Information,
Journal Year:
2024,
Volume and Issue:
15(12), P. 814 - 814
Published: Dec. 18, 2024
In
modern
digital
infrastructure,
cyber
systems
are
foundational,
making
resilience
against
sophisticated
attacks
essential.
Traditional
cybersecurity
defenses
primarily
address
technical
vulnerabilities;
however,
the
human
element,
particularly
decision-making
during
attacks,
adds
complexities
that
current
behavioral
studies
fail
to
capture
adequately.
Existing
approaches,
including
theoretical
models,
game
theory,
and
simulators,
rely
on
retrospective
data
static
scenarios.
These
methods
often
miss
real-time,
context-specific
nature
of
user
responses
threats.
To
these
limitations,
this
work
introduces
a
framework
combines
Extended
Reality
(XR)
Generative
Artificial
Intelligence
(Gen-AI)
within
gamified
platform.
This
enables
continuous,
high-fidelity
collection
behavior
in
dynamic
attack
It
includes
three
core
modules:
Player
Behavior
Module
(PBM),
Gamification
(GM),
Simulation
(SM).
Together,
modules
create
an
immersive,
responsive
environment
for
studying
interactions.
A
case
study
simulated
critical
infrastructure
demonstrates
framework’s
effectiveness
capturing
realistic
behaviors
under
attack,
with
potential
applications
improving
response
strategies
across
sectors.
lays
foundation
adaptive
training
user-centered
development
infrastructure.
Student
simulation
supports
educators
to
improve
teaching
by
interacting
with
virtual
students.
However,
most
existing
approaches
ignore
the
modulation
effects
of
course
materials
because
two
challenges:
lack
datasets
granularly
annotated
materials,
and
limitation
models
in
processing
extremely
long
textual
data.
To
solve
challenges,
we
first
run
a
6-week
education
workshop
from
N
=
60
students
collect
fine-grained
data
using
custom
built
online
system,
which
logs
students'
learning
behaviors
as
they
interact
lecture
over
time.
Second,
propose
transferable
iterative
reflection
(TIR)
module
that
augments
both
prompting-based
finetuning-based
large
language
(LLMs)
for
simulating
behaviors.
Our
comprehensive
experiments
show
TIR
enables
LLMs
perform
more
accurate
student
than
classical
deep
models,
even
limited
demonstration
approach
better
captures
granular
dynamism
performance
inter-student
correlations
classrooms,
paving
way
towards
''digital
twin''
education.
Robotics,
Journal Year:
2025,
Volume and Issue:
14(3), P. 33 - 33
Published: March 13, 2025
This
study
presents
an
approach
for
developing
digital
avatars
replicating
individuals’
physical
characteristics
and
communicative
style,
contributing
to
research
on
virtual
interactions
in
the
metaverse.
The
proposed
method
integrates
large
language
models
(LLMs)
with
3D
avatar
creation
techniques,
using
what
we
call
Tree
of
Style
(ToS)
methodology
generate
stylistically
consistent
contextually
appropriate
responses.
Linguistic
analysis
personalized
voice
synthesis
enhance
conversational
auditory
realism.
results
suggest
that
ToS
offers
a
practical
alternative
fine-tuning
creating
accurate
responses
while
maintaining
efficiency.
outlines
potential
applications
acknowledges
need
further
work
adaptability
ethical
considerations.
