Abstract
Lung
cancer,
a
leading
cause
of
global
mortality,
demands
combat
for
its
effective
prevention,
early
diagnosis,
and
advanced
treatment
methods.
Traditional
diagnostic
methods
face
limitations
in
accuracy
efficiency,
necessitating
innovative
solutions.
Large
Language
Models
(LLMs)
Natural
Processing
(NLP)
offer
promising
avenues
overcoming
these
challenges
by
providing
comprehensive
insights
into
medical
data
personalizing
plans.
This
systematic
review
explores
the
transformative
potential
LLMs
NLP
automating
lung
cancer
diagnosis.
It
evaluates
their
applications,
particularly
imaging
interpretation
complex
data,
assesses
achievements
associated
challenges.
Emphasizing
critical
role
Artificial
Intelligence
(AI)
imaging,
highlights
advancements
screening
deep
learning
approaches.
Furthermore,
it
underscores
importance
on‐going
encourages
further
exploration
this
field.
Human Behavior and Emerging Technologies,
Год журнала:
2024,
Номер
2024(1)
Опубликована: Янв. 1, 2024
Incorporating
AI
tools
like
ChatGPT
into
higher
education
has
been
beneficial,
yet
the
extent
of
user
satisfaction
with
quality
information
provided
by
these
tools,
known
as
(UIS)
and
(IQ)
theory,
remains
underexplored.
This
study
introduces
a
UIS
model
specifically
designed
for
ChatGPT’s
application
in
educational
sector
based
on
multidimensions
IQ
theory.
Drawing
from
established
we
crafted
centered
around
seven
essential
factors
that
influence
effective
use
ChatGPT,
aiming
to
guide
educators
learners
overcoming
common
challenges
such
plagiarism
ensuring
ethical
AI.
Data
was
collected
Indonesian
university
participants
(
N
=
508)
analyzed
using
structural
equation
modeling
Smart‐PLS
4.0.
The
results
reveal
completeness,
precision,
timeliness,
convenience,
format
are
most
influential
driving
ChatGPT.
Interestingly,
our
research
indicated
accuracy
reliability
information,
typically
deemed
paramount,
were
not
primary
concerns
academic
Our
findings
recommend
cautious
approach
integrating
education.
We
advocate
strategic
recognizes
its
innovative
potential
while
acknowledging
limitations,
responsible
contexts.
balanced
perspective
is
crucial
fabric
without
compromising
integrity
or
quality.
Reversely
computed
dynamic
temporary
weights
introduce
a
novel
and
significant
enhancement
to
the
adaptability
accuracy
of
large
language
models.
By
dynamically
recalculating
key
hidden
layers
during
inference,
our
methodology
significantly
improves
model’s
performance
across
various
natural
processing
tasks.
Experimental
results
demonstrated
substantial
increases
in
accuracy,
response
time,
computational
efficiency
when
compared
baseline
performance.
The
integration
enabled
model
adjust
its
internal
parameters
real-time,
resulting
more
precise
context-aware
predictions.
Statistical
analysis
confirmed
significance
these
improvements,
providing
robust
validation
for
proposed
enhancements.
This
research
not
only
advances
state-of-the-art
optimization
but
also
paves
way
intelligent
adaptable
AI
systems.
Future
work
will
address
overhead
explore
broader
applicability
other
neural
network
architectures.
Knowledge-Based Systems,
Год журнала:
2024,
Номер
301, С. 112293 - 112293
Опубликована: Июль 31, 2024
Customer
service
is
an
important
and
expensive
aspect
of
business,
often
being
the
largest
department
in
most
companies.
With
growing
societal
acceptance
increasing
cost
efficiency
due
to
mass
production,
robots
are
beginning
cross
from
industrial
domain
social
domain.
Currently,
customer
tend
be
digital
emulate
interactions
through
on-screen
text,
but
state-of-the-art
research
points
towards
physical
soon
providing
person.
This
article
explores
feasibility
Transfer
Learning
different
domains
improve
chatbot
models.
In
our
proposed
approach,
transfer
learning-based
models
initially
assigned
learn
one
initial
random
weight
distribution.
Each
model
then
tasked
with
learning
another
by
transferring
knowledge
previous
domains.
To
evaluate
a
range
19
companies
such
as
e-Commerce,
telecommunications,
technology
selected
interaction
X
(formerly
Twitter)
support
accounts.
The
results
show
that
majority
improved
when
at
least
other
domain,
particularly
those
more
data-scarce
than
others.
General
language
observed,
well
higher-level
similar
knowledge.
For
each
domains,
Wilcoxon
signed-rank
test
suggests
16
have
statistically
significant
distributions
between
non-transfer
learning.
Finally,
explored
for
deployment
robot
platforms
including
"Pepper",
semi-humanoid
manufactured
SoftBank
Robotics,
"Temi",
personal
assistant
robot.
Abstract
Lung
cancer,
a
leading
cause
of
global
mortality,
demands
combat
for
its
effective
prevention,
early
diagnosis,
and
advanced
treatment
methods.
Traditional
diagnostic
methods
face
limitations
in
accuracy
efficiency,
necessitating
innovative
solutions.
Large
Language
Models
(LLMs)
Natural
Processing
(NLP)
offer
promising
avenues
overcoming
these
challenges
by
providing
comprehensive
insights
into
medical
data
personalizing
plans.
This
systematic
review
explores
the
transformative
potential
LLMs
NLP
automating
lung
cancer
diagnosis.
It
evaluates
their
applications,
particularly
imaging
interpretation
complex
data,
assesses
achievements
associated
challenges.
Emphasizing
critical
role
Artificial
Intelligence
(AI)
imaging,
highlights
advancements
screening
deep
learning
approaches.
Furthermore,
it
underscores
importance
on‐going
encourages
further
exploration
this
field.
