►
Description
Presented by: Elvis Saravia
This presentation illustrates how GitHub, Codespaces, GitHub Copilot, and the open source community accelerated and democratized machine learning (ML) education. Elvis will discuss how GitHub achieved this through the DAIR.AI initiative and other initiatives. This session will also showcase how to use GitHub technologies to make ML education more accessible and extendable.
As always, feel free to leave us a comment below and don't forget to subscribe: http://bit.ly/subgithub
Thanks!
Connect with us.
Facebook: http://fb.com/github
Twitter: http://twitter.com/github
LinkedIn: http://linkedin.com/company/github
About GitHub
GitHub is the best place to share code with friends, co-workers, classmates, and complete strangers. Millions of people use GitHub to build amazing things together. For more info, go to http://github.com
A
A
A
Also
a
contributor
of
paper
stick
code
and
dear.ti,
so
paper.
C
code
is
a
place
where
we
organize
machine
learning,
code
for
ML
papers
and
dear.ti
is
basically
an
initiative.
Community
Driven
initiative
to
democratize,
machine
learning,
education,
I'm.
Also
an
independent
researcher.
I
do
a
bit
of
research
for
topics
around
NLP
like
language
models.
A
You
can
also
find
me
in
places
like
Twitter
and
GitHub
as
well.
So
I
want
to
talk
a
bit
about
my
involvement
in
ml
education
as
I've
introduced
already
dear.
The
ti
is
an
initiative
that
I'm
very
heavily
involved
and
it's
a
Community
Driven
initiative,
meaning
a
lot
of
folks
from
around
the
world.
They
come
together,
put
their
minds
together
and
basically
collaborate
to
provide
tutorials
for
the
community,
and
these
tutorials
are
around
Ai
and
machine
learning.
So
we
have
a
lot
of
those
and
tutorials
and
repositories
where
we're
basically
democratizing
machine
learning
education.
A
We
do
open
source,
we
open
source,
all
our
tutorials,
we
open
source,
all
our
research
as
well,
and
we
have
even
also
contributed
to
you
know,
like
data
sets
for
machine
learning.
We
have
also
been
involved
in
machine
learning
research
as
I've
said,
and
one
place
where
I
spend
a
lot
of
time.
Thinking
about
how
to
organize
machine
learning
knowledge
is
fair.
A
Pc
code,
I'm,
a
contributor
but
I
also
work
closely
with
that
team,
and
basically
our
mission
is
to
organize
scientific
knowledge
and
converting
that
information
into
useful
knowledge,
and
they
do
a
lot
of
community
building
collaborate,
a
lot
with
researchers
to
organize
content
better
and
to
make
sure
that
we
are
pushing
for
open
code
and
open
source
code,
around
machine
learning
papers
and
machine
learning.
Research
in
general
also
do
a
bit
of
data
science
and
NLP
research
as
well.
So
the
topic
of
this
presentation
is
around
ml
education.
A
When
I
say
democratizing
about
education,
I
mean
a
couple
of
things
and
I'll
get
to
that
in
a
bit,
but
I
also
want
to
start
with.
What's
a
state
of
ml
education
today,
as
I
think
really
hard
about
this
problem,
and
how
do
we
use
the
tools
today?
How
do
we,
you
know,
work
with
a
community
to
help
accelerate
ml
education
and
improve
it
in
general,
we
need
to
understand,
what's
the
state
of
things
and
there's
a
lot
of
really
good
content
already
online.
Obviously,
there
is
free
content.
A
A
More
specifically,
are
you
know,
good
examples
of
enablers
to
help
accelerate
ml
education
and
we're
going
to
talk
about
that
here,
and
we
are
even
going
to
demo
it
about
how
AI
tools
like
code,
spaces
and
GitHub
are
accelerating
ml
education,
so
GitHub
and
copilot
offers
I
think
a
right.
The
right
set
of
tools,
the
right
approach
to
making
sure
that
whatever
education
we
are
democratizing
is
accessible
right.
It's
easy
shareable.
