►
From YouTube: DevoWorm (2021, Meeting 18): Neural Cellular Automata, Regeneration, and Developmental Collinearity.
Description
Presentations on Neural Cellular Automata and Regenerative Capacity (with discussion). Differentiation tree finding and regeneration, Mathematics of DevoWorm, HOX collinearity, body plans, and shape. Attendees: R Tharun Gowda, Vrutik Rabadia, Jesse Parent, Richard Gordon, Susan Crawford-Young, Krishna Katyal, Bradly Alicea, Shruti Raj Vansh Singh, Yash Vadi, and Mainak Deb.
A
B
B
A
B
C
C
B
A
A
Anyways,
so
welcome
to
the
meeting
it's
a
lot
smaller
today.
Well,
maybe
people
will
come
and
join
us
later
in
a
little
bit
but
anyways
I
wanted
to
get
started.
So
what
do
I
do
things?
I
was
gonna
do
first
here,
let's
see,
why
don't
I
go
into
the
submissions?
First,
let's
see
where
we
are
with
that.
A
Okay,
so
oh,
we
got
another
person,
oh
here's,
dick
hi,
dick,
so
we've
got
a
bunch
of
things.
Let's
see
if
we
have
what
we
have
outstanding,
what
we
have
submitted.
So
we
have
the
the
netsigh
or
network
2021
submissions,
those
the
embryo
networks
and
connectome
submission
was
accepted
and
that's
being
worked
on
right
now.
A
We
also
have
the
growth
form
and
theory
of
deep
learning,
abstract
that's
submitted
to
net
neuroscience
and
if
it
doesn't
make
it
in
there
we
can
submit
it
somewhere
else
pretty
easily.
You
can
turn
around
and
turn
it
around
and
send
it
somewhere
else,
and
also
this
euler
cycles
for
life
submission
to
topo
nets,
which
again
I
got
it
down
to
a
two
page,
abstract,
and
so,
if
that
doesn't
make
it
in
there,
it's
it's
a
lot
easier
to
sort
of
expand
out
and
submit
somewhere
else.
A
D
A
Model
of
the
embryo,
and
then
this
these
are
von
neumann
neighborhoods
for
cellular
automata.
You
know
those
again
we
can
use.
Each
of
these
is
like
a
figure
write,
some
something
underneath
it
describing
it
a
little
bit
and
why
it's
useful,
and
then
you
know
that
would
make
some
sort
of
document
that
we
would
submit
somewhere
there.
You
know
we
can
add
additional
different.
You
know
we
can
add
different
things
into
these
little
ovals,
so
I'm
trying
to
think
of
a
another
set
of
equations.
A
That
might
be
useful,
maybe
like
some
sort
of
reaction
diffusion
model.
If
we
want
to
add
that
in
and
so
that,
I'm
you
know,
I'm
just
kind
of
working
out
in
this
poster
format,
just
to
get
it
into
the
and
getting
into
the
shape
of
like
thinking
about
how
to
organize
these
things.
So
I
mean
it's
not
going
to
end
up
like
looking
like
this,
I
might
make
a
we
might
make
a
poster,
that's
a
little
bit
less
busy
but
anyways.
I
just
wanted
to
point
that
out.
F
E
C
E
A
Oh
yeah,
so
we
have
dick
and
josh
and
verdick
and
trudy
here.
So
hello
we're
just
going
over
the
submissions.
I
want
to
do
that
first,
so
we
have
those
things
and
then
we
have.
Let's
see
we
have
the
boring
billion,
which
is
a
potential
book
contribution,
and
this
work
is
about
this
idea
of
you
know
long-term
revolution.
A
What
was
happening
a
billion
years
ago
on
earth,
and
how
did
we
get
things
like
you
know,
developmental
systems
or
complex
life?
So
that's
that's
the
scope
of
that
area.
Boring
billion
specifically
looks
at
that
long
period
between
the
sort
of
the
origin
of
life
and
then
the
emergence
of
maybe
something
more
complex,
like
a
eukaryotic
cell.
So
there
was
this
period
of
life
where
it
was
very,
very
simple
and
why
you
know.
