►
From YouTube: DevoWorm (2021, Meeting 2): Assorted Organizations Topics, Visualized Theory, and Physics of Life
Description
DevoWorm (2021, Meeting 2): Overview of New and Upcoming Publications, Finalized GSoC Projects, Growth, Form, Morphogenesis, and Machine Learning, Physics of Embryogenesis and the Molecular Milieu. Attendees: Susan Crawford-Young, Ujjwal Singh, Krishna Katyal, Shruti Rajvanshsingh, and Bradly Alicea.
A
B
B
A
B
A
A
A
Hello,
shreddy
and
so
we're
contributing
to
this
chapter
or
this
this
special
issue.
So
this
is
a
paper
that
again,
like
you
know,
this
is
something
that's
been
a
couple
years
in
the
making.
So
if
you
want
to
know
how
sometimes
how
papers
sort
of
caring
around
you
know
different
places
that
this
is
a
good
example
of
it,
this
paper,
you
know
it's
it's
one
of
the
series
of
papers
that
we
did
from
the
secondary
data
from
c
elegans
that
I
think
everyone
in
the
group
is
familiar
with.
A
C
A
A
Look
at
other
properties
of
embryogenesis
present,
some
sort
of
maybe
some
sort
of
data
extraction
data
analysis,
but
then
we
provide
a
framework
for
theoretical
exploration.
So
so
the
paper
you
know
it
starts
in
by
doing
a
there
are
a
lot
of
graphs
in
this
paper,
a
lot
of
supplemental
materials
at
the
end,
and
then
we
get
into
some
analysis
of
differentiation,
trees
and
lineage
trees.
A
Here's
a
graph
of
like
different
terminally
differentiated
cells,
so
these
are
cells
that
are
differentiated
into
things
that
are
going
to
become
tissues,
so
you
have
different
classes
of
cell
that
emerge
early
versus
late.
So
we've
characterized
these
here
we
have
some
heat
maps
of
some
all
the
cells
and
c
elegans
that
dif
that
that
divides.
A
These
terminally
differentiated
cells,
broken
them
out
by
functional
class
and
then
looked
at
the
time
that
they
terminally
divide
or
differentiate,
and
so
you
can
see,
though,
if
you
compare
different
classes
of
cell,
that
there
are
differences
in
when
they
they
sort
of
emerge
in
the
embryo.
So,
for
example,
the
hypodermal
cells
emerge
rather
early.
They
emerge
in
the
205
minute
to
240
minute
window
of
a
lot
of
them.
A
A
I
think
there's
neuron
versus
interneuron
and
there's
neuron
versus
muscle
and
centitium.
So
all
these
are
these
are
kind
of
useful
for
people
who
want
to
go
in
and
think
about
basically
characterizing
c
elegans,
quantitatively
and
then
moving
on
to
something
like
simulation
or
some
theory
that
they
want
to
test,
and
so
then,
at
the
end
of
the
paper.
A
Oh,
this
is
a
nice
graph
here,
where
we're
doing
these
things
called
differentiation
maps
where
you
have
you
start
off
with
this
is
like
a
two-dimensional
view.
This
is
like
a
left-right
view
versus
an
anterior
posterior
view
of
the
of
the
embryo,
so
you
have
two
different
anatomical
dimensions
here
and
then
you're
looking
at
how
they
move
across
those
dimensions
over
time.
So
your
p0
is
your
initial
condition.
It's
the
single
cell
version.
A
Then
you
get
you
divide
into
the
two
cell
stage.
Here
is
just
showing
the
anterior
end
of
that.
So
this
is
a
b.
Then
this
is
abp
and
aba,
and
then
you
keep
getting
these
divisions
and
they
go
in
different
directions
and
you
can
calculate
an
angle
of
this
and
so
it's
all
averaged
out,
but
you
get
an
idea
of
how
these
cells
move
around
in
space
as
they're
dividing,
and
so
this
gives
you
sort
of
a
map
to
that
and
generated
maps
up
to
like
128
cells.
A
So
you
can
see
when
you
get
to
higher
cell
numbers
that
they're
almost
like
trajectories
across.
So
you
know,
there's
a
lot
of
crowding,
but
this
is
representative
of
this.
Is
the
cell
centroid,
these
black
dots
and
then
or
I
guess,
they're,
green
and
red,
but
they
move
around.
You
know
in
different
ways,
and
so
you
were
going
to
say
something
susan.
A
A
A
So,
like
you
know
the
third
dimension,
at
least
from
the
data
we
have
doesn't
show
I
mean
they
don't
move
like
you
know
they
kind
of
move
up
to
maybe
another
level.
You
know,
but
they
don't
you
know
it's
not.
It
doesn't
explain
a
lot
of
the
variants
but
anyways.
That's
that's.
You
know
something
we
do
with
c
elegans.
It's
it's
somewhat
informative.
C
A
A
A
And
so
I'm
not
sure
we've
been
talking
about
this
for
a
couple
years
where
you
have
these
angles,
and
that
may
be
interesting
because
you
know
what
is
the
angle
of
division?