It
also
has
that
collaborative
nature
to
it
and
you
can
build
upon
very
easily
so
I
think
these
are
some
of
the
foundation
blocks.
A
A
So
one
of
those
areas
is
accessibility.
Accessibility
could
be
as
simple
as
providing
content
in
a
different
language
right
for
everyone.
Accessibility
might
mean
something
different
I.
Think
accessibility
is
an
interesting
one,
because
accessibility,
not
only
in
terms
of
language
but
the
tools
as
well.
Where
are
we
providing
this
educational
content?
How
we
provide
it
really
matters
right
if
we
are
going
to
provide
it
in
books
or
we're
going
to
provide
any
courses
that
may
not
be
as
accessible
to
some
students
and
we
need
to
make
it
as
accessible
as
we
could.
A
Extensibility
is
also
very
interesting
because,
when
you're
thinking
about
accelerating
ml
education,
what
you
want
to
do
is
you
want
to
help
a
community
to
find
content.
You
want
to
encourage
your
community
to
build
upon
the
content
that
exists
as
well,
so
you
want
to
have
a
set
of
best
practices
where
students
can
come.
You
know
maybe
help
the
Next
Generation
to
keep
building
upon
this
and
build
upon
that
and
I
think
in
the
science
domain.
That
works
really
well.
A
I
think
that's
something
that's
missing
and
we
need
to
work
on
and
tools
like
code
species
are
already
helping
out
with
this
and
also
I
think
it
needs
to
be
collaborative
whatever
solution.
We're
going
to
use
needs
to
be
collaborative
in
nature
and
what
I
mean
by
that
is
that
you
know
when
you
have
a
set
of
tutorials
or
set
of
content
right,
anything,
that's,
educational,
any
material,
that's
educational!
It
needs
to
have
an
element
of
collaboration
collaboration
because
it's
a
community
that's
going
to
help
build
this.
A
You
know
for
the
Next,
Generation
and
I
think
places
like
GitHub
and
their
other
wonderful
places
as
well,
where
information
is
shared,
offers
the
right
tooling,
to
make.
You
know,
collaboration,
much
easier,
I.
Think
this
probability
is
an
interesting
one.
It's
an
era.
We
think
about
a
lot.
You
know
working
for
PPC
code
and
that
team
very
closely
I
think
about
this
problem,
a
lot.
How
do
we
actually
make
content
discoverable
as
well,
because
we
could
have
really
good
content?
A
You
know
we
could
solve
accessibility,
we
could
solve
extensibility,
we
could
solve
collaboration
as
well,
but
if
it's
not
discoverable,
then
you
know
it
serves
no
purpose.
So
we
also
need
to
organize
the
content
better
present
it
better
and
make
it
more
discoverable
for
everyone
and
I
believe
GitHub
and
code
spaces,
and
a
bunch
of
other
tools
are
already
enabling
this
in
very
interesting
ways.
A
So
talk
a
bit
about
what
copilot
is
in
the
demo.
I'll
get
deeper
into
that,
but
at
a
high
level,
compilot
is
basically
an
AI
peer
programmer.
It
allows
you
to
easily
give
it
a
set
of
instructions
and
it
gives
you
code,
it
generate
code
and
that's
AI
powered
it's
you
know.
That's
the
the
technique
behind
us
called
open.
The
iconics
and
CODIS
is
enabling
this
code
generate
generation
tools.
A
So
it's
an
excellent
tool
for
learning
as
well
and
for
teaching
right.
So
you
can
imagine
someone
comes
and
has
co-pilot
installed.
They
can
learn
themselves,
but
also
use
it
as
a
tool
to
Showcase
what's
possible
with
it
and
you
know,
give
instructions
and
so
on
code.
Space
is
another
interesting
one
and
I
think
that's
what
I'm
excited
mostly
about
today
offers
the
right
environment
offers
the
right
set
of
tools
to
put
everything
together
in
one
place
right.
A
So
when
you're
working
the
code
lives
there,
the
tools
are
there,
libraries
are
there,
everything
is
in
one
place
and
it's
not
everywhere,
and
that
makes
things
much
more
accessible.