Why
is
that?
Why
is
it
the
case
that
it
was
like
that
for
so
long
yeah,
george
nikolowsky.
A
So
that's
that's!
You
know,
that's
actually
something
we
might
do
a
book
or
you
know
getting.
You
know
somewhere
moving
forward
on
that,
a
little
bit
more.
Also,
there's
this
kindle
book,
which
is
an
amazon
book.
This
is
something
I
think
krishna
mentioned,
but
this
is
for
these.
Let
me
make
a
note
here.
A
But
that's
so
that's
the
main
conference
and
again,
like
I
said
it's
pretty
competitive
but
it'll.
You
know
if
you
don't
get
get
it
in
there,
which
I
mean
you
know
you
probably
won't.
You
can
submit
it
to
one
of
their
sort
of
their
satellite
workshops
are
often
good
or
you
know,
submitted
somewhere
else
with
some
reviews
so
we'll
see.
If
anyone
wants
to
submit
to
that,
you
know
go
ahead
and
do
so.
A
A
It's
run
by
a
group
of
people,
a
largely
european
collaborators,
but
they're
doing
a
lot
of
different
workshops
on
these
bio-robotic
systems
and
you
might
be
interested
if
you
have
something
in
that
area.
You
might
be
interested
in
submitting
something.
So
I
think
they
want,
like
a
2000
word
submission
by
the
end
of
the
month
by
may
30th,
and
you
have
to
submit
to
the
workshop
chairs
so
go
to
the
website
and
they
have
the
workshop
chairs
listed
there.
A
E
A
A
A
You
know,
an
organism
that
expresses
different
modes
of
development
that
come
from
different
organisms,
and
so
the
question
is
you
know
it?
Can
you
see
this
in
its
gene
in
organisms
genome
and
you
know,
there's
a
whole.
You
know
we
can
talk
more
about
this.
This
is
a
little
bit
deeper
sort
of
problem.
So,
if
you're
interested,
we
can
talk
more
about
the
background
on
this
requires.
E
A
A
Okay,
so
that's
yeah
and
then
there's
this.
These
molecular
level
simulations
of
diatoms,
so
the
stuff
that
we've
been
doing
with
diatoms
and
this
meeting
has
mainly
been
on
the
cellular
level,
but
there
are
also
molecular
level
simulations
that
can
be
done,
and
so
that's
what
this
entry
is
yeah
we're
accepting
that
now
yeah.
Okay,
so
that's
that's
a
outstanding
submission,
so
yeah.
If
anyone
has
anything
to
add,
let
me
know
so
next
I'd
like
to
invite
anyone
who
wants
to
present.
A
E
A
A
B
B
F
H
H
H
H
All
these
images,
you
can
see
that
it
starts
from
a
single
point,
the
initial
input
image
that
is
just
one
pixel
in
the
in
the
whole
matrix
one
one
cell
and
it
grows
and
grows
to
us
to
a
point
till
it
looks
like
not,
if
not
exactly
because,
of
course
it
requires
some
more
fine
tuning
but
yeah.
It
grows
to.
H
A
A
H
H
A
H
E
A
E
A
E
A
E
A
A
E
A
E
A
A
B
A
F
A
And
then
you
know,
the
other
thing
too
is
like
the
gridlines
that
darcy
thompson
used
essentially
their
coordinate
systems.
So
I
know
that
people
who
applied
to
project
number
three.
I
had
two
applications
and
they
both
use
different
versions
of
coordinate
systems,
and
so
that's
something
that
you
know
we
can
follow
up
on.
You
know
talking
about
because
you
know
you're,
basically
taking
images
and
projecting
them
to
a
coordinate
system,
and
then
all
that
would
be
would
be
like
you'd
have
maybe
two.
You
know
two
different
sets
of
images.
A
Maybe
two
different
species
of
embryo
you'd
map,
those
to
a
coordinate
system,
and
then
you
just
compare
those
coordinate
systems
or
you'd
trans
you'd
make
some
sort
of
transformation
from
one
to
the
other.
So
that's
how
that
would
work,
and
then
you
could
do
the
same
thing
with
cellular,
automata
and
images
of
an
embryo.