You
know
you
know
if
you
might
think
it's
kind
of
trivial,
but
actually
you
know
maybe
it
matters
where
the
centroid
of
the
cell
goes
in.
A
What
direction
is
it
going
at
an
angle,
or
is
it
going
straight
out
or,
and
so
you
could
actually
predict
some
of
the
physics
too,
of
some
of
the
cells
like
you
know
what
is
the
physical
process
that
leads
to
something
like
this,
where
it's
like
a
180
degree
angle
around
where
it's
going
in
left
versus
you
know
almost
like
a
left
right,
split
versus
something
like
this,
where
it's
very
narrow.
Let
me
find
an
example
of
a
narrow
one.
A
A
So
we
kind
of
talk
about
visualization
as
well,
so
we're
kind
of
creating
visualizations
here
and
you
know
what
can
we
say
something
about
that
and
then
future
directions
and
analysis.
So
this
is
kind
of
touches
on
some
of
the
machine
learning
stuff
for
segmentation
and
a
couple
other
papers.
I
could
have
cited
in
this
because
I
found
after
I
you
know
how
you
always
try
to
cite
as
much
as
you
can,
but
you
always
find
that.
A
C
A
Then
finally,
each
of
these
figures
are
considered
in
terms
of
theory
construction.
So
this
is
a
nice
area
that
I
don't
think
a
lot
of
biological
papers
usually
talk
about,
but
there's
a
whole
area
of
study
called
theory
construction,
and
so
this
is
actually
sort
of
people
talking
about
how
to
construct
theories.
A
It's
not
you
know
just
going
out
and
saying
I'm
going
to
make
a
theory
of
this
phenomenon.
I
see
in
my
data.
This
is
a
whole
formal
process
of
theory,
building
your
theory
construction,
and
so
we
have
each
we
go
through
each
graph
and
we
talk
about
like
these
different
relationships
and
then
you
know
maybe
how
you
can
apply
the
standards
of
theory,
construction
to
these
different
relationships,
and
so
this
I'm
not
going
to
go
deeply
into
that.
A
But
this
kind
of
you
know
talks
about
how
that
process
works
and
then
kind
of
maybe
suggesting
a
future
direction,
which
is
category
theory,
which
is
a
tool
for
mathematics.
That
is
rather
recent.
But
it's
it's
something
like
set
theory,
but
it's
much
more.
It's
much
different
than
set
theory
actually
mathematically.
A
It
allows
you
to
build
these
categories
of
different
functions
and
different
groups
and
then
apply
them
to
things
like
dynamical
systems
or
other
types
of
approaches,
and
so
I
had
a
longer
there
was
a
longer
more
to
be
said
on
this,
but
they
didn't
like
it,
so
it
got
cut
down
but
there,
but
that's
an
area
that
I
know
I
think
jesse
and
I
have
talked
about
category
theory.
So
that's
something
that
might
be
a
future
direction
if
people
are
interested
is
the
category
theory
of
development
where
you're
applying
this
this
technique.
A
Now
I
do
warn
people
that
it's
very
it's
somewhat
inscrutable,
because
it's
a
mathematical
field
of
mathematics,
so
you
know
that's
something
that
you
don't
necessarily
you
know
know
maybe
in
advance
and
you
get
involved,
and
you
don't
know
why
you
got
signed
up
for
it.
But
you
know
I
don't
know
if
anyone's
talking
about
this
right
now,
so
that
would
be
a
sort
of
a
frontier
area.
A
So
so
this
is,
these
are
the
there
well,
there
are
some
sources
of
data,
because
this
is
a
informatics
journal.
They
wanted
us
to
have
our
data,
our
secondary
data,
that
we
have
on
the
in
the
github
repos
and
then
the
references
and
then
there's
some
supplemental
materials
at
the
end.
So
again,
if
you
join
late,
I'll
put
the
link
to
the
paper
in
the
chat.
A
A
Oh
yeah:
well,
we
could
yeah,
we
could
apply
some
of
the
principles,
the
vassal
area
bacillary.
Of
course,
like
you
know,
every
organism
has
its
own
sort
of
mode
of
development
and
growth,
and-
and
so
you
know,
you
could
apply
it,
but
it's
you'd
have
to
modify
the
mouth
method
somewhat,
but
yeah
we
might
do
that.
We
have
some
quantitative
data
for
basal
area,
so
we
could
look
into
that.
So
yeah
we
have
some
new
data
from.
A
From
our
collaborator
in
germany,
thomas
and
he's
you
know,
he's
got
some
more
movies
that
we're
going
to
analyze
later
on,
and
so
we'll
look
at
that,
and
there
is
it's
a
lot
more
there's
a
lot
more
to
the
data
than
even
we
had.
You
know
back
when
we
did
the
machine
learning
paper
last
year,
which
he
generated.
Some
data
for
he's
got
some
more
even
more
interesting
data
to
look
at.
So
that's
going
to
be
another
interesting
area,
we'll
have
to
get
more
into
that.
A
Maybe
later
this
month
into
february,
we'll
get
into
that
those
data,
but
they
yeah.
