bioRxiv (Cold Spring Harbor Laboratory),
Journal Year:
2024,
Volume and Issue:
unknown
Published: May 17, 2024
Size
estimation
is
a
hard
computer
vision
problem
with
widespread
applications
in
quality
control
manufacturing
and
processing
plants,
livestock
management,
studies
on
animal
behaviour.
Typically,
image-based
size
facilitated
by
either
well-controlled
imaging
conditions,
the
provision
of
global
cues,
or
both.
Reference-free
challenging
because
objects
vastly
different
sizes
can
appear
identical
if
they
are
similar
shape.
Here,
we
attempt
to
implement
automated
reference-free
body
facilitate
large-scale
experimental
work
key
model
species
sociobiology:
leaf-cutter
ants.
Leaf-cutter
ants
suitable
testbed
for
size-estimation,
their
workers
differ
both
shape;
principle,
it
therefore
possible
infer
mass,
proxy
size,
from
relative
proportions
alone.
Inspired
earlier
E.O.
Wilson,
who
trained
himself
discern
ant
worker
visual
cues
alone,
used
various
deep
learning
techniques
achieve
same
feat
automatically,
quickly,
at
scale
single
reference
image:
Wilson
Only
Looks
Once
(WOLO).
Utilizing
over
3
million
hand-annotated
computer-generated
images,
set
neural
networks
-
including
regressors,
classifiers,
detectors
were
estimate
mass
image
cut-outs.
The
WOLO
approximately
matched
human
performance,
measured
small
group
experts
non-experts,
but
about
1000
times
faster.
Further
refinement
may
enable
accurate,
high-throughput,
non-intrusive
weight
scale,
so
eventually
contribute
more
nuanced
comprehensive
understanding
complex
division
labour
that
characterises
polymorphic
insect
societies.
Philosophical Transactions of the Royal Society B Biological Sciences,
Journal Year:
2025,
Volume and Issue:
380(1922)
Published: March 20, 2025
Division
of
Labour
(DoL)
among
group
members
reflects
the
pinnacle
social
complexity.
The
synergistic
effects
created
by
task
specialization
and
sharing
duties
benefitting
raise
efficiency
acquisition,
use,
management
defence
resources
a
fundamental
step
above
potential
individual
agents.
At
same
time,
it
may
stabilize
societies
because
involved
interdependence
collaborators.
Here,
I
review
conditions
associated
with
emergence
DoL,
which
include
existence
(i)
sizeable
groups
enduring
membership;
(ii)
improving
performance;
(iii)
low
conflict
interest
owing
to
correlated
payoffs.
This
results
in
(iv)
combination
intra-individual
consistency
inter-individual
variance
carrying
out
different
tasks,
creates
(v)
some
degree
mutual
members.
DoL
typically
evolves
‘bottom-up’
without
external
regulatory
forces,
but
latter
gain
importance
at
later
stage
evolution
Owing
feedback
processes,
cause
effect
are
often
difficult
disentangle
evolutionary
trajectory
towards
structured
well-developed
their
Nevertheless,
entail
one-way
street
complexity,
retrogression
getting
increasingly
more
agents
depend
on
each
other
progressing
stages
evolution.
article
is
part
theme
issue
‘Division
labour
as
key
driver
evolution’.
Size
estimation
is
a
hard
computer
vision
problem
with
widespread
applications
in
quality
control
manufacturing
and
processing
plants,
livestock
management,
research
on
animal
behaviour.
Image-based
size
typically
facilitated
by
either
well-controlled
imaging
conditions,
the
provision
of
global
cues,
or
both.
Reference-free
remains
challenging,
because
objects
vastly
different
sizes
can
appear
identical
if
they
are
similar
shape.
Here,
we
explore
feasibility
implementing
automated
reference-free
body
to
facilitate
large-scale
experimental
work
key
model
species
sociobiology:
leaf-cutter
ants.