A
You
could,
just
you
know,
find
major
points
and,
and
you
know,
put
them
into
a
coordinate
system
once
you
have
it
in
the
coordinate
system,
then
you
can
do
the
mathematical
transformations
in
that
system,
and
I
can't
remember
the
two
systems
that
we
had.
I
know
krishna
was
one
of
the
applicants
and
he
did
some
work
on
that.
But
anyways
we'll
talk
more
about
that.
Maybe
next
week
I'll
have
to
pull
up.
Maybe
some
information
more
information
about
that.
But,
okay,
so
does
anyone
else?
Have
anything
else
he'd
like
to
present.
A
G
D
I
A
J
J
D
D
J
J
J
C
A
Young
children
can
regenerate
lost
fingertips,
yeah,
there's
a
if
you're
young
enough
and
you
get
your
tip
of
your
finger
cut
off.
You
can
regenerate
it,
but
of
course
you
can't
do
that
after
a
certain
age-
and
you
can't
read-
I
mean
there's
certain
things:
humans
can
regenerate
like
skin,
and
you
know
healing.
A
So
so
you
know
it's
like
there's
something
about
like
age
and
the
type
of
organ
that
it
is,
or
maybe
what
the
sort
of
the
molecular
program
of
the
cell
is
so
like
liver
is
pretty
what
they
call
pluripotent,
which
means
it's
sort
of
its
gene
expression
profile
is
a
lot
closer
to
say
what
you
would
find
in
a
stem
cell
or
in
a
cancer
cell
than
the
rest
of
your
organs.
So
the
regenerative
potential
is
maybe
a
little
bit
higher
and
they
don't
know
it
could
be
yeah
a
lot
of.
J
J
A
Aside
from
like
synaptic
plasticity
that
you
can
see,
so
I
mean
there
are
a
lot
of
different.
This
is
a
pretty
interesting
area
and
people
are
always
trying
to
figure
out
like
you
know.
Obviously,
if
you
could
regenerate
your
limbs
or
damaged
organs
or
whatever
it
would
be
great,
but
people
are
still
trying
to
figure
out
how
that
works,
and
so
the
yeah
one
way.
E
A
G
A
A
They
think
that
you
know
if
you
stimulate
the
spinal
cord
enough,
the
cells
in
the
spinal
cord
that
they'll
start
to
regenerate
and
sort
of
move
like
you
know
how
our
nerves
and
axons
usually
do
in
development,
where
they
start
to
migrate
towards
each
other
and
make
a
connection,
and
so
they
they
try
to
do
all
the
sort
of
mimicry
through.
You
know
like
biogels
and
things
like
that.
A
D
A
E
A
I
A
E
F
G
E
A
A
B
A
That'd
be
great,
be
great
and
then
jesse
says
I'd
be
interested
in
that
kind
of
tree,
finding
our
functionality
of
regeneration,
yeah.
We
can
talk
more
about
that
and
then
krsna's
shruti
says
krishna.
You
are
the
one
with
back
issues:
you'd
be
a
better
match,
so
that
was
referring
to
regenerating
spinal
cord
or
whatever
jesse's
is
volunteering
versus
volunt
folding
or
volunteering?
A
D
A
A
Diseases
and
a
lot
of
them
try
different
things
like
they
want.
You
know
you
would
you
would
try
anything
to
sort
of
regain
your
function
or
to
stay
alive
and
sometimes
people
you
know,
there's
a
lot
of
you
know
a
lot
of
times.
It's
ethical,
you
know,
and
sometimes
it's
experimental
and
sometimes
people
are
just
not
ethical
about
it.
So
it's
like
the,
especially
in
like
stem
cell
biology.
There
are
some
people
who
are
not,
as
you
know
so,
and
you
know,
there's
a
whole
set
of
issues
there.
A
Dick
said
how
about
your
regeneration
paper
cited.
So
can
you
krishna?
Can
you
create
a
list
of
well?
I
think
you
had
the
list
there
of
the
papers,
but
can
you
get
it
into
a
more
formal
format?
That'd
be
great
yeah,
okay,
so
I
wanted
to
go.
