Rapid
prediction
of
environmental
chemistry
properties
is
critical
towards
the
green
and
sustainable
development
chemical
industry
drug
discovery.
Machine
learning
methods
can
be
applied
to
learn
relations
between
structures
their
impact.
Graph
machine
learning,
by
representations
directly
from
molecular
graphs,
may
enable
better
predictive
power
than
conventional
feature-based
models.
In
this
work,
we
leveraged
graph
neural
networks
predict
molecules.
To
systematically
evaluate
model
performance,
selected
a
representative
list
datasets,
ranging
solubility
reactivity,
compare
commonly
used
methods.
We
found
that
achieved
near
state-of-the-art
accuracy
for
all
tasks
and,
several,
improved
large
margin
over
models
rely
on
human-designed
features.
This
demonstrates
powerful
tool
do
representation
chemistry.
Further,
compared
data
efficiency
networks,
providing
guidance
selection
dependent
size
datasets
feature
requirements.
Frontiers in Sustainable Food Systems,
Год журнала:
2024,
Номер
8
Опубликована: Фев. 23, 2024
Microalgae
are
emerging
as
a
sustainable
source
of
bioproducts,
including
food,
animal
feed,
nutraceuticals,
and
biofuels.
This
review
emphasizes
the
need
to
carefully
select
suitable
species
highlights
importance
strain
optimization
enhance
feasibility
developing
algae
resource
for
food
biomaterial
production.
It
discusses
microalgal
bioprospecting
methods,
different
types
cultivation
systems,
biomass
yields,
using
wastewater.
The
paper
advances
in
artificial
intelligence
that
can
optimize
algal
productivity
overcome
limitations
faced
current
industries.
Additionally,
potential
UV
mutagenesis
combined
with
high-throughput
screening
is
examined
strategy
generating
improved
strains
without
introducing
foreign
genetic
material.
necessity
multifaceted
approach
enhanced
acknowledged.
provides
an
overview
recent
developments
crucial
commercial
success
Sustainability,
Год журнала:
2023,
Номер
15(20), С. 14884 - 14884
Опубликована: Окт. 15, 2023
The
goal
of
this
study
is
to
use
machine
learning
methodologies
identify
the
most
influential
variables
and
optimum
conditions
that
maximize
biochar,
bio-oil,
biogas
yields
for
slow
pyrolysis.
First,
experimental
results
reported
in
37
articles
were
compiled
into
a
database.
Then,
an
explainable
approach,
Shapley
Additive
exPlanations
(SHAP),
was
employed
find
effects
descriptors
on
targets,
it
found
higher
biochar
can
be
obtained
at
lower
temperatures
using
biomass
with
low
volatile
matter
high
ash
content.
Following
that,
decision
tree
classification
used
discover
leading
levels
generalizable
path
yield
where
maximum
particle
diameter
less
than
or
equal
6.5
mm
temperature
greater
912
K.
Finally,
association
rule
mining
models
created
associations
very
yields,
among
many
findings,
discovered
larger
particles
cannot
converted
bio-oil
efficiently.
It
then
concluded
methods
help
determine
best
pyrolysis
production
renewable
sustainable
biofuels.
Energies,
Год журнала:
2024,
Номер
17(20), С. 5191 - 5191
Опубликована: Окт. 18, 2024
Biofuel
has
received
worldwide
attention
as
one
of
the
most
promising
renewable
energy
sources.
Particularly,
in
many
countries
such
U.S.
and
Brazil,
first-generation
ethanol
from
corn
sugar
cane
been
used
automobile
fuel
after
blending
with
gasoline.
Nevertheless,
order
to
continuously
increase
use
biofuels,
efforts
are
needed
reduce
cost
biofuel
production
its
profitability.
This
can
be
achieved
by
increasing
efficiency
a
sequential
process
consisting
multiple
operations
feedstock
supply,
pretreatment,
fermentation,
distillation,
transportation.
study
aims
at
investigating
methodologies
for
predicting
yields,
which
is
earliest
step
stable
sustainable
production.
this
reviews
yield
estimation
approaches
using
machine
learning
technologies
that
focus
on
gradually
improving
accuracy
big
data
computer
algorithms
traditional
statistical
approaches.
Given
it
becoming
increasingly
difficult
stably
produce
feedstocks
climate
change
worsens,
research
developing
predictive
modeling
raw
material
supply
latest
ML
techniques
very
important.
As
result,
will
help
researchers
engineers
predict
yields
various
techniques,
contribute
efficient
chain
design
based
accurate
predictions
feedstocks.
The
applications
of
algae,
such
as
in
biofuel
production,
carbon
capture
and
utilization,
assorted
microalgae
production
due
to
their
high
nutrient
content,
wastewater
treatment,
bioremediation,
make
algae
cultivation
extremely
popular
industries.
Various
process
parameters
algal
characteristics
operating
conditions
the
affect
yield
productivity.
feedstock
are
directly
correlated
product.
For
example,
bio-oil
produced
is
related
ultimate
proximate
analysis;
similarly
treatment
with
dependent
upon
organic
inorganic
content
water.
Therefore,
it
essential
optimize
these
processes
enhance
productivity
identify
efficient
method
for
producing
high-quality
products
minimal
wastage.
This
can
be
accomplished
by
using
machine
learning
(ML),
one
most
recently
developed
tools
modeling
a
multiple
inputs
predict
output
accurately
without
conducting
tedious
experiments.
ML
widely
applied
predictive
growth
optimization,
recovery,
real-time
decision
support
systems,
quality
control
biomass,
energy
efficiency
many
more.
incorporation
playing
critical
role
evolution
farming
applications.
chapter
examines
different
artificial
intelligence
(AI)
ML-based
algorithms
product
enhancement,
gaining
important
insights
into
biotechnology.
