Deep Learning Neural Networks in the Cloud

International Journal of Advanced Engineering, Management and
Science (IJAEMS)
Peer-Reviewed Journal
ISSN: 2454-1311 | Vol-9, Issue-10; Oct, 2023
Journal Home Page: https://ijaems.com/
DOI: https://dx.doi.org/10.22161/ijaems.910.2
This article can be downloaded from here: www.ijaems.com 9
©2023 The Author(s). Published by Infogain Publication, This work is licensed under a Creative Commons Attribution 4.0 License.
http://creativecommons.org/licenses/by/4.0/
Deep Learning Neural Networks in the Cloud
Burhan Humayun Awan
burhanhumayunawan@yahoo.ie
Received: 05 Sep 2023; Received in revised form: 08 Oct 2023; Accepted: 15 Oct 2023; Available online: 25 Oct 2023
Abstract— Deep Neural Networks (DNNs) are currently used in a wide range of critical real-world
applications as machine learning technology. Due to the high number of parameters that make up DNNs,
learning and prediction tasks require millions of floating-point operations (FLOPs). Implementing DNNs
into a cloud computing system with centralized servers and data storage sub-systems equipped with high-
speed and high-performance computing capabilities is a more effective strategy. This research presents an
updated analysis of the most recent DNNs used in cloud computing. It highlights the necessity of cloud
computing while presenting and debating numerous DNN complexity issues related to various architectures.
Additionally, it goes into their intricacies and offers a thorough analysis of several cloud computing
platforms for DNN deployment. Additionally, it examines the DNN applications that are already running on
cloud computing platforms to highlight the advantages of using cloud computing for DNNs. The study
highlights the difficulties associated with implementing DNNs in cloud computing systems and provides
suggestions for improving both current and future deployments.
Keywords— Deep Learning, Neural Networks, Cloud Computing.
I. INTRODUCTION
Numerous pattern recognition applications in real-world
fields like e-commerce, manufacturing, medicine and
health, and autonomous vehicles are now being developed
using deep neural networks (DNNs). However, due to their
extensive parameter requirements, DNNs pose significant
computing demands, especially during training. DNNs
typically have millions of parameters. As an illustration,
popular DNNs like AlexNet have 60 million parameters,
while VGG-16 has 138 million. A DNN with 175 billion
parameters that required a lengthy seven months to train
was used in a recent OpenAI project for natural language
processing (NLP) [1]. As a result, it is impossible to train a
large DNN using a single isolated computer. High-
performance computing tools are necessary for the efficient
training of DNNs.
DNN deployment on cloud platforms has increased in
popularity recently. These cloud computing platforms are
extremely fast and memory-capable high-performance
computing systems. On a variety of cloud machine learning
(ML) platforms, such as Google Colab and Amazon Web
Services (AWS) Deep Learning, training can be effectively
carried out in reasonable amounts of time. Centralized
servers powered by cloud computing provide a lot of
computing power, a lot of data storage, fast processing, low
latency, and high availability. DNNs for online applications
can be deployed thanks in large part to cloud computing.
Fig.1: Deep Neural Network with hidden layers
The implementation of cloud platforms for computationally
intensive tasks is discussed in various recent survey
publications cited in the current manuscript. These survey
papers are divided into three main categories: applications,
performance improvement technologies, and security
technologies. Yan et al.'s [2] study of security technologies
covered a range of solutions for preventing harmful
assaults, including talks of countermeasures. Recent

Awan International Journal of Advanced Engineering, Management and Science, 9(10) -2023
security methods, including encryption, access structure,
and fine-grained trace mechanisms, were introduced by
Nita and Mihailescu [3] and Sun et al. [4]. Gai et al.'s [5]
analysis of the functionality of blockchain-enabled
integrated hardware and software in cloud data centers was
centered on the integration of blockchain with clouds.
Regarding performance improvement, Pupykina et al. [7]
investigated memory management methods in cloud
computing, and Xu et al. [6] assessed technologies for
managing virtual machine performance. Offloading
technologies for job optimization across cloud and edge
systems were covered by Wang et al. [8]. In order to address
problems and look toward the future, Xu et al. [6] provide
an overview of computational distribution strategies for
managing virtual machines in the cloud. Deep
reinforcement learning-based cloud resource scheduling
solutions were discussed by Zhou et al. [9] and Feng et al.
[10].
Regarding applications, Bera et al. [12] reviewed cloud
computing applications in smart grids, while Khan et al.