Actually
I
wanted
to
go
to
the
papers
now
and
I
have
something
that
we
just
we're
talking
about
axolotl
and
I
found
something
else
about
axolotls,
so
this
is
kind
of
a
follow-up
here.
A
It's
about
metamorphosis
of
axolotls,
so
this
is
old
paper
by
julian
huxley
in
nature.
1920,
it's
metamorphosis
of
axolotl
caused
by
thyroid
feeding,
so
this
isn't
regeneration,
but
this
is
actually
where
you're
sort
of
at
a
certain
point
in
development
and
you
feed
the
you
feed
the
axolotl,
a
certain
type
of
food
and
it
metamorphoses
into
a
different
phenotype.
A
A
In
this
case,
if
they're
eating
this
diet,
the
special
diet,
they
actually
do
metamorphose
into
a
different
phenotype,
and
so
this
is
the
phenotype
that
it
has,
and
I
think,
we've
well
kristen
showed
some
pictures
of
the
axolotl
under
normal
circumstances,
but
this
is
the
metamorph
metamorphic
version
of
the
axolotl,
so
that
I
mean
I
thought
I'd
bring
that
up,
because
it
was
a
it's
an
interesting
thing.
We're
talking
about
regeneration,
but
there's
also,
this
issue
of
metamorphosis
and
metamorphosis
is
an
interesting
topic.
That's
just
one
form
of
it.
D
E
A
Right
and,
of
course,
the
the
stuff
that
michael
levin
is
done
with
with
flat
worms.
You
know
that
he's
been
able
to
take
like
in
a
flatworm.
You
can
just
blow
away
everything,
but
a
single
cell,
and
then
a
single
cell
can
become
a
new
flatworm.
But
what's
interesting
is
those
flatworms
have
like
this
associative
memory?
That's
retained,
even
if
you
just
take
a
single
cell
of
the
old
flatworm
and
regenerate
a
new
flatworm.
A
I
I
E
D
A
Okay,
yeah,
that
would
be
interesting
yeah.
So,
let's
see
I'll
go
to
another
paper
here
we
had
some.
Actually
I
wanted
to
mention
this.
I
don't
know
if
sherdy
saw
this
yet
so
there
was
that
paper
she
just
presented
on
the
one
on
aerocellular
automata,
and
I
posted
this
in
the
slack
this
week.
This
is
the
follow-up
paper
on
that.
A
So
this
is
the
adversarial
reprogramming
of
neural
cellular
automata,
and
this
is
this
is
where
they
take
those
neural
ca
models
and
they
use
adversarial
data
to
train
them,
so
they
use
different
types
of
inputs
that
are
adversarial,
meaning
that
they're
intentionally
sort
of
cryptic
or
they're
intentionally
wrong,
and
so
this
is,
of
course
we
do
this
in
neural
networks,
to
you
know,
train
them
against
attacks
or
against
false
positives
and
so
adversary.
So
they
have
an
adversarial
mnist
cellular
automata
here,
so
they
have.
A
They
introduce
information,
not
just
the
training
information
that
looks
like
the
training
set,
but
examples
that
aren't
explicitly
aren't
supposed
to
be
part
of
the
training
set,
but
they're
they're
used
as
sort
of
part
of
the
training
set.
So
you
know
an
eight
is
what
you
may
be,
what
you
would
achieve
as
a
desired
morphology,
but
maybe
like
a
7
that
looks
kind
of
like
a
t.
A
You
know
that
that's
not
really
what
we
want,
so
you
would
introduce
that
as
well,
and
then
the
system
would
have
to
differentiate
between,
say
something
like
that
and
like
an
actual
seven.
So
this
is
this
kind
of
goes
through
a
lot
of
this
they're
they're
doing
this.
They
say
in
this
experiment.
The
goal
is
to
create
an
adversarial
ca
that
can
hijack
the
cell's
collective
classification
consensus
to
always
classify
innate,
so
in
this
case
they're
using
eight
as
a
way
to
like
sort
of
hijack
what
the
cells
are
doing.