They
look
pretty
good
and
we
can
apply
different
quantitative
methods
to
them
so
like
in
that
paper
we
did
some
machine
learning
visual
on
asthma,
we're
doing
some
work
on
that
and
then
thomas
actually
did
some
biomechanical
analysis
of
the
cells
looking
at
how
they
move
and
so
looking
at
the
physical
parameters,
and
so
we
have
that.
A
But
then
you
know
we
also
are
interested
in
how
they
sort
of
we
don't
want
to
say
think
necessarily,
but
how
they
sort
of
have
a
model
of
cognition
or
a
model
of
like
you
know,
behavior
and
so
we'll.
We
want
to
look
at
that
as
well
and
now,
of
course,
the
thing
is
in
most
theories
of
behavior:
they
don't
really
consider
morphology
and
but
in
bacillaria,
because
it
doesn't
really
have
a
brain.
It's
a
collective
set.
You
know
it's
a
bunch
of
cells,
doing
some
sort
of
collective
movement.
A
The
physics
become
much
more
important,
and
so
you
know
that's
that's
something.
To
think
about
as
well.
Is
that
we
need
to
think
about
how
we
can
you
know
maybe
lend
some
discovery
to
some
physical
parameters.
You
know
either
through
visualization
or
something
else
so
so
yeah
we
could
apply
some
of
those
methods.
A
So
we
have
a
bunch
of
I've
created
a
list
of
things
that
maybe
we
can
target,
and
these
are
you
know,
deadlines,
but
they're
also,
you
know
so
they're
things
that
have
deadlines,
and
you
know
we
want
to
make
sure
we
meet
them,
but
we
also
want
to
make
sure
we
know
kind
of
where
what
we
have
to
go
out,
and
so
this
is
a
list
that
I
came
up
with
for
diva
worm,
and
this
is
don't
worry
about
what
this
says.
This
is
another
group,
but
I
made
a
sep.
A
C
A
Okay,
so
now
everyone
should
have
access,
but
if
you
don't,
you
can
ask
for
them,
so
we
have
eight
things
on
the
list
so
far
and
again,
these
are
things
that
are,
you
know
we
might
meet
them.
We
might
not.
We
might
go
to.
You
know,
try
to
submit
whatever
this
in
this
submission
category
is
to
another
place.
A
A
evolution
is
a
meeting
that
is
basically
evolutionary
biology,
so
they're
interested
in
a
lot
of
topics
related
to
evolution
and
it's
going
to
be
online
this
year,
which
is
why
I,
you
know
point
it
out
to
people.
So
if
you
can,
you
know
you
think
there
might
be
some
interesting
ideas
we
could
pursue.
A
A
But
this
is,
you
know
it's
a
nice
opportunity
to
present
on
some
work
that
we've
done
in
the
group.
I'm
not
sure.
We've
done
anything,
that's
really
been
focused
on
evolution,
but
it
might
be
an
opportunity
to
to
do
something
submit
some
abstract
that
we
know
we
might
think
of
some
idea
that
we
want
to
submit.
A
A
So
if
you
go
to
this
link,
you'll
find
more
information
about
this,
and
it's
just
a
flash
talk.
I'm
going
to
give
with
you
know,
I'm
going
to
put
everyone
on
who's
done.
Work
on
diva
learn
as
a
co-author
and
then
it's
a
like
a
five
to
ten
minute
talk
on
the
stevo
learn
platform,
so
that
should
get
the
word
out
about
divalern
I'll,
just
basically
go
over
some
of
the
things
we've
talked
about.
A
You
know
in
terms
of
the
evil
learning
software,
some
of
the
stuff
with
the
devozu,
some
of
the
other
things
of
the
data
science
tutorials
and
how
it
can
be
used
for
education,
because
it's
they're
interested
in
online
education.
So
this
conference,
if
you're
interested
in
attending,
is,
I
think,
free
to
attend.
A
C
A
A
What's
that
one
okay,
this
is
it
I
think
so,
okay,
I
have
a
different
name
to
it
here,
but
this
is
the
thing:
we've
been
working
on
morphogenetic
patterns
in
the
theory
of
deep
learning,
so
this
is
an
abstract
we're
working
on.
I
was
thinking
of
submitting
it
somewhere
else,
but
that
didn't
happen,
so
we're
going
to
shoot
for
another
venue
when
working
on
this
abstract
and
again.
A
E
A
A
So
you
know
we
start
talking
about
the
special
properties
of
deep
learning
networks
and
what
they
do
and
then
you
know
how
they
relate
to
development.
So
so
it's
still
pretty.
I
think
we
need
to
go
through
a
couple
more
rounds
of
revision
on
this,
but
I
think
it's
it's
an
interesting
area.
I've
seen
some
things
also.
A
I've
talked
to
some
people
regarding
you
know
this
topic
and
it
seems
like
there's
something
there.
Nobody
knows
what
it
is,
but
we're
just
kind
of
at
the
cusp
of
understanding
what
it
is.
I
think
the
the
my
opinion
of
what
it
is
is
that
these
networks,
so
we
did
some
work,
we've
done
some
work
on
we're
using
cellular
automata
to
generate
patterns
that
look
like
embryos,
and
so
you
can
do
this.
You
can
use
these
one
of
these
cellular
automata
models.