[11] provided an overview of mobile cloud topologies and
the advantages of cloud computing. An overview of cloud
computing architectures for cyber-physical systems was
presented by Cao et al. [13], evaluating numerous
applications. Notably, the development of DNNs in cloud
computing systems is not a specific emphasis of these
survey studies.
This research intends to close this gap by offering a
thorough analysis of cloud computing methods for DNN
deployment, along with considerations of difficulties and
potential future research areas. The article's remaining
sections are structured to present different DNN
mechanisms, the need for cloud computing, popular cloud
platforms for deploying DNNs, specific DNN applications
implemented in cloud systems, difficulties in current DNN
deployments using cloud computing systems, and
opportunities for improving current DNN deployments on
cloud systems. Finally, a summary of the findings in brings
the article to a close.
Fig.2: Deep learning architectures in emerging cloud computing architectures
(Source: Fatsuma Jauro, 2020. Deep learning architectures in emerging cloud computing architectures. Volume 96, 106582,
ISSN 1568-4946, https://doi.org/10.1016/j.asoc.2020.106582)
Computational Complexities in Deep Learning
A single deep neural network (DNN) consists of a
significant number of parameters, demanding a large
amount of storage memory. A DNN's training and execution
phases both need a considerable time commitment. This
section explores several popular DNN architectures, such as
multilayer perceptron’s (MLPs), convolutional neural
networks (CNNs), and graph neural networks (GNNs), all

of which have intricate architectural designs and a wide
range of DNN parameters. Because DNN training requires
extensive computation, it is impractical to use a single
isolated computer for this purpose, which highlights the
vital role that cloud computing plays in aiding DNN
training.
An extensively discussed neural network is the multilayer
perceptron [24–26]. Each of MLP's levels, which include an
input layer, hidden layers, and an output layer, has a
collection of perceptron components, also known as
neurons. An MLP with two hidden layers, an input layer,
and an output layer is shown in Figure 3.
Fig.3: Multi-Layer Perceptron’s Typical Topology
Longer computational times are required to optimize an
MLP when number of layers are higher. Multiple apps have
been created as a result of using MLP in the cloud. For
instance, the study described in [28] concentrated on
developing a forecasting model that made use of multiple
input variables generated from several daily basic food price
kinds. Using the Amazon Cloud Services infrastructure, this
model sought to forecast the Surabaya consumer price
index. The multilayer perceptron technique was used in the
study to build a prediction system with a hidden layer, an
epoch, and a set number of neurons. Similar to this, a
different study [29] divided regions afflicted by cancer and
stored relevant information in the cloud using transfer
learning-based cancer segmentation (TL-CAN-Seg)
technology. A unique MLP and an altered Levenberg-
Marquardt (LM) algorithm were used to interpret complex
picture patterns and accomplish precise classification of
areas affected by breast cancer, improving the accuracy of
breast cancer diagnosis.
Recurrent neural networks (RNNs) are more effective than
MLPs at handling temporal data, including text and
sequentially correlated time series. The result from the
previous phase in the sequence serves as the input for the
next step in RNNs' unique versions of neural processing
units [30]. RNNs' hidden state, which is used for iterative
processing, catches and holds onto data all the way through
the sequence. RNNs have the ability to learn by storing,
retrieving, and using historical data to make predictions.
Model interactions at different temporal scales endow
memory. Information from all earlier steps is captured by
the aforementioned concealed state. In order to produce the
output at a particular step in the sequence, the trained RNN
can combine the input sequence and the hidden state.
Despite being created more than a decade ago, RNNs still
have issues with memory storage and computational time
limits [31].
Figure 4 depicts a typical CNN with input/output layers,
hidden layers, and a fully connected network [41,42].
Convolution, activation, and pooling layers are all included
in a hidden layer. The activation function assists in learning
nonlinear input patterns from the convolution output, the
pooling layer consolidates the outputs of the activation
function into a single value, and the convolution layer
retrieves input features within this layer. To make
categorization easier, relevant features are retrieved after
numerous convolution and pooling procedures.
Fig.4: Typical Convolutional Neural Network (CNN)
(Source:https://www.analyticsvidhya.com/blog/2022/01/co
nvolutional-neural-network-an-overview/)
A CNN typically consists of millions of network
parameters, each of which must be determined over the
course of a long period of time. In particular, ShuffleNet,
GoogLeNet, DenseNet, ResNet, AlexNet, VGGNet, and
ConvNet are a few CNN models that include millions of
network parameters and need a lot of training time. For
instance, it takes more than two weeks to train ConvNet on
a machine with four NVIDIA Titan Black GPUs.