A
So,
no
matter
what
the
cell
is
encountering,
it's
always
classifying
an
eight
for
some
reason,
and
this
is
of
course,
if
you
had
like
a
you
know,
I
mean
you
can
even
think
in
the
biological
context.
If
a
cancer,
you
know
a
bunch
of
cancer
cells
wanted
to
hijack
the
biological
system
in
a
in
a
in
an
organism
to
take
over
or
a
virus
wanted
to
hijack
the
dna
replication
machinery
of
a
cell.
A
You
know
those
are
kinds
of
adversarial
attacks
on
the
on
the
biological
system,
so
they're
they're
kind
of
looking
at
this
problem,
and
I
think
it's
really
ingenious
because
it's
you
know,
I
mean
we
don't
really
think
about
things
like
I
just
mentioned
as
adversarial
attacks,
but
they
really
are
on
a
biological
system
and
then
how
do
you
model
that
in
something
that
you
can
see
like
the
output?
This
is
a
really
interesting
extension
of
this
work.
So
I
don't
know
surety.
This
is
a
distilled
article.
A
A
A
A
A
So
this
this
is
like
really
busy,
but
I
think
you
can
get
the
idea
here.
Is
that
we're
going
to
be
talking
about
something
called
the
body
plan
and
the
body
plan
goes
back
to
the
german
term
ball
plan,
which
is
this
idea
that
there
are
certain
sort
of
architectures
that
emerge
in
development
that
defines
certain
groups
of
organisms
and
so
they're
as
a
small
in
a
relatively
small
number
of
body
plans
in
the
tree
of
life
relative
to
all
the
diversity
that
we
see,
and
so
these
are
conserved
throughout
evolution.
A
So
there's
a
common
origin
to
each
body
and
those
body
plans
emerge
pretty
early
in
evolution
as
it
turns
out.
They
emerged
some
somewhere
back
in
you
know
pretty
far
back.
You
know
at
least
several
million
years
to
hundreds
of
millions
of
years.
They
don't
have
any
incentive
to
change
because
they're
successful
solutions.
A
And
so
the
the
question
that
comes
up
here
is
that
you
have
this:
you
know
how
do
you,
why
is
this
conserved
and
how
does
this
sort
of
map
to
the
genetic
substrate?
So
here's
a
segmented
organism
here.
This
is
like
some
sort
of
segmented
organism
and
you
can
see
these
circles
and
they're
lettered
a
through
e,
and
so
these
lettered
circles
are
genes
that
are
ordered.
A
A
A
So
this
co-linearity
this
this
physical
linkage
is
actually
it
enforces
this
fitness.
So
if
you
move
the
head
state
of
the
middle,
the
organism
would
not
be
able
to
survive,
because
you
know
you
can't
really
have
a
head
in
the
middle
and
survive.
These
things
have
to
be
in
a
certain
order.
So
so
this
paper
kind
of
gets
into
this
idea
that
maybe
physic
physics
has
a
role
to
play
here
so
hox
gene
collinearity
is
a
multi-scalar
property
of
many
animal
phyla,
particularly
important
in
embryogenesis.
A
A
This
hox
gene
collinearity
is
observed
to
spatial
collinearity
when
the
hox
genes
are
located
in
the
order.
Hox
1,
hox,
2,
ox
3,
so
they're
all
sort
of
ordered
in
that
sequence
and
then
corresponding
sequence
of
ontogenetic
units,
so
the
antigenetic
units
are
basically
the
mapping
to
the
different
segments
of
the
organism.
So
that's
this
e1
e2
e3,
and
so
this
is
located
along
the
anterior
posterior
axis
of
the
embryo.
As
you
saw,
you
go
from
the
head
to
the
tail,
so
expression
of
hox
1
occurs
in
e1,
hops,
2
and
e2
and
so
forth.
A
Besides
this
spatial
collinearity,
a
temporal
collinearity,
has
also
been
observed
in
many
vertebrates.
So
this
means
that
there's
a
collinearity
in
terms
of
time,
so
that
means
that,
in
this
case
the
first
hox1
or
this
first
module
cox1
is
expressed
in
e1
and
then
later
on,
hox2
is
expressed
in
e2,
followed
by
hox3
is
expressed
in
e3.
