Analytical Chemistry,
Journal Year:
2024,
Volume and Issue:
unknown
Published: July 15, 2024
Deep
learning
(DL)
is
becoming
more
popular
as
a
useful
tool
in
various
scientific
domains,
especially
chemistry
applications.
In
the
infrared
spectroscopy
field,
where
identifying
functional
groups
unknown
compounds
poses
significant
challenge,
there
growing
need
for
innovative
approaches
to
streamline
and
enhance
analysis
processes.
This
study
introduces
transformative
approach
leveraging
DL
methodology
based
on
transformer
attention
models.
With
data
set
containing
approximately
8677
spectra,
our
model
utilizes
self-attention
mechanisms
capture
complex
spectral
features
precisely
predict
17
groups,
outperforming
conventional
architectures
both
group
prediction
accuracy
compound-level
precision.
The
success
of
underscores
potential
transformer-based
methodologies
enhancing
techniques.
Advanced Optical Materials,
Journal Year:
2024,
Volume and Issue:
unknown
Published: July 3, 2024
Abstract
All‐day
passive
daytime
radiative
coolers
(PDRC)
offer
a
promising
solution
for
energy‐free
cooling
of
buildings
and
devices.
This
study
investigates
the
use
various
cellulose‐derivative
networks
to
achieve
optimal
stable
performance.
These
results
showed
that
mixed
cellulose
ester
network
has
maximum
solar
reflectance
97%.
While
acetate
infrared
emissivity
96%
in
atmospheric
transparency
window
band,
which
is
near‐perfect
emitter,
nitrocellulose
shows
highest
temperature,
with
significant
reduction
14
°C
from
ambient
temperature
power
124
W·m
−2
during
at
night
7.7
72.8
.
also
analyzes
dampness's
effect
on
performance
networks.
The
drops
≈
3
(from
11.3
°C)
when
relative
humidity
day
exceeds
30%
observed.
findings
indicate
capacity
material
absorb
water
surrounding
air
significantly
influences
its
as
cooler,
primarily
due
changes
optical
properties.
an
important
insight,
it
highlights
need
consider
environmental
factors
like
sample
hydrophobicity
PDRC
systems.
Analytical Chemistry,
Journal Year:
2024,
Volume and Issue:
unknown
Published: July 15, 2024
Deep
learning
(DL)
is
becoming
more
popular
as
a
useful
tool
in
various
scientific
domains,
especially
chemistry
applications.
In
the
infrared
spectroscopy
field,
where
identifying
functional
groups
unknown
compounds
poses
significant
challenge,
there
growing
need
for
innovative
approaches
to
streamline
and
enhance
analysis
processes.
This
study
introduces
transformative
approach
leveraging
DL
methodology
based
on
transformer
attention
models.
With
data
set
containing
approximately
8677
spectra,
our
model
utilizes
self-attention
mechanisms
capture
complex
spectral
features
precisely
predict
17
groups,
outperforming
conventional
architectures
both
group
prediction
accuracy
compound-level
precision.
The
success
of
underscores
potential
transformer-based
methodologies
enhancing
techniques.
