of
the
shin
code.
This
transition
takes
place
at
three
application
level.
The
first
one
is
the
running
agreement
level.
F
F
F
It
should
be
noted
how
this
email
is
the
administrator
and
the
maintainer
of
a
set
of
service,
including
you
know
the
front
end,
the
led,
the
kibana
graffana
from
a
fuse
and
swagger.
In
addition,
the
two
mobiles
networks
operators
include
the
functionalities
to
maintain
the
and
monitoring
the
the
language.
F
One
of
the
most
remarkable
points
of
this
project
is,
you
know,
is
the
integration
with
other
open
source
projects
that
belong
to
the
hyper
layer,
mentoring
or
mentorship
program,
specifically
the
project,
analyzing,
hyper
layer,
fabulous
transactions
and
logs
using
elasticsearch
and
kibana.
We
include
this
project
as
part
of
of
you
know,
of
our
project
and
in
the
demo,
it's
possible
to
see
how
the
this
this
integration
sorry
now
it's
time
to
to
have
a
a
demo
time.
F
First,
one
you
can
see
the
transaction
committed
into
the
ledger
once
the
same
code
is
installed
and
instantiated
the
attempt
to
access
for
the
user
fail,
because
the
smas
need
to
register
it
before
now,
usma
register,
admins
and
user
of
the
mobile
network's
operator,
the
user
one
now
is
able
to
register.
For
instance,
one
mobile
network
operator.
F
Now
can
be
verified
how
the
transaction
is
committed
into
the
lecture
now
the
drumming
agreement.
Let
me
let
me
stop
here
the
the
the
video
as
you
can
see
here.
You
know
the
the
mobile
network
operates
on
number
two
proposed:
a
roaming
agreement-
registration,
as
you
can
see
here
these
this
this
thing
represents
the
the
roaming
agreement
output,
template
that
was
provided
by
the
nlp
engine.
F
F
Again,
the
transaction
is
committed
into
the
letter,
and
now
this
mobile
network
operator
is
able
to
propose
an
article
based
again
on
the
running
agreement.
Output,
template
loaded
from
the
nlp
ending
here,
as
you
can
see,
a
standard
class
is
defined,
variables
are
introduced
and,
in
addition,
variations
are
defined.
F
F
F
F
In
addition,
the
architectural
design
document
of
the
system-
also,
we
create
a
simple
user
interface
as
proof
of
concept
that
will
show
the
step
of
the
agreement
drafting
and
finally,
we
publish
an
academic
conference
paper
with
the
main
finding
of
the
project
for
the
future
awards.
We
have
included.
F
F
B
B
So
we
come
at
the
very
early
stage
where
we
try
to
digitize
the
agreement
text
and
extract
it
into
those
features
through
the
nlp
model
and
then
use
the
chain
code
to
build
the
actual
digital
version
of
the
roaming
agreement
and
make
it
actionable.
So
the
next
phases
in
terms
of
settlement
disputes
resolution
can
benefit
from
having
this
digitized
version
of
the
roaming
agreement
text,
which
I
think
that's
the
point
where
santiago
mentioned
that
right
now,
it's
more
of
a
generalist
approach.
B
B
So
just
wanted
to
put
that
into
the
context,
and
I
think
also
it's
good
to
mention
that
next
week
we
are
meeting
with
the
gsma
to
try
to
find
out
some
synergy
between
what
we've
worked
on
and
the
bwr
group
at
the
gsma
and
see
if
how
we
can
maybe
contribute
with
our
work
that
santiago
built
into
the
into
the
gsma
the
bwr
initiative.
So
yeah
back
to
you
in
case.
C
C
D
B
B
We
are
also
that's
something.
Maybe
I
would
like
to
mention
to
you
webinar
and
david.
We
are
building
a
more
detailed
solution,
design
document.
We
can
share
that
document
with
you
and
maybe
we
can
maybe
have
it
sort
of
as
a
solution
brief
or
I
don't.
I
don't
know
what
exactly
the
the
best
name
to
call
it.