A
So
that
means
that
what
happens
in
this
diagram
is
that
this
a
gene
is
expressed
first
and
this
whole
thing
is
like
a
single
unit.
You
know
it's
in
the
embryo,
but
the
a
gene
is
expressed
first,
so
the
head
is
the
genes
that
define
the
head
are
expressed
first
and
then
the
genes
that
define
the
second
segment
are
expressed
next,
and
so
you
get
this
definition
from
head
and
then
eventually
to
tail
where
the
cells
are
sort
of
given
the
signals
to
differentiate
and
form
structure.
A
So
that's
what
they're
finding
here
they
use
a
biophysical
model
to
test
this.
They
say,
according
to
the
biophysical
model,
physical
forces
are
created
which
pull
the
hox
genes,
one
after
the
other,
driving
them
to
a
transcription
factory
domain
which
in
which
they
are
transcribed,
and
this
is
a
lot
of
molecular
biology
jargon,
but
basically
the
hox
genes
are
pulled
as
a
in
in
during
their
expression
into
the
machinery
and
they're.
This
is
done
sequentially.
A
So
this
is
a.
This
is
a
simulation
study
and
they're
just
kind
of
trying
to
test
some
of
these
ideas.
Symmetry
is
a
physical
mathematical
property
of
matter.
It
was
explored
in
depth
by
no
other
formulated
than
a
groundbreaking
theory.
That
applies
to
all
sizes
of
matter,
and
then
they
just
kind
of
talk
about
this.
So
they
use
that
as
a
way
to
use
hox
gene
collinearity.
A
It
explains
the
origins
of
this
not
only
along
the
ap
axis,
like
I
showed
you,
but
also
in
animals
with
circular
symmetry,
and
there
are
all
sorts
of
marine
invertebrates
that
have
circular
symmetry.
So,
instead
of
having
this
anterior
posterior
symmetry,
they
also
have
a
circular
symmetry
where
this
head
and
tail
meet,
and
then
there
would
be
like
maybe
four-fold
symmetry
where
you'd
have
you'd
divide
your
circle
up
into
different,
you
know
slices
and
then
each
of
those
slices
would
be
they
would
correspond
to
a
certain
hox
gene.
A
So
that's
the
idea,
and
so
they
get
into
this
they
kind
of
show
their
work
here.
They
show
an
example
in
this
figure
they
show
concentration
of
the
gene
product
from
each
hops
gene.
They
show.
You
know
other
things
here.
They
kind
of
show
this
at
the
micro
scale
and
macro
scale,
and
then
they
they
have
this
elastic
spring
expansion
model
for
their
physical
model,
and
then
they
show
okay,
then
they
show
the
circular
organization
of
volcano.
Derms
is
an
example
of
circuit.
A
You
know
the
circular
symmetry
they
show
the
barnsley
leaf
as
a
self-similar
design.
So
the
idea
being
that
like,
if
I
took
this-
and
this
wasn't
shown
in
this
figure
too
much.
But
if
you
go
to
the
barnsley
fern,
which
her
barnsley
leaf,
which
is
a
fractal
structure,
you
can
imagine
each
one
of
these
segments
having
their
own
set
of
hox
genes,
which
is
a
little
bit.
I
don't
know
if
that
actually
happens
in
nature,
but
they're
using
this
as
an
example.
A
So
you
know
each
of
these
branching
mechanisms
would
be
a
hox
gene
and
then
each
of
these
branching
mechanisms
would
have.
It
would
have
the
same
hox
gene,
but
it
would
have
a
different
expression
pattern,
and
so
you
can,
you
know
you
can
repurpose
these
hox
genes
for
different
scales
of
of
morphogenesis.
A
So
that's
all
I
wanted
to
talk
about
that
paper.
The
other
one
I
wanted
to
talk
about,
then,
is
this
a
paper
development,
origin
of
animal
body
plans,
and
so
again
I
s
like,
as
I
said,
the
body
plan
is
comes
from
this
term
ba
plan,
which
is
a
german
term,
and
the
german
embryologist
coined
it
to
describe
these
basically
very
similar
types
of
organization
across
large
parts
of
the
tree
of
life,
and
so
they
this
paper
is
pretty
long,
but
they
talk
a
lot
about.
