IET Signal Processing,
Journal Year:
2023,
Volume and Issue:
2023, P. 1 - 17
Published: Oct. 23, 2023
Chemical
contents,
the
important
quality
indicators
are
crucial
for
modeling
of
sintering
process.
However,
lack
these
data
can
result
in
biasedness
state
estimation
It,
thus,
greatly
reduces
accuracy
modeling.
Although
there
some
general
imputation
methods
to
tackle
lackness,
they
rarely
consider
interoutputs
correlation
and
negative
impacts
caused
by
incorrect
prefilling.
In
this
article,
a
novel
sparse
multioutput
Gaussian
convolution
process
(MGCP)
framework
is
proposed
imputation.
MGCP
flexibly
mine
relationships
between
outputs
sharing
latent
function
different
kernels.
Moreover,
penalization
terms
designed
weaken
false
relationship
outputs.
To
generalize
long-period
case,
dynamic
time
warping
term
introduced
keep
global
similarity
original
estimated
series.
Compared
with
several
existing
methods,
method
shows
great
superiority
raw
material
contents
real-world
data.
ACM Computing Surveys,
Journal Year:
2024,
Volume and Issue:
unknown
Published: Nov. 23, 2024
There
is
an
urgent
need
in
many
application
areas
for
eXplainable
ArtificiaI
Intelligence
(XAI)
approaches
to
boost
people’s
confidence
and
trust
Artificial
methods.
Current
works
concentrate
on
specific
aspects
of
XAI
avoid
a
comprehensive
perspective.
This
study
undertakes
systematic
survey
importance,
approaches,
methods,
domains
address
this
gap
provide
understanding
the
domain.
Applying
Systematic
Literature
Review
approach
has
resulted
finding
discussing
155
papers,
allowing
wide
discussion
strengths,
limitations,
challenges
methods
future
research
directions.
Advanced Engineering Informatics,
Journal Year:
2023,
Volume and Issue:
58, P. 102191 - 102191
Published: Sept. 24, 2023
Developing
a
comprehensive
data-driven
strategy
for
evaluating
the
organisational
culture
in
companies
to
foster
digital
innovation
involves
multi-criteria
decision-making
(MCDM)
problem.
This
needs
consider
various
characteristics
that
influence
success,
assign
significance
weights
each
characteristic,
and
recognise
distinct
cultures
may
excel
different
aspects
necessitates
proper
handling
of
data
variations.
Hence,
provide
organisations
seeking
align
cultural
practises
with
objectives
valuable
insights,
this
study
aims
develop
an
MCDM
model
benchmarking
innovation.
The
decision
matrix
is
formulated
based
on
intersection
evaluation
list
companies.
developed
two
phases.
Firstly,
new
weighting
model,
q-rung
picture
fuzzy-weighted
zero-inconsistency
(q-RPFWZIC),
assessing
under
fuzzy
sets
environment.
Secondly,
simple
additive
(SAW)
using
extracted
characteristics.
results
indicate
characteristic
C6
(corporate
entrepreneurship)
has
highest
weight,
value
0.161,
while
C3
(employee
participation,
agility
organizational
structures)
C7
(digital
awareness
necessity
innovations)
lowest
weight
0.088.
Company
A2
secures
top
rank
score
0.911,
satisfying
eight
characteristics,
whereas
company
A7
holds
last
order,
only
one
obtaining
0.101.
In
evaluation,
several
scenarios
were
considered
sensitivity
analysis
test
100%
increment
values
validate
reliability
results.
Micromachines,
Journal Year:
2025,
Volume and Issue:
16(5), P. 541 - 541
Published: April 30, 2025
Extracting
defect
profile
parameters
from
measured
images
poses
a
significant
challenge
in
extreme
ultraviolet
(EUV)
multilayer
metrologies,
because
these
are
crucial
for
assessing
printing
behavior
and
determining
appropriate
repair
strategies.
This
paper
proposes
to
reconstruct
reflected
field
intensity
of
phase
assisted
by
transfer
learning
with
fine-tuning.
These
generated
through
simulations
using
the
rigorous
finite-difference
time-domain
(FDTD)
method.
The
VGG-16
pre-trained
model,
known
its
robust
feature
extraction
capability,
is
adopted
fine-tuned
map
parameters.
results
demonstrate
that
proposed
approach
accurately
reconstructs
parameters,
thus
providing
important
information
mask