You
can
model
like
some
sort
of
pattern
formation.
A
A
But
you
know
deep
learning
has
that
same
generative
quality,
and
so
you
know
there's
a
sort
of
a
parallel
there
between
the
two
generative
types
of
systems,
the
biological
one,
which
is
development,
and
this
deep
learning,
one
which
is
computational
and
so
there's
something
there.
That
is,
you,
know
kind
of
what
this
abstract
is
about,
unlocking
that
relationship,
and
so
we've
got
a
lot
of
citations
in
here
already
so
so
back
to
the
list
of
submissions.
A
A
So
you
know
if
we
want
to
do
you
know
if
we
want
to
get
in
touch
with
people,
maybe
who
are
interested
in
the
math
part
of
it,
but
not
necessarily
know
much
about
the
biology
part,
or
you
know,
they're
biologists
who
come
to
applied
math
meetings,
people
sometimes
biologists,
with
a
pretty
good
background
in
math.
So
this
is
so.
I
think
this
was
facilitate
a
lot
of
discussion
if
nothing
else,
and
so
this
is
the
link
to
this
conference.
A
A
A
That's
something
that
we
owe
to
mathematics
of
diatoms
and,
of
course,
if
you've
been
in
the
group,
you
know
that
we've
submitted
a
a
chapter
proposal
on
this.
We
also
did
a
talk
at
neuromatch
on
this,
which
I
think
established
because
it
was
well
received
by
you
know
the
people
there,
and
these
are
people
who
are
neuroscientists,
but
also
a
lot
of
molecular
biologists.
So
they
understood,
like
you,
know,
the
track
that
we
were
in.
They
understood
like
the
molecular
aspects
of
it,
but
they
were
actually
pretty.
A
I
thought
they
would
be
like
really
against
the
idea
of
non-neural
cognition.
They
actually
seem
to
like
it.
So
that's
that's.
Coming
up
and
again
I
can't
remember
who's
an
author
on
this
paper,
but
we
have
a
number
of
authors
and
we'll
be
circulating.
Drafts
so
there
might
still
be
a
chance
to
become
involved
with
this.
If
you're
interested-
and
I
know
usual
is
involved
and
I'm
involved-
and
I
I
think
maybe
susan's.
A
A
But
it's
it's
a
it's
a
biological
focus,
so
this
is
not
like
mathematics
or
modeling.
This
is
biology,
the
biology
of
c
elegans,
and
it's
really,
you
know
the
these
always
hosted
at
ucla
in
california,
and
they
would
always
have
a
nice
group
of
people.
A
couple
hundred
people
they'd
have
a
what
they
called
the
worm
show
you
can
check
it
out
online
where
they
did
this.
It
was
like
a
about
a
hour-long
skit
show
where
they
would
do
different
things,
thematic
things
around
c
elegans.
A
You
know
it
was
very
cute,
and,
and
so
it
was
a
it's
a
nice
community
of
people
and
they
it's,
it
basically
covers
all
the
biology
of
c
elegans
from,
like
you
know,
evolutionary.
You
know,
experimental
evolution
to
development
to
starvations.
You
know
where
they
study
different
mutants.
They
study
effects
on
mutants
like
starvation
or
behavior.
A
A
You
know
it's
like
a
clearinghouse
of
c
elegans
research,
because
it's
a
model
organism.
We
have
this,
there's
this
interest
in
using
c
elegans
for
a
lot
of
different
things.
So
now
open
worm
actually
had
a
workshop
at
the
c
elegans
meeting.
I
think
in
20
it
was
a
workshop
on
d
on
machine
learning,
and
so
I
don't
know
if
that's
available
on
the
open
worm
website,
but
I
can't,
if
you're
interested
in
that
I
can
direct
you
to
that.
A
I
don't
have
the
link
right
now,
but
yeah
so
openworm
did
a
machine
learning
workshop
there
and
it
was
actually
a
pretty
a
pretty
good
set
of
talks.
A
lot
of
it
was
on
just
simulating
c
elegans.
So
they
kind
of
said
it
was
machine
learning,
but
it
was
really
simulation,
but
yeah.
So
I
mean
we
should
that's
something.
We
should
definitely
do
because
there's
a
lot
of
exposure
we
had
in
the
cl
within
this
community.
A
I
did.
I
I've
actually
been
there
once
and
I
I
did
a
couple
of
posters
there.
So,
but
this
year
it's
going
to
be
virtual,
and
so
the
abstract
submission
is
march
25th.
So
if
we
think
of
something,
maybe
that's
more
biological
in
nature-
or
you
know
even
something-
that's
computational
but
is
very
grounded
in
biology,
we
might
submit
it
there's
for
an
abstract,
and
so
the
abstracts,
I
think,
are
like
posters
and
presentations.
So
you
know
we.
A
Maybe
if
we
get
a
couple
of
submissions
for
that,
maybe
we
can
work
it
down
into
two
or
three
submissions,
and
then
you
know
one
get
one
or
many.
You
know
a
couple
get
accepted
and
they
could
accept
as
a
post
or
presentation,
and
then
we
can
do
that
by
the
way,
if
you've
ever
done
a
poster
presentation
in
some
of
these
virtual
conferences,
they've
done
a
lot
of
they've.
