Sharq.,
Journal Year:
2023,
Volume and Issue:
15(4), P. 25 - 31
Published: Oct. 1, 2023
Facial
expression
recognition
using
deep
learning
methods
has
been
one
of
the
active
research
fields
in
last
decade.However,
most
previous
works
have
focused
on
implementation
model
laboratory
environment,
and
few
researchers
addressed
real-world
challenges
facial
systems.One
implementing
face
system
real
environment
(e.g.webcam
or
robot)
is
to
create
a
balance
between
accuracy
speed
recognition.Because,
increasing
complexity
neural
network
leads
an
increase
model,
but
due
size
decreases.Therefore,
this
paper,
we
propose
recognize
seven
main
emotions
(Happiness,
sadness,
anger,
surprise,
fear,
disgust
natural),
which
can
speed.Specifically,
proposed
three
components.First,
feature
extraction
component,
features
input
images
are
extracted
combination
normal
separable
convolutional
networks.Second,
integration
integrated
attention
mechanism.Finally,
merged
used
as
multi-layer
perceptron
expression.Our
approach
evaluated
public
datasets
received
via
webcam
Electronics,
Journal Year:
2025,
Volume and Issue:
14(9), P. 1748 - 1748
Published: April 25, 2025
As
the
e-commerce
industry
rapidly
expands,
number
of
users
and
items
continues
to
grow,
making
it
increasingly
difficult
capture
users’
purchasing
patterns.
Sequential
recommendation
models
have
emerged
address
this
issue
by
predicting
next
item
that
a
user
is
likely
purchase
based
on
their
historical
behavior.
However,
most
previous
studies
focused
primarily
modeling
sequences
using
IDs
without
leveraging
rich
item-level
information.
To
limitation,
we
propose
sequential
model
called
ITS-Rec
incorporates
various
types
textual
information,
including
titles,
descriptions,
online
reviews.
By
integrating
these
components
into
representations,
captures
both
detailed
characteristics
signals
related
motivation.
built
self-attention-based
architecture
enables
effectively
learn
long-
short-term
preferences.
Experiments
were
conducted
real-world
Amazon.com
data,
proposed
was
compared
several
state-of-the-art
models.
The
results
demonstrate
significantly
outperforms
baseline
in
terms
Hit
Ratio
(HR)
Normalized
Discounted
Cumulative
Gain
(NDCG).
Further
analysis
showed
reviews
contributed
performance
gains
among
components.
This
study
highlights
value
incorporating
features
recommendations
provides
practical
insights
enhancing
through
richer
representations.
User Modeling and User-Adapted Interaction,
Journal Year:
2024,
Volume and Issue:
unknown
Published: June 6, 2024
Abstract
User
intent
modeling
in
natural
language
processing
deciphers
user
requests
to
allow
for
personalized
responses.
The
substantial
volume
of
research
(exceeding
13,000
publications
the
last
decade)
underscores
significance
understanding
prevalent
models
AI
systems,
with
a
focus
on
conversational
recommender
systems.
We
conducted
systematic
literature
review
identify
frequently
employed
From
collected
data,
we
developed
decision
model
assist
researchers
selecting
most
suitable
their
Furthermore,
two
case
studies
assess
utility
our
proposed
guiding
modelers
developing
Our
study
analyzed
59
distinct
and
identified
74
commonly
used
features.
provided
insights
into
potential
combinations,
trends
selection,
quality
concerns,
evaluation
measures,
datasets
training
evaluating
these
models.
offers
practical
domain
modeling,
specifically
enhancing
development
introduced
provides
structured
framework,
enabling
navigate
selection
apt
methods
Research Square (Research Square),
Journal Year:
2023,
Volume and Issue:
unknown
Published: Aug. 11, 2023
Abstract
Context:
User
intent
modeling
is
a
crucial
process
in
Natural
Language
Processing
that
aims
to
identify
the
underlying
purpose
behind
user’s
request,
enabling
personalized
responses.
With
vast
array
of
approaches
introduced
literature
(over
13,000
papers
last
decade),
understanding
related
concepts
and
commonly
used
models
AI-based
systems
essential.
Method:
We
conducted
systematic
review
gather
data
on
typically
employed
designing
conversational
recommender
systems.
From
collected
data,
we
developed
decision
model
assist
researchers
selecting
most
suitable
for
their
Additionally,
performed
two
case
studies
evaluate
effectiveness
our
proposed
model.
Results:
Our
study
analyzed
59
distinct
identified
74
features.
provided
insights
into
potential
combinations,
trends
selection,
quality
concerns,
evaluation
measures,
frequently
datasets
training
evaluating
these
models.
Contribution:
contributes
practical
comprehensive
user
modeling,
empowering
development
more
effective
Conversational
Recommender
System,
can
perform
efficient
assessment
fitting
frameworks.