They
kind
of
lay
out
some
of
these
factors.
A
They
talk
about
hux
clusters,
which
are
these
groups
of
hox
genes.
They
talk
about
head
formation,
eye
formation,
patterning
and
this
kind
of
gets
really
deep
into
the
technical
aspects
of
how
these
hox
genes
and
how
these
types
of
mechanisms
relate
to
body
plans
and
sort
of
the
development
of
body
plans.
A
And
then,
of
course,
here
you
have
major
developmental
innovations
leading
to
the
origin
of
bilateral,
so
one
of
these
body
plans
comes
with
a
bunch
of
innovations
that
occur
different
points
in
the
tree
of
life.
So
you
have
these
different
fact.
These
different
things
that
sort
of
enable
development
of
bioteria
and
then
that
enables
these
body
plants
to
emerge.
A
Yeah,
so
this
is
a
pretty
long
involved
paper,
but
I
wanted
to
mention
it
because
they
do
go
through
a
lot
of
different
types
of
tissue,
morphogenesis
and
and
other
things
in
in
this
same
vein,
in
the
same
vein
as
this
model
so
and
then
finally,
I
wanted
to
see
if
we
have
anything
else
in
here,
I
wanted
to
talk
about.
Maybe
I
will
talk
about
this,
so
this
is
a
blog
post
that
came
up
this
week
and
this
is
on
the
node.
A
So
this
is
something
that
we've
talked
about.
In
the
past,
we
published
a
blog
post
on
this.
This
is
a
blog
run
by
the
society
for
developmental
biology,
and
this.
This
blog
post
talks
about
embryomorphogenesis
is
a
play
whose
outcome
is
result
is
the
result
of
a
complex
and
delicate
plot
made
of
balances
and
agreements
among
many
actors,
the
execution
of
the
genetic
program,
biochemical
communication
among
cells,
mechanical
forces,
energy
consumption
geometry
and
all
this
other
stuff.
A
This
process
is
by
no
means
smooth
and,
in
some
developmental
stages,
dramatic
sudden
shifts
in
the
properties
of
the
structure
of
the
embryo
occur.
So
they
talk
about
this.
This
is
sort
of
their
research
experience,
they're
talking
about
some
of
these
sudden
shifts
and
and
how
they're
investigating
some
of
these,
and
so
they,
you
know
they
kind
of
talk
about
their
case
and
actually
they
get
into
this
issue
of
phase
transitions.
A
A
They
know
the
viscosity
of
the
embryo
tissue
drops
by
more
than
an
order
of
magnitude
when
morphogenesis
starts.
So
this
when
the
cells
invade
the
yolk
and
take
it
over.
There's
this
sort
of
phase
transition
from
a
solid
non-deformable
state,
a
fluid,
highly
deformable,
and
so
that's
their
sort
of
motivating
issue,
and
then
they
apply
this
model
of
phase
transitions
to
it
and
they
have
a
number
of
different
papers
here.
A
So
if
this
is
an
area
you're
interested
in
there's
a
nice
bibliography
here-
and
they
also
have
some
of
the
papers
that
they're
referring
to
in
here-
so
this
is-
I
don't
know
if
the
link
is
in
this
pdf-
oh
here
it
is
so
it's
here
I'll
just
put
it
put
it
in
the
chat
for
you
and
so
there
it
is,
and
then
let
me
go
through
the
chat
messages
here.
We
have
a
couple
all
right,
so
yeah
I
had
to
leave
now.
Thank
you,
shreedy
for
sharing
slides
a
great
week.
A
Thank
you.
Yosh
for
attending
rudick
elsa
has
to
go.
Is
an
amazing
presentation
shruti.
So
again,
thank
you
verdict,
susan
says
I'll,
get
you
a
review
paper
and
face
transitions,
tissue,
rheology
and
embryonic
organization.
So
I
yeah
thanks
that
you
should
send
me
that
and
then
this
is
the
link
to
the
node
blog
post.