A
lot
of
these
venues
have
done
a
lot
of
innovation
to
make
that,
like
nice,
a
nice
experience
probably
nicer
than
a
physical
experience.
A
You
know
you
present
your
poster
virtually
you
could
do
it
like
in
a
video
or
like
in
a
virtual
room.
I've
done
this
a
couple
times.
Actually
so
it's
it
can
be
done
and
we'll
see
how
they
handle
the
virtual
format,
but,
and
then
there's
this
evil
learned
paper
now
I've
I've
tagged
icml,
which
is
a
machine
learning
conference
for
this.
A
But
I'm
not
really
sure
about
that,
because
I
don't
know
if
it's
something
that
is
like
it
seems,
you're
a
meta
for
a
machine
learning
conference,
but
if
we
want
to
make
icml,
which
is
international
conference
of
machine
learning,
there's
a
call
for
papers
and
the
deadline
is
actually
very
soon
for
this.
So
it's
january
28th
for
the
abstract
and
then
february
4
for
the
paper,
it's
a
weird
setup,
but
basically
it's
like.
We
have
the
28th,
you
submit
an
abstract
and
then
you
have
submit
the
accompanying
paper
on
the
fourth.
A
I
don't
know
why
they
do
it
that
way,
but
it's
coming
up
very
soon.
If
we
want
to
do
this-
and
this
is
again
this
evil
and
paper
which
we
would
have
to
make-
I
think
we'd
have
to
expand
it
by
about
threefold.
To
do
this,
which
isn't
hard
it's
just
we
have
to
you
know
I
have
to
spend
some
time
on
it.
A
Going
through
and
we'll
have
to
do,
some
collaborating
on
that
to
make
sure
that
it
fits
the
the
format,
but
I
don't
know
it
might
be
a
bit
meta
for
something
like
that,
if
you're,
if
you're
interested
in
machine
learning,
let
me
know
about
that.
If
that's
something
that
you
are
interested
in
or
not,
if
we
should
try
to
make
that
it
in
any
case,
if
we
don't
make
that
deadline,
we
can
send
the
diva
learn
paper
somewhere
else.
A
It's
not
you
know
it's
not
the
end
of
the
it's,
not
the
only
thing
we
can
do
with
it
by
any
means,
so
it's
just
an
opportunity.
We
should
try
or
can't
you
know.
Maybe
we
could
try
and
I
don't
think
that
would
be
a
very
good
match
for
the
c
elegans
conference,
for
example,
but
it
might
be
better
for
like
a
machine
learning
conference.
A
A
F
A
Like
complex
networks,
things
like
that
those
are
things
that
they're
interested
in,
so
we
might
work
something
out
on
that
as
well,
so
stay
tuned
on
that,
I'm
not
really
sure
and
again
this.
This
shows.
I
think
this
list
how
this
is
useful.
You
know,
if
you
find
in
you,
find
an
event
or
find
some
opportunity.
A
A
A
A
A
A
A
I
think
this
is
actually
the
description
that
I
center
here.
This
short
one
I
and
then
what
I
did
was
I
put
a
link
to
this
document
in
the
description,
so
when
a
student
is
going
to
the
website
that
they're
going
to
put
it
on
they'll,
see
this
description
and
then
they'll
be
able
to
go
to
this
document
for
more
details.
A
So
these
documents
that
we
definitely
want
to
add
to
them
over
time,
because
students
are
going
to
be
engaging
with
us
on
these
proposals
and
we
want
to
give
them
as
much
information
as
we
can
about
it,
preferably
in
some
something
like
this,
which
is
like
a
doc
sort
of
approach
where
they
can
look
at
the
docs
and
get
some
ideas.
So
this
is
the
description
that
they're
going
to
get
on
the
website.
A
A
A
A
A
So
this
is
a
description
of
sort
of
the
idea
that
you
have
these
this
microscopy
data,
and
then
you
have
these
different
views
and
you're
basically
warping
the
views
to
create
this
continuous
or
this
mosaic
that
we
can
explore
and
I
didn't
really
put
down.
I
mean
I
said
you
must
be
proficient
in
c,
plus,
plus
or
python,
or
maybe
topological
techniques
and
it'll
be
up
to
the
student
to
propose
a
viable
solution.
Then
I
give
them
some
readings,
and
you
know
I
don't
know
we'll
see
if
people
apply
to
that.
So
now.
A
Incf
will
publish
these
ideas
sometime,
I
think
next
month
and
then
we'll
be
they'll,
be
asking
questions,
they'll
be
sending
inquiries
about
it
and
then
I
think
there's
like
a
discourse
like
I
said,
a
discourse
channel
where
they
ask
questions
and
then,
like
you,
have
to
answer
them
and
I
think
a
couple
of
us.
I
don't
know
if
someone
I
I
know,
I'm
a
part
of
the
discourse
channel
but
I'll
I'll
give
maybe
a
couple
other
people
access,
so
they
can.
A
Now
all
these
projects
might
not
get
accepted,
but
you
know
that's
that's
the
world
of
gsoc.