It
also
amplifies
the
other
key
things
that
I
mentioned,
such
as
discoverability
and
that
and
extensibility
as
well
so
code
space
is,
in
short,
is
just
like
a
fast
Cloud
developer
environment,
and
you
know
it
provides
an
environment
where
I
believe
really
Powers
a
male
education.
It
does
solve
the
problem
of
accessibility.
A
So,
in
short,
what
I've
said
is
that
code
space
is
proscopilot
and
many
other
AI
power
tools
are
now
accelerating
education,
and
it's
great
to
see
that
it's
great
news,
because
not
only
now
do
we
have
the
right
tools
to
help.
You
know:
accelerate
the
education,
but
also
you
know,
build
a
set
of
good
practices
around
that
as
we
keep
thinking
about
this
problem
of
ml
education
in
the
next
few
minutes,
I
will
go
through
a
demo.
A
I
will
show
you
how
we
can
use
code,
spaces
and
compilot
to
really
Empower
Learners,
and
you
know
have
a
set
of
tools
that
allow
students
and
Learners
to
hit
the
ground
running
rather
than
being
frustrated
about
how
to
install
Tools
in
this
demo.
I
want
to
demonstrate
how
we
can
use
GitHub
tools,
code,
space
and
compilot
to
help
with
machine
learning
education,
so
we
have
created
together
with
the
community,
a
set
of
ml
notebooks.
You
can
find
them
here
in
the
organization
here
at
the
organization.
A
A
So
one
problem
is:
how
do
we
make
this
content
accessible,
easy
to
access
and
to
build
upon,
and
also
we
are
interested
in
a
solution
that
makes
things
collaborative
as
well?
So
if
someone
wants
to
build
upon
some
of
these
notebooks
and
contribute
to
the
repository
to
keep
building
the
content
and
keep
democratizing
the
education,
what
do
we
need
for
that
and
I
think
code?
Space
plus
you
know,
co-pilot
and
a
bunch
of
other
tools
are
really
allowing
this
to
happen
in
a
very,
very
interesting
way.
A
So
what
I
want
to
demonstrate
here
is
how
do
we
go
about
using
these
tools
like
code
space
on
GitHub
to
accelerate
ml
education
and
democratize
it?
So
in
this
repo
we
have
a
couple
of
notebooks.
So
if
you
go
into
the
notebook
folder,
you
will
find
a
couple
of
the
notebooks
here
that
we
have
created,
and
what
you
will
notice
here
is
that
we
have
this
code
button,
which
is
where
you
would
be
able
to
create
a
code
space
instance.
A
A
So,
in
order
for
us
to
create
that
environment,
where
we
have
all
the
tools
installed
and
all
the
libraries
that
are
necessary
to
run
all
the
notebooks
in
this
repo,
we
could
you
know,
hit
in
that
code
button
that
green
button
here
and
create
the
code
space
here,
I'm
interested
in
creating
a
code
space
based
on
a
configuration
that
I'm
offering.
So
the
configuration
is
in
this
Dev
container
folder.
You
can
check
that
out
in
once
we
go
to
that
option.
A
We
can
now
hit,
create
and
create
the
code
space
instance
based
on
a
few
options.
We
are
using
main
as
the
branch,
but
you
can
use
any
branch
and
the
dev
container
configuration
we
are
using
the
one.
That's
provided
as
part
of
the
repo
region
doesn't
matter
here
and
machine
type
as
well,
but
those
are
options
you
have
and
then
you
can
create
the
post
office.
A
Once
you
hit
that
button,
then
you
will
be
taken
to
the
code
space
instance
there's
a
process,
obviously
that
you
know
is
in
charge
of
creating
the
environment
based
on
that
configuration
file
that
you
specified
and
it
creates.
You
know
this
instance,
for
you.
This
instance
contains
some
of
the
tools
that
you
will
need
to
make
those
notebooks
that
are
part
of
this
Reaper
runnable.
If
you
notice,
when
I
say
notebooks
it's
basically,
what
you
have
here
is
something
that
looks
like
Visual
Studio
code
for
those
of
you
that
have
used
that.