Additionally, OpenAI's CNN, also known as neural
architecture search (NAS), requires six months to train
when using 8 P100s in parallel scaling and has a remarkable
175 billion parameters for natural language processing.
While purchasing many computers for a single CNN proves
economically inefficient, training a big CNN on a sole
computer is prohibitively time-consuming and unfeasible.
Therefore, the use of CNNs in cloud infrastructure has been
proven in a number of applications [49-56].

GNN is a result of graph representation learning [57–59],
which entails transforming and comprehending nodes and
edges in a graph into a continuous space with fewer
dimensions. GNN treats non-Euclidean domains using
elaborate data structures that reflect the relationships
between these entities [60]. This is true even if it works with
Euclidean 1-D sequences like texts and 2-D grids like
images [61]. In GNNs, the underlying data structure is
represented by a graph with the formula G = (V, E), where
V denotes the set of vertices or nodes and E denotes the
edges joining them. It is possible for the relationship (u, v)
E to be symmetric or asymmetric. Graphs can be
homogenous, like social networks like Facebook friends, or
diverse, like knowledge graphs. The incorporated elements
or graphic topologies may also alter dynamically over time.
GNNs' exponential node expansion increases
computational complexity and memory use significantly
[62]. Notably, some sizable GNNs, like GemNet-XL,
include billions of parameters [69]. Existing GNN models
are judged insufficient for large-scale graphs that
incorporate intricate topologies because they have only
been evaluated on small graphs [70]. Examples of GNN
implementation in cloud infrastructure have shown that
graph analysis is scalable and effective in a variety of
applications, such as recommender systems, traffic flow
prediction, industrial IoT, privacy preservation, and matrix
completion [71–76].
It is essential to comprehend the causes of latency in the
training and inference phases of various DNN architectures
if you want to deploy them in the cloud with the least
amount of delay possible. Using methods like dilated
convolutions in CNNs, for instance, one can increase the
network's receptive field without adding more parameters
or layers, which lowers computational cost and inference
delay [77,78]. The use of backpropagation gradients
through time, which can be computationally expensive, is
not necessary with randomization-based learning
techniques like echo state networks (ESN) [79].
Additionally, GNN and CNN pruning strategies can reduce
the amount of parameters and computations needed,
resulting in quicker inference times [80,81].
Cloud computing architectures for deep learning based
applications
Because deep neural network (DNN) architectures are
complex and demand a large number of parameters, training
and execution periods are prolonged. As a result, it is
impossible to train or deploy DNN using a single standalone
computer. Cloud computing offers a practical answer to this
problem for these kinds of resource-intensive computations.
Cloud computing meets the demanding computational
needs of several DNN implementations and training tasks
by offering substantial computing power and abundant data
storage, eventually helping customers using DNNs in
intense applications [82].
The structure and composition of cloud data centers are
outlined in the next section. The next part discusses
frequently used commercial cloud computing platforms for
the deployment of DNN and provides an overview of public
or volunteer cloud computing platforms. The section also
discusses frequently used cloud streaming systems, offering
light on how data streaming is implemented there.
Researchers studying deep learning (DL) who are looking
for reliable, affordable, and quick computing platforms for
DNN development will find this comprehensive material to
be of particular value.
- Cloud data centers
Data storage and computing are handled by cloud data
centers or remote clouds, including backhaul and core
networks [83]. A typical cloud computing architecture,
shown in Figure 5, is made up of cloud users, internet
network providers, and cloud service providers. Users
provide computational data over network service providers,
which servers then receive. This data is processed using
cloud resources, ensuring that users have enough access to
a shared pool of resources in response to their requests. This
use of cloud resources makes it possible to offer adaptable
processing power and storage, which eventually helps well-
known cloud-based companies like Amazon and Google
Cloud stay profitable [84, 85].
Fig.5: Typical Cloud Data Center Architecture
(Source: Elmirghani. 2018. GreenTouch GreenMeter Core
Network Energy-Efficiency Improvement Measures and
Optimization. Journal of Optical Communications and
Networking. 10. 10.1364/JOCN.10.00A250)
Users' requests for calculations are spread across a variety
of cloud platforms with numerous data centers [86].