It's
also
a
bit
shorter
this
year,
so
we'll
see
what
we
can
actually
do.
That's
why
I
left
these
open-ended
because
the
then
there's
this
whole
thing
where
they
build
a
schedule
where
they
have
to
propose
a
solution
in
a
certain
amount
of
time.
A
Now,
there's
one
more
project
that
I
couldn't
get
it
in
the
shape
that
I
wanted
it's
for
this.
This
is
the
devonet.
This
is
the
one
that
krishna
proposed
and
it
actually
did
match
the
first
one
quite
a
bit,
so
I
didn't
want
to
submit
it
as
a
separate
project,
but
I'm
thinking
that
we
could
do
something
with
this.
A
I'm
not
sure
what,
because
I
don't
think
this.
The
way
it's
written.
It'll
fit
into
10
weeks,
and
I
couldn't
figure
out
a
way
to
do
it
like
like
to
make
it
match
into
the
because
you
have
to
have
like
a
you
have
to
propose
it
in
a
way
that
incorporates
something
neural
or
something
about
the
brain
very
explicitly
so.
A
But
I
think,
though,
that
we
can
use
this
for
like
something
else.
It's
got
some
nice
references
and
we
we
did
have
talked
about
like
we
have
this
pre-trained
models
theme.
So
I'm
not
sure
you
know
where
we
go
with
the
pre.
I
think
it's
definitely
something
we
should
keep
pushing
on,
but
I
don't
exactly
know
how
we'll
do
it,
but
we
need
to.
We
need
to
work
on
this.
I
think
a
little
bit
more,
maybe
tailor
it
to
one
of
the
opportunities
I
talked
about
conference
wise
or
you
know,
they're
always.
A
G
E
A
G
It
was,
I
guess
it
was
more
than
five
six
months
back
user
cycle
of
like
presentation.
I
think
so,
I'm
not
really
sure.
Okay,
you
presently
it
was.
You
know
we
can.
You
know
present
it
in
the
evolution,
because
it
was
showing
that
how
some
of
the
sea
animals
have
their
shell
in
the
shape
of
circles
and
that
sports
that
you
use
does
it
strike.
G
When
you
presented
it,
so
what
we
can
do
that,
I
guess
that
it
can
be.
You
know
very
well
versed
into
evolution
that
how
the
organism
you
know
are
evolving
and
the
geometry
played
a
key
role
in
how,
for
example,
that
how
their
shell
was,
you
know
was
that
was,
you
know,
considered
merely
for
protection,
but
it
had
deep
mathematics
beyond
that.
That's
why
that
shell
was
stable.
A
Yeah,
that
might
be
a
good
idea.
Yeah.
I
have
to
look
back
and
I
think
that's
a
theme
that,
like
actually
I'm
gonna,
move
to
some
papers
next
and
I
think
we're
kind
of
touching
on
that
theme
in
some
of
these
papers.
So
we'll
see
that
yeah.
But
I
remember
looking
back
at
the
list
from
last
year
and
there
a
lot
of
themes
that
yeah
we
haven't
gone
back
to
we've
kind
of
just
kind
of
touched
on
them
and
moved
on,
but
yeah
there's
a
lot
to
explore
in
those
areas.
A
A
A
Basically,
the
physics
of
development,
so
scientists
are
pushing
forward
their
understanding
of
the
role
of
mechanical
forces
in
the
body
from
embryo
to
adult,
and
so
this
is
a
zebrafish
embryo
and
this
is
an
imaging
where
they're
imaging
some
of
the
cells.
So
this
is
what
it
looks
like
in
a
certain
stage
of
development.
Here
you
have
the
head.
You
have
the.
A
And
you
have
the
part
of
the
body
here,
so
you
know
the
zebra
fish
is
going
to
extend
out.
This
will
be,
the
tails
will
be
the
head
and
but
they're
try.
What
they're
talking
about
in
this
little
feature
is
they're
talking
about
forces
and
physical
forces
that
you
know
push
this
embryo
push
the
cells
around
so
that
map
that
I
showed
you
in
the
in
the
paper.
A
You
know
it
has
these
little.
You
know
arrows.
It
looks
like
a
map
of
different
lines
radiating
out
from
dots.
So
what
that
would
be
would
be
one
of
these
cell
centroids,
which
are
these
colored
dots
here
and
it
would
be
if
it
divides,
and
it
has
two
daughter
cells.
They
move
in
a
certain
orientation
away
from
that
original
position,
and
the
question
is:
if
all
cells
are
doing
this,
you
know
what
and
what
you
know.
How
does
that
result
in
a
pattern
of
cells?
And
then
you
know
how
are
those
angles
governed?
A
A
A
A
This
is
a
nice
graph
for
pressure
to
develop.
This
is
an
figure
that
shows
a
couple
of
different
images
here.
This
is
a
mammalian
embryo,
that's
sculpted,
by
force,
but
from
pressurized
water
bubbles,
embryos
arrange
their
cells
around
a
cavity,
isolating
the
cells
that
become
the
fetus
bubbles
appear
and
then
pull
to
create
this
pattern.
A
B
B
B
A
Yeah,
that's
that's
good.