A
You
know
vs
code
on
the
browser,
so
what
we
want
to
do
here
is
I
want
to
show
you
how
you
can
easily
use
the
code
space
instance
to
make
this
content
runnable
for
anyone,
and
in
this
code
space
instance
in
this
environment
right
I
have
already
installed.
You
know
the
tool
the
tools
have
already
been
installed,
but
now
I
need
to
install
the
libraries
that
I'll
need
to
run
these
notebooks
and
they're,
mostly
python
libraries,
so
I've
used
conda.
A
You
know
as
a
tool
to
create
environments
where
I
can
install
the
necessary
Library
libraries
to
run
notebooks.
So
I've
done
that
ahead
of
time
it
takes
a
bit,
but
in
this
demo,
won't
have
too
much
time
to
to
do
that.
But
I've
done
it
ahead
of
time
and
I've
created
a
environment
called
NLP
and
I've
activated
that
here
so
now,
I
can,
as
an
example,
I
would
like
to
run
this
notebook.
A
You
can
see
here
that
I
can
run
through
the
code
very
easily.
Everything
is
already
installed
and
it's
allowing
me
to
you
know
interact
like
you
would
do
with
a
jupyter
notebook
locally,
so
that's
very
powerful
because
everything
lives
here.
The
content
of
the
notebook
is
not
that
important,
but
as
I'm
going
through
the
cells.
Something
change
in
this
right
in
this
repo
and
I
can
easily
check
Source
control
and
if
there's
any
changes
that
I
want
to
push
back
to
the
original
repository
I
can
easily
do
it
through
here
as
well.
A
I
won't
go
through
all
the
code
cells
here,
but
you
can
try
this
on
your
own,
but
it's
definitely
something
that
is
runnable
and
it
allows
you.
You
know
pretty
much
like
out
of
the
box
to
run
code
and
push
changes
to
repository,
allows
a
collaborative
environment
very
easily
to
spin
up
and
I.
Think
that's
a
beautiful
experience
for
anyone,
that's
learning
and
also
for
anyone,
that's
teaching
as
well.
A
A
So
that's
the
really
the
powerful
I
think
feature
of
something
like
code
spaces,
combined
with
other
tools
like
a
pilot
management
of
GitHub
repositories
and
so
on.
Thank
you
for
listening.
The
last
thing
I
want
to
talk
about
is
the
future
of
ml
education.
I've.
Given
a
couple
of
examples
of
the
tools
you
can
use
to
help
accelerate
ml
education,
but
we're
just
getting
started
with
all
of
this
I
we
need
a
community
to,
as
I
said,
help
build
best
practices,
keep
improving
the
tools.
A
You
know
give
feedback
in
creating
the
content
that
will
accelerate
education
in
ml
space
and
I'm
really
excited
about
the
future
of
ml
education.
As
we
see
the
community
get
together,
sharing
ideas
and
making
content
more
accessible,
I
think
with
the
right
set
of
tools
and
keep
improving
those
tools.
A
It
is
making
things
more
accessible,
it's
making
information,
more
accessible,
I,
think
education
will
improve,
it
will
improve
and
it's
a
cycle.
Another
thing
that
I
want
to
highlight
to
wrap
up
this
talk
is
how
you
can
contribute
right.
What
do
we
do?
What's
the
next
step,
you
can
think
of
Education
as
a
powerful
enabler
for
the
tools
that
you're
building
as
well
and
your
adoption
right
have
in
the
past,
I've
worked
as
a
professional
trainer
and
I've
seen
how
powerful
it
is
when
you
put
together
a
set
of
really
good
educational,
tutorials
and
material.
A
We
also
need
to
think
about
accessibility,
as
I
said
how
to
make
content
accessible.
When
you
build
a
tool,
you
build
a
library
and
you
open
source
it.
What
are
you
offering
as
part
of
that
launch
or
as
part
of
that
tool,
to
help
folks
learn
about
those
tools?
You
know
and
contribute
eventually
right.