The
blood
diagnosis
of
diabetes
mellitus
(DM)
is
accurate,
but
invasive.
Attenuated
Total
Reflectance
by
Fourier
Transform
Infrared
Spectroscopy
(ATR-FTIR)
a
green
technology
adopted
in
the
detection
several
diseases
resulting
non-invasive
and
accurate
diagnosis.
analysis
ATR-FTIR
data
using
deep
learning
techniques
like
Convolutional
Neural
Network
(CNN)
promising.
However,
challenges
to
find
optimized
architectures
are
barely
explored
literature.
In
this
paper,
we
propose
an
Evolutionary
Architecture
Search
technique
able
CNN
for
salivary
spectra
type
2
DM
Genetic
Algorithm
as
optimization
approach.
Oceans,
Journal Year:
2024,
Volume and Issue:
5(3), P. 398 - 428
Published: June 21, 2024
Microplastics
are
ubiquitous
in
marine
environments
and
have
been
documented
across
all
ocean
compartments,
especially
surface
waters,
the
world.
Even
though
several
studies
identify
presence
of
microplastics
world’s
five
oceans,
there
remains
an
overt
problem
large
inconsistencies
their
sampling,
extraction,
consequent
quantification.
Despite
complexity
these
methodologies,
researchers
tried
to
explore
microplastic
abundance
waters.
Using
a
systematic
review
approach,
dataset
was
derived
from
73
primary
undertaken
since
year
2010
following
Oslo
Paris
Conventions
(OSPAR)
guidelines
monitor
harmonise
debris.
The
results
showed
differences
distribution
waters
oceans.
overall
concentration
oceans
ranged
between
0.002
62.50
items/m3,
with
mean
2.76
items/m3.
highest
found
Atlantic
(4.98
items/m3),
while
least
observed
Southern
Ocean
(0.04
items/m3).
While
challenging,
this
paper
recommends
harmonisation
separation,
identification
methods
globe
aid
design
appropriate
mitigation
strategies
for
reducing
plastic
pollution.
Foods,
Journal Year:
2025,
Volume and Issue:
14(10), P. 1670 - 1670
Published: May 9, 2025
As
one
kind
of
‘probable
human
carcinogen’
(Group
2B)
compound
classified
by
the
International
Agency
for
Research
on
Cancer,
3-MCPD
is
mainly
formed
during
thermal
processing
food.
Tedious
pretreatment
techniques
are
needed
existing
analytical
methods
to
quantify
3-MCPD.
Hence,
a
nondestructive
sensing
technique
that
offers
low
noise
interference
and
high
quantitative
precision
must
be
developed
address
this
problem.
Following
this,
Fourier
transform
infrared
spectroscopy
association
with
an
convolutional
neural
network
(CNN)
model
was
employed
in
investigation
measurement
oil
samples.
Before
building
CNN
model,
NL-SGS-D2
utilized
enhance
feature
extraction
capability
eliminating
background
noise.
Under
optimal
hyperparameter
settings,
calibration
achieved
determination
coefficient
(R2C)
0.9982
root
mean
square
error
(RMSEC)
0.0181
validation,
along
16%
performance
enhancement
enabled
stepwise
hybrid
preprocessing
strategy.
The
LODs
(0.36
μg/g)
LOQs
(1.10
proposed
method
met
requirements
detection
samples
Commission
Regulation
issued
EU.
superior
traditional
contributed
quality
monitoring
edible
industry.
Journal of the Chinese Chemical Society,
Journal Year:
2025,
Volume and Issue:
unknown
Published: May 9, 2025
Abstract
Vibrational
spectroscopy
is
a
cornerstone
in
molecular
analysis,
offering
detailed
insights
into
chemical
compositions
and
dynamics.
Recent
years
have
witnessed
paradigm
shift
with
the
integration
of
deep
learning,
which
excels
automatically
extracting
intricate
patterns
from
raw
spectral
data,
bypassing
traditional
preprocessing
steps.
This
synergy
has
significantly
enhanced
precision
speed
applications
ranging
material
science
to
biomedical
diagnostics.
review
comprehensively
explores
transformative
impact
learning
on
vibrational
modeling,
emphasizing
its
superiority
over
machine
approaches.
However,
interplay
between
still
presents
significant
challenges,
including
demand
for
massive
labeled
datasets,
risk
overfitting,
particularly
limited
samples,
inherently
black‐box
nature
models.
To
address
these
this
highlights
recent
breakthroughs
that
leverage
unique
two
fields.
For
instance,
transfer
enables
knowledge
across
domains,
mitigating
need
extensive
data.
Generative
adversarial
networks
synthetically
expand
datasets
by
capturing
complex
inherent
spectra.
Physics‐informed
neural
integrate
spectroscopic
principles
directly
model
architectures,
bridging
gap
physical
data‐driven
Additionally,
enhancing
interpretability
through
techniques
like
attention
mechanisms
saliency
mapping
critical
trustworthy
deployment,
especially
high‐stakes
where
domain‐specific
can
guide
validate
predictions.
not
only
encapsulates
advancements
but
also
distills
best
practices
development,
experimental
design
tailored
hyperparameter
tuning
robustness,
validation
protocols
ensure
reliability
cheminformatics.
provides
an
overview
latest
research
past
2
offers
future
directions
modeling
face
big
data
challenges.