I
hadn't
even
thought
about.
Well,
I
mean
this
is
a
interesting
area
and
you
know
it's
again
one
of
these
things
that
we
don't
really
understand,
but
we
understand
some
of
the
you
know.
Some
of
the
things
are
starting
to
come
out
and
it's
starting
to
really
change
our
conception
of
how
this
process
works.
So
so
yeah.
A
Is
like
when
you
like,
take
a
piece
of
metal
and
you
heat
it,
and
then
you
can
start
hammering
it
into
a
different
shape.
So
you
have
this
heating
and
cooling
process
where
you
heat
the
the
metal
to
a
certain
temperature.
Usually,
like
you
know,
a
couple
hundred
degrees
and
then
you
hit
it
with
a
hammer
and
you
can
shape
it
because
it's
no
longer
really
stiff.
A
Process
of
annealing
and
they're
using
a
metaphor
for
it
here,
so
this
is
making
hearts
and
minds.
This
is
basically
talking
about
how
you
make
a
heart
and
a
brain.
Basically,
so
in
drosophila,
which
is
the
fruit
fly,
this
group
examined
heart
formation
and
embryos.
This
is
a
crucial
event
when
two
pieces
of
a
tissue
come
together
to
form
a
tube
that
will
ultimately
become
the
heart.
A
A
There's
this
process
where,
in
development,
these
things
work
very
synchronously
in
like
clockwork
in
some
ways,
but
there's
also
a
lot
of
error
correction,
because
it's
not
a
perfect
process.
So
we
want
to
know,
like
you
know,
is
this
a
process.
This
is
especially
interesting
for
like
simulation,
because
simulation
depends
largely
on
our
model
of
the
phenomena.
So,
if
we're
wrong
on
our
model
with
a
phenomenon,
then
it's
hard
to
really
kind
of
do
as
a
proper
simulation.
B
B
A
Yeah
and
there's
also
a
lot
of
work
on
like
regenerative
medicine
that
is
in
informative
to
sort
of
the
development.
If
people
heard
of
like
stem
cell
transplantation,
where
they
will
transplant
stem
cells
into
into
heart
muscle
and
stem
cells,
then
will
differentiate
into
heart
muscle
and
you
know,
serve
as
a
way
to
repair
heart
muscle.
A
So
they
can
do
things
like
that,
and
you
know,
relies
on
a
lot
of
signaling
between
cells
and
there's,
something
called
a
cell
niche,
which
is
where,
if
the
cells
are
together
and
forming
a
tissue,
other
cells
depending
on
the
organism
will
go
along,
and
so
that
can
actually
be
used
not
just
for
regeneration
but
also
like
they've
done.
Experiments
where
they've
taken
cells
from
other
parts
of
the
embryo
and
put
them
into
emerging
tissues
and
those
cells
have
essentially
changed
their
fate
into
that
new
fate.
A
So
a
lot
of
this
a
lot
of
development
is
about
like
what
they
call
cell
fate
and
there's
a
lot
in
terms
of
like
plasticity
and
things
like
that
that
are
applicable
so
yeah.
It's
definitely
a
really
interesting
area,
and
so
it
doesn't
just
involve
physics,
but
physics
is,
is
a
integral
part
of
this
so
again
you're
going
to
study
things
like
skin
cancer
as
well,
and
so,
when
you
study
cancer,
you're,
really
kind
of
studying
a
form
of
development.
A
That
might
be
a
little
weird
to
say,
but
you're,
basically
studying
the
same
types
of
processes
where
there's
like
a
lot
of
morphogenesis
and
there's
a
lot
of
changing
cell
fates
and
things
like
that.
But
you
know
physics
plays
a
role
in
this
as
well,
and
so
physical
forces
could
explain
why
some
tumors
are
benign
and
others
can
spread,
and
so
we
don't
really
know
why
that
is,
but
we
have
some
we're
getting.
I
think
better
at
understanding
that
process,
and
so
so
this
is.
This
is
again.
A
E
C
A
Yeah
yeah,
I
yeah,
I
think,
like
a
lot
of
the
stuff
that
we're
talking
about
is
really
kind
of
well
described
using
animations
and
I've
really
wanted
to
kind
of
bring
more
animations
into
the
group,
but
I
haven't
been
able
to
find
a
good
way
to
do
that
so,
but
I
mean
I
know
what
you're
talking
about
is
definitely
like
it's
cool,
but
I
wish
we
had
an
animation
of
it
like
we
could
see.
You
know
how
it
works.
A
A
So
there's
this
paper
discovering
the
power
of
single
molecules,
and
so
this
is
another
physic
physical
biology
paper.
This
one
is
mechanical
manipulations
of
single
biological
molecules
and
revealed
highly
dynamic
and
mechanical
processes
at
the
molecular
level.
Recent
developments
have
permitted
examination
of
the
impact
of
torque
on
these
processes.
A
Visualization
of
detailed
molecular
motions
enabling
studies
of
increasingly
complex
systems.
Here
we
highlight
some
recent
important
discoveries,
so
this
is
another
paper
where
we
have
like
that's
a
review.
It's
it's
a
high
level
review.