B
A
C
F
B
C
B
Yeah
yeah,
I
think
yeah
that
could
be,
but
in
terms,
I
think
one
of
the
issues
that
we
faced.
We
we
we
didn't,
have
anyone
who's
a
legal
person
in
our
team
and
just
as
I
said,
we
borrowed
this
concept
from
financial
markets
and
I
know
how
they
built
it
in
financial
markets.
It
was
legal
team
that
they
were
analyzing
text,
not
using
nlp,
but
people
analyzing
text
and
identifying
variables,
variations,
etc.
B
B
E
G
Complimentary
question
to
the
first
part:
basically,
if
we
can
build
a
generalized
negotiation
agreement
within
the
negotiation
system
on
hyperledger
fabric,
that
can
be
where
you
can
drop
in,
like
the
template
of
the
contract
and
then
the
negotiation
process
in
terms
of
as
presented
in
this
scenario,
each
party
kind
of
suggests
changes
and
then
the
process
moves
on
and
everybody
agrees,
and
then
you
have
at
the
final
stage
you
have.
You
have
a
contract
that
everybody
agrees
on
to
me.
B
B
We
targeted
javascript
like
an
nbm
package
that
will
enable
transforming
that
nlb
model
into
a
library
that
will
allow
easily
to
pick
article
2
and
then,
once
you
pick
article
2
that
will
populate
what
are
the
variables.
What
are
the
different
variations,
and
this
is
actually
what's
done
by
santiago,
but
in
not
as
as
a
library,
but
he
coded
that
into
the
front-end.
B
So
our
work
maybe
is
to
extract
that
into
a
library.
So
if
we
would
like
to
build
that
generalized
model,
what
will
happen?
You
will
take
different
sla
agreements?
You
will
take
different
roaming
agreements
with
and
then
you
will
build
an
nlb
model
for
each
one
of
these
and
then
build
some
packages
that
once
you
select
what
type
of
agreement
it
will
select
it
will.
B
G
G
B
Yeah,
in
that
case
yeah,
I
totally
agree.
So
that
would
be
like
the
chain
code
will
be
the
most
generalized
system
here
that
will
handle
just
the
process
of
adding
articles
specifying
the
ins
and
outs
of
each
article
and
then
handle
the
negotiation
piece,
which
is,
if
you
accept
it,
if
you
propose
change,
if
you
want
to
delete
it,
reject
it
wherever,
but
again,
it
would
have
to
start
with
building
the
different
models
for
the
different
use
cases
that
you
would
like
this
system
to
handle.
B
So
without
without
that
model,
you
don't
know
how
the
text
is
interpreted
because
differently.
There
are
specific
things
that
you
can
capture,
maybe
in
roaming
agreements,
but
one
if
you
go
to
a
real
estate
brokering
agreement.
That
would
be
maybe
different
kind
of
variables
and
and
text
that
you
need
to
analyze
to
come
up
with
the
the
exact
model
and
the
other
thing
is
which
we
mentioned
already,
that
we
borrowed
the
concept
from
another
industry.
So
definitely
a
generalized
model
can
be
built.
B
It's
just
you
need
to
to
do
that
batch
processing
at
the
beginning
by
building
your
nlp
nab
model.
So
if
you
build
that
generalized
chain
code,
then
you
would
start
every
now
and
then
to
go
into
a
specific
domain.
Analyze
lots
of
legal
texts
and
come
up
with
the
model
that
that
can
be
used
and
fair
to
say
is
that
telecom
running
agreements
are
not
that
complicated.
H
Yeah,
thank
you
and
thanks
ahmad
and
santiago
for
the
presentations
very
helpful.
So
thank
you
for
that.
So
I
know
this
is
a
proof
of
concept
project.
So
it's
not
a
production
environment,
but
it's
a
good
starting
point
to
carry
over
to
production.
So
my
question
is:
when
we
move
over
to
production
governance
question,
so
there
is
a
roaming
agreement
that
that
comes
into
existence
right
in
the
example
in
the
presentations
between
orange
and
telefonica.