Leaf-cutter
ants
suitable
testbed
for
estimation,
their
workers
differ
both
shape;
principle,
it
therefore
possible
infer
mass—a
proxy
size—from
relative
proportions
alone.
Inspired
earlier
E.O.
Wilson,
who
trained
himself
discern
ant
worker
from
visual
cues
alone,
deployed
deep
learning
techniques
achieve
same
feat
automatically,
quickly,
at
scale,
images:
_Wilson
Only
Looks
Once_
(WOLO).
Using
150,000
hand-annotated
100,000
computer-generated
images,
set
convolutional
neural
networks
were
estimate
mass
image
cutouts.
The
best-performing
WOLO
achieved
errors
as
low
11
%
unseen
data,
approximately
matching
exceeding
human
performance,
measured
small
group
experts
non-experts,
but
about
1000
times
faster.
Further
refinement
may
thus
enable
accurate,
high
throughput,
non-intrusive
behavioural
work,
so
eventually
contribute
more
nuanced
comprehensive
understanding
rules
that
underpin
complex
division
labour
characterises
polymorphic
insect
societies.
Ecology and Evolution,
Journal Year:
2024,
Volume and Issue:
14(4)
Published: April 1, 2024
Abstract
Ants
are
crucial
ecosystem
engineers,
and
their
ecological
success
is
facilitated
by
a
division
of
labour
among
sterile
“workers”.
In
some
ant
lineages,
workers
have
undergone
further
morphological
differentiation,
resulting
in
differences
body
size,
shape,
or
both.
Distinguishing
between
changes
size
shape
not
trivial.
Traditional
approaches
based
on
allometry
reduce
complex
3D
shapes
into
simple
linear,
areal,
volume
metrics;
modern
using
geometric
morphometrics
typically
rely
landmarks,
introducing
observer
bias
trade‐off
effort
accuracy.
Here,
we
use
landmark‐free
method
large
deformation
diffeomorphic
metric
mapping
(LDDMM)
to
assess
the
co‐variation
mandibles
head
capsules
Atta
vollenweideri
leaf‐cutter
ants,
species
exhibiting
extreme
worker
size‐variation.
Body
mass
varied
more
than
two
orders
magnitude,
but
atlas
created
via
LDDMM
μ‐CT‐derived
mesh
files
revealed
only
distinct
capsule
shapes—one
for
minims
(body
<
1
mg)
one
all
other
workers.
We
discuss
functional
significance
identified
variation,
its
implications
evolution
polymorphism
.
Frontiers in Ecology and Evolution,
Journal Year:
2024,
Volume and Issue:
12
Published: June 21, 2024
Phenotypic
plasticity
is
a
fundamental
property
of
developing
organisms
and
thought
to
play
an
important
role
in
diversification.
Plastic
responses
themselves
are
remarkably
diverse
respond
wide
range
environmental
factors.
Here
I
focus
on
response
variation
nutrition
insects
since
1)
widespread
factor
that
impacts
most
organisms,
2)
models
study
phenotypic
plasticity.
First,
provide
brief
overview
our
current
mechanistic
understanding
the
regulation
nutritionally
cued
insects,
both
traditional
as
well
emerging
model
systems.
Then,
explore
evolutionary
mechanisms
enabling
diversification
across
taxa,
emphasizing
gene
duplication
regulatory
network
co-option.
Furthermore,
examine
nutrition-responsive
phenotypes
suites
multiple
traits
develop
coordinated
manner.
argue
how
these
integrated
at
molecular
level
can
shed
light
evolution
complex
phenotypes.
Finally,
discuss
potential
challenges
opportunities
further
plasticity,
its
regulation,
evolution.
Size
estimation
is
a
hard
computer
vision
problem
with
widespread
applications
in
quality
control
manufacturing
and
processing
plants,
livestock
management,
studies
on
animal
behaviour.
Typically,
image-based
size
facilitated
by
either
well-controlled
imaging
conditions,
the
provision
of
global
cues,
or
both.
Reference-free
challenging,
because
objects
vastly
different
sizes
can
appear
identical
if
they
are
similar
shape.