Resource-intensive computations are made possible by
resource sharing within and between data centers. To
increase processing power, a distributed cloud can also be
connected to hybrid clouds, public clouds, and edge

computing installations. User requests can be distributed
among close-by data centers in an effort to reduce data
transmission delay. The architecture of a distributed cloud
consists of numerous sub-clouds. Based on the availability
of resources, a central controller within the distributed cloud
distributes computing workloads among various sub-
clouds.
- DL in the cloud
The availability of customizable computing resources, a
critical component for a variety of DL workloads
characterized by different degrees of compute needs across
distinct activities and datasets, is one of the main advantages
of cloud computing. Numerous cloud providers offer
services to meet these needs, such as the auto-scaling
features of Amazon EC2 and the scale-up and scale-out
capabilities of Microsoft Azure. These services make it
possible to run DL workloads efficiently even while
working with limited cloud resources.
Parameter server
To enable the scalability of distributed machine learning
applications within cloud data centers, parameter servers
(PSs) have been developed [87]. To train deep neural
networks, PSs have been incorporated into a number of deep
learning frameworks, including TensorFlow and MXNet.
Even with developments, total reliability cannot be assured
because cloud data centers may have server outages. To
ensure the orderly completion of learning tasks in such
contexts, preventative procedures are required to handle
job-sharing and backups [88]. The collective, which
consists of a server group and a worker group, is the center
of the PS framework. Both have job schedulers and worker
nodes, which work together in DNN training, while the
latter also includes a manager and server nodes. Consistent
hashing is used to express shared parameters as vectors of
(key, value). Operational data created by nodes is
communicated to the server, which then distributes global
information to each node. This framework provides
flexibility in guaranteeing consistent data when the
algorithm is not sensitive to data inconsistencies, increasing
the reliability of the PS framework. It allows asynchronous
tasks and dependencies by initiating the necessary methods.
Due to its widespread connectivity with a sizable and
dynamic resource pool, the PS architecture is more suitable
with heterogeneous production data centers and public
clouds than alternative techniques like AllReduce5. But the
original PS architecture has some drawbacks, such as
imbalances, elasticity constraints, and static parameter
assignments. It is not possible to incorporate more available
resources into ongoing training activities, and workload
distribution across nodes frequently does not maximize
resource capacity. Several strategies have been put out to
deal with these limitations and improve cloud computing's
capabilities. For instance, Proteus, an elastic PS framework
created to scale up training on public clouds, was introduced
by Harlap et al. [89]. The framework dynamically assigns
PSs and personnel using three transitional stages,
maximizing cost reductions, particularly when temporary
revocable resources become available. Litz [90] created
logical executors, which map physical nodes to control the
executor states of specific applications, in an additional
effort to increase elasticity. This method also includes
micro-tasks for determining dependencies and assigning
micro-tasks accordingly.
Advanced learning frameworks
Deep neural network (DNN) training in cloud environments
will be more dependable and effective thanks to proposed
improvements in learning frameworks. For workload
scheduling and dynamic resource scaling, the DL-driven
framework DL2 [91] combines supervised learning and
reinforcement learning techniques. The DNN is initially
taught offline to assimilate resource allocation patterns from
prior judgments, and reinforcement learning is then used to
improve the training of the DNN. An exploitation-
exploration method was used by Chen et al. [92] to propose
the dynamic PS load distribution scheme known as PSLD.
The plan consists of three stages: gathering data on each PS,
having workers create performance profiles, and choosing
PSs based on performance indicators and communication
time.
At the same time, Wang et al. [93] suggested the elastic
parameter server (EPS), a simple solution that enables
dynamic resource allocation and deallocation for increased
resource use and training efficiency. To improve scalability
and optimize resource utilization, this strategy incorporates
heuristic scheduling modes like incoming job scheduling
and running job scheduling. Additionally, more specialized
frameworks have been created, with a particular emphasis
on DL workloads on private clouds. In order to reduce
waiting times, Hu et al. [94 - 95] implemented an
optimization technique based on the idea of training
progress and integer programming to handle resource
scaling challenges particular to AWS and Huawei clouds.
- Data streaming for the cloud
The continuous streaming of real-time data is essential for
supporting deep learning (DL) applications in various
sensor networks and control systems, such as those found in
autonomous vehicles or smart grids. For activities like
recognition or decision-making, it is essential to acquire
measurements or data as soon as possible. Lack of access to
the most recent data versions may jeopardize system
functionality and lower safety standards. Given the huge
amounts of data volumes contained in each sample, it is

impractical to retrain deep neural networks (DNNs) using a
standalone computer, highlighting the necessity of data
streaming in a cloud setting.