H
B
So
one
one
of
the
things
that
we
are
looking
at
is
to
look
into
private
data
collections,
and
in
that
case,
that
will
be
something
where
the
roaming
agreement
is
is
is
just
built
inside
the
chain
code
actually-
and
you
kind
of,
let
me
try
to
find
the
best
word
describe
this.
You
can
yeah,
you
generate
the
the
actual
text
based
on
what's
inside
the
chain
code
and
since
it's
a
private
data
collection,
so
that
means
it's
only
the
two
operators
that
they
have
access
to
this
to
this
agreement.
B
Yeah
yeah-
currently
it's
not
there
currently
like
gsma,
is
the
the
authority
that
handles
or
manages
the
entire
network
and
everything
is
plain
accessible
to
everyone.
But
our
concept
is
to
advance
this
and
build
it
in
on
top
of
private
data
collections,
where
data
will
become
only
accessible
to
the
two
operators
that
they
are
right
so.
H
B
B
H
C
B
A
D
B
D
Lot
yeah:
my
point
is
a
little
bit
different.
What
I'm
saying
is
because
vpn
mentioned
that
it's
good
for
outreach.
We've
done
actually
some
of
that
to
prove
that
the
previous
work
can
be
quite
useful,
so
santiago
has
used
already
existing
stuff
and
hopefully
for
this
project
it
can
take
different
branches.
D
So
a
couple
of
suggestions,
for
example,
where
were
really
good
already
sorry,
I
forgot
names
anas
and
I
think
I
forgot
sorry
for
the
other
person's
name.
So
these
are
quite
useful
suggestions.
Some
of
the
stuff
here,
I
must
say,
is
basically
future
work.
Okay,
so
please
do
bear
in
mind
that
this
is
just
a
proof
of
concept
projects.
Not
it's
not
really
a
large-scale
project,
so
it
can
go
into
different
directions,
one
of
them.
For
example,
sorry
it
was
jason,
yeah,
jason,
john
one
of
them.
D
One
of
them
could
be,
for
example,
this
project
can
be
the
basis
or
the
foundation
for
another
project,
so
somebody
can
maybe
extend
it
to
something
else
like,
for
example,
into
interoperability,
for
example
between
different
providers
of
of
blockchain
technologies
such
as
ethereum
and
so
on.
Also,
of
course,
packaging.
This
stuff,
from
putting
it
into
libraries
and
reusable
modules,
that's
also
really
good
yeah.
So
this
is
actually
just
a
one.
One
point
I
wanted
to
to
highlight
also
regarding
the
material
and
the
existing
stuff
that
already
is
there.
C
D
C
C
So
if
we
can
work
on
that
part
like
as
a
separate
project
or
something
that's
what
they
do
in
different
foundations,
for
example
cntt.
This
is
with
linux,
foundation,
networking
and
gsm.
For
for
what
telco
cloud
should
be
like
in
the
implementation
part.
So
so,
if
we
can
do
something
like
this,
that
what
we
can
propose
to
gsma
regarding,
we
are
interested
in
building
some
reference
architecture,
implementations
and
models,
so
that
it
will
be
like
a
reference
point
for
other
operators
as
well.
C
B
B
Yeah
well,
let's
see,
I
think
they
they
already
work
with
two
or
three
operators
and
two
vendors.
I
think
I
think
you
you
already
hosted
the
shemet
and
someone
from
xenoverse.
They
delivered
a
presentation
to
the
group
yeah
in
the
past
few
months.
Yeah
we
will
discuss
that
with
them
and
see
if
they
yeah,
if
they
would
like,
maybe
to
set
up
a
small
group
that
can
like
bring
synergies.
A
F
F
A
In
santiago
feel
free
to
share
that
on
the
groups
list,
you
know
not.
Everybody
in
the
group
was
able
to
join
the
call
today
so
there's
over
200
people
on
the
list.
So
if
you
want
to
have
a
conversation
there
you're
certainly
welcome
to,
and
anybody
who
is
not
on
the
group's
list
but
would
like
to
take
part
in
discussions.
It's
an
open
list
and
you're
certainly
welcome
to
subscribe.