Here,
we
attempt
to
implement
automated
reference-free
body
facilitate
large-scale
experimental
work
key
model
species
sociobiology:
leaf-cutter
ants.
Leaf-cutter
ants
suitable
testbed
for
size-estimation,
their
workers
differ
both
shape;
principle,
it
therefore
possible
infer
mass,
proxy
size,
from
relative
proportions
alone.
Inspired
earlier
E.O.
Wilson,
who
trained
himself
discern
ant
worker
visual
cues
alone,
used
various
deep
learning
techniques
achieve
same
feat
automatically,
quickly,
at
scale
single
reference
image:
Wilson
Only
Looks
Once
(WOLO).
Utilizing
over
3
million
hand-annotated
computer-generated
images,
set
neural
networks---including
regressors,
classifiers,
detectors---were
estimate
mass
image
cut-outs.
The
WOLO
networks
approximately
matched
human
performance,
measured
small
group
experts
non-experts,
but
were
about
1000
times
faster.
Further
refinement
may
enable
accurate,
high-throughput,
non-intrusive
weight
scale,
so
eventually
contribute
more
nuanced
comprehensive
understanding
complex
division
labour
that
characterises
polymorphic
insect
societies.
bioRxiv (Cold Spring Harbor Laboratory),
Journal Year:
2024,
Volume and Issue:
unknown
Published: Nov. 8, 2024
ABSTRACT
Social
insects
offer
powerful
models
to
investigate
the
mechanistic
foundation
of
elaborate
individual
behaviors
comprising
a
cooperative
community.
Workers
leafcutter
ant
genus
Atta
provide
an
extreme
example
behavioral
segregation
among
many
phenotypically
distinct
worker
types.
We
utilize
complex
system
cephalotes
test
molecular
underpinnings
programming
and,
in
particular,
extent
plasticity
reprogramming.
identify
specific
neuropeptides
as
mediators
division
labor
A.
,
finding
two
associated
with
characteristic
leafcutting
and
brood
care.
Manipulation
via
genetic
knockdown
or
by
injection
these
led
stark
loss
gain
each
behavior
transcriptomic
shifts
predicted
direction,
that
is,
towards
gene
pathways
expressed
natural
caste.
also
compare
transcriptomes
those
orthologous
workers
eusocial
mammal,
naked
mole
rat
H.
gaber
revealing
global
similarities
between
caste-biased
expression
link
roles
our
studied
ants.
This
work
underscores
essential
function
establishing
social
remarkable
Journal of Umm Al-Qura University for Applied Sciences,
Journal Year:
2023,
Volume and Issue:
10(1), P. 225 - 231
Published: Aug. 14, 2023
Abstract
Leaf-cutter
ants
(
Atta
and
Acromyrmex
species)
exhibit
complex
social
organizations
that
have
fascinated
scientists
for
decades.
The
leaf-cutter
belong
to
a
subgroup
of
fungus-growing
ants,
which
live
with
fungus
inside
their
nests.
behaviours
exhibited
by
these
are
closely
linked
organization,
involves
intricate
division
labour,
caste
systems,
cooperative
tasks.
This
review
article
provides
an
overview
the
behaviors
associated
organization
ants.
It
explores
various
aspects
including
foraging
behavior,
hitchhiking
behaviour,
hygienic
environmental
influences.
leaf
cutter
instinct
forage,
as
they
walk
around
nest
cut
fragments
from
plants,
then
transport
those
jaws
go
back
cultivate
special
garden
within
colony.
By
synthesizing
existing
knowledge,
this
highlights
relationships
between
structure
shedding
light
on
underlying
mechanisms
evolutionary
implications.
Furthermore,
it
identifies
research
gaps
proposes
future
directions
studying
ant
behaviors,
emerging
techniques
interdisciplinary
approaches.
correlating
showcase
highly
organized
societies.
In
conclusion,
contributes
our
understanding
fascinating
implications
in
broader
context
insect