So,
if
you're
interested
in
some
of
these
different
things
that
single
molecules
are
doing
in
the
in
the
biology
of
of
in
biology,
then
this
is
a
good
paper.
A
So
they
talk
about.
You
know
largely
about
dna
based
motor
proteins,
so
these
are
proteins
that
are
involved
in
doing
things
like
unzipping,
helicase
and
and
in
transcription
and
a
lot
of
other
things.
So
this
is
it's
an
interesting
system
because
it's
sort
of
physics
at
a
different
scale
than
cells
and
they
talk
specifically
about
torque,
and
so
the
physical
properties
of
dna
are
critical
to
cellular
processes.
A
And
so
you
know
dna
is
wrapped
up
in
this
helicase.
It's
it's
not
only
twisted
into
a
double
helix,
but
it's
folded
up
into
this
into
this
into
the
nucleus,
and
so
you
know
you
have
to
be
able
to
unfold
it
unzip
it
copy
it
transcribe
it
and
then
take
all
that
zip
it
back
up
and
put
it
back
into
the
stuff
it
back
into
the
nucleus.
So
it's
a
lot
of
complicated.
A
A
A
Review
article
2021,
so
it's
brand
new,
it's
basically
sort
of
a
position
paper
on
what
machine
learning
can
do
for
developmental
biology
and
they
point
out
that
developmental
biology
has
grown
into
a
data
intensive
science,
so
you're
doing
a
lot
of
high
throughput
imaging
and
multi-omics,
which
is
like.
You
know
these
different
sequencing
techniques
and
I
guess.
C
A
Just
omics
is
just
a
catchy
term
that
they've
used
for
it.
So
now
we're
talking
about
applying
machine
learning
to
this
you
know
it
allows
us
to
make
sense
of
large
data
such
as
minimal
human
intervention,
they're
tests
such
as
image
segmentation,
super
resolution
microscopy
and
cell
clustering,
and
so
they
introduce
key
concepts,
advantages
and
limitations
of
machine
learning
and
some
application
domains.
So
this
is
one
person's
view
of
this.
A
A
So
their
focus
is
mainly,
though
on
microscopy
and
then
on
growing,
an
interdisciplinary
community,
so
the
challenges
of
talking
between
biologists
and
computer
scientists
who
specialize
in
machine
learning.
So
this
is
just
laying
out
the
landscape
and
I
think
it's
good
just
to
look
at
it.
If
you
really
are
not
clear
on
what
the
relationship
is
between
these
areas,
this
is
a
good
place
to
go
for
that.
A
So
it's
you
know
it's
it's
aimed
at
that
audience
and
you
know
so
it's
accessible
to
biologists,
but
it
also
kind
of
lays
out
that
landscape
of
how
we
apply
machine
learning.
Now,
that
being
said,
it's
not
the
ultimate
truth.
I
think
we
have
in
our
group.
We
have
a
lot
of
different
directions
that
are
beyond
the
scope
of
this
review.
So
it's
it's,
but
it's
nice,
it's
a
broad,
open
area.
I
think
there
are
a
lot
of
opportunities
to
sort
of
stake
out
a
space
in
that
area.
So.
D
E
G
G
H
A
A
It's
a
very
interesting
book
because
you
know
darcy
thompson.
This
was
kind
of
like
almost
pre-darwinian.
In
some
ways
he
didn't
really
integrate
a
lot
of
what
darwin
was
talking
about,
but
yet
there's
a
lot
that,
like
you
know,
you
could
integrate.
You
know
natural
selection
and
things
like
that
with
this
kind
of
work.
A
It's
definitely
a
different
area
and
there's
been
some
work
following
up
on
it
as
well
that
we
could
talk
about
you
know,
but
also
this
sort
of
idea
of
geometry
as
being
important
and
and
these
ratios
as
being
important,
which
people
have
talked
about
things
like
elementary
scaling
and
things
like
that.
But
it's
like
there's
something
more
there.
It's
it's
still
kind
of
like
it
still
isn't
fully
integrated.
I
think
there's
something
here:
yeah,
there's
something
very
interesting
in
this
area.
It's
just
a
question
of
like
what
exactly
it
is.
G
A
A
A
A
A
And
so
then
you
can
describe
the
difference
in
the
morphology
and
you
know
people
I
mean
I
don't
know
how
much
people
I
imagine
in
morphometrics
people
use
things
that
are
similar,
but
I'm
not
sure
if
people
have
actually
explored
like
the
relationship
between
these
mappings,
like
it's
a
little
weird
to
think
about.
But,
like
you
know,
if
you
have
like
say
one
mapping
versus
another
mapping,
how
do
you
describe
that
relationship
like
is
it
like?
B
B
B
B
C
B
E
B
B
B
E
B
B
B
A
C
A
Okay,
okay,
so
that's
all
we
have,
I
think
we're
done
for
today.
Again,
if
you
have
anything
to
add
to
the
list
of
things
or
you
know,
maybe
I'll
add
the
thing
on
darcy
thompson
or
the
on
growth
and
forum,
and
then
I
think,
if
krishna
wants
to
do
his
presentation
on
his
stuff
next
week,
we
can
do
that
as
well.