Cloud data streaming platforms
Platforms for streaming cloud data have found use in
government and academic settings for the analysis of data
gathered by sensor networks. To make geographic data
processing easier, the Southern California Earthquake
Center, for instance, has set up a geophysical sensor
network [96]. This network's tens of thousands of sensors
continuously and quickly sample data. In order to better
understand climate change and create systems for predicting
earthquakes and inland flooding, geospatial data must be
collected. Furthermore, to study seismic and hydrological
features in North America, the Geodesy Advancing
Geosciences and EarthScope (GAGE) GPS network uses
information from more than a thousand GPS sensors [97].
Similar to this, the US National Science Foundation has
provided funding for the creation of a worldwide sensor
network that will largely be used to study climate change
and the cycling of carbon [98].
Data-streaming approaches
Given that many deep neural networks (DNNs) operate in
dynamic contexts, it is essential that they constantly absorb
new information or undergo retraining. A data-streaming
technique has been developed to handle this issue and
determine whether streaming data is required for changing
DNN parameters [101]. This method incorporates a
strategic trade-off between training expenses and
performance to decide if DNNs need to be updated. Its use
with TensorFlowOnSpark for three online learning
workloads has shown a decrease in total processing time.
Similar to this, Ashfahani et al. [102] have suggested a data-
streaming approach to modify network topologies in
response to fresh input. This method makes it easier for
network nodes to grow and shrink dynamically, improving
performance while reducing complexity. In comparison to
previous approaches, comparative evaluations on standard
datasets have shown improved network performance and
decreased network complexity [102].
A similar advancement is the introduction of an incremental
high-order DL model by Li et al. [103] that is designed to
adapt to high-frequency data streams. The strategy
efficiently minimizes adaptation time by translating data
into a high-order tensor space, and then uses first-order
approximations to reduce the time-consuming parameter
incrementation frequently associated with iterative
procedures. DNNs may now adapt to dynamic situations
better than traditional iterative approaches, effectively
satisfying real-time needs. A unique fuzzy neural network
has been introduced by Pratama et al. [104] that
automatically incorporates fuzzy rules from data streams,
using a simplification procedure to merge unnecessary
hidden layers and control network growth.
Results from experiments show that this approach
successfully limits network size while maintaining
performance standards. In addition, Nguyen et al. [105]
have developed a sensor network for gathering maritime
data, using a deep recurrent neural network combined with
streaming data to monitor fishing activities, spot smuggling,
forecast maritime pollution, and improve maritime traffic
safety and security in real time. This comprehensive method
effectively handles noisy and infrequently sampled data in
maritime environments by combining latent variable
modeling and data streaming to capture key elements within
maritime dynamics.
In order to provide real-time predictions, DNN designs in
cloud environments must incorporate minimal inference
latencies. Smaller DNN designs or the use of accelerators,
together with the storage of data features in low-latency
storage locations for offline precomputing predictions, can
improve serving latency. Additionally, essential for
adjusting to newly streamed data and improving DNN
performance over time is the implementation of incremental
training [106]. Using model artifacts from well-known,
publicly accessible DNNs, this entails routine updates of
DNNs based on fresh streams of data, enabling updates
without the requirement for retraining from scratch.
II. APPLICATIONS OF DNNS IN THE CLOUD
Deep neural network (DNN) deployment in the cloud has
become widely used in a wide range of applications. The
sections that follow provide an overview of a few of these
applications, including wireless capsule endoscopy, travel,
natural language processing (NLP), business intelligence
(BI), cybersecurity, anomaly detection, and mobile-cloud-
assisted implementations. Tables that give a brief synopsis
of the application's content follow these subsections. These
descriptions accurately identify the complex research issues
posed by the combination of cloud systems and DNNs.
- Natural language processing
The application of artificial intelligence (AI) to the interplay
of computers and human language is known as natural
language processing (NLP). It entails the creation of
computational models and algorithms that provide
computers the ability to effectively comprehend, decode,
and produce human language. NLP includes a wide range
of activities, including sentiment analysis, speech
recognition, language translation, natural language
comprehension, and more. Its uses span from text analysis

tools and chatbots to virtual assistants and language
translation programs [107 - 115].
- Business intelligence
The technology, systems, and procedures utilized for the
gathering, integrating, analyzing, and presenting of business
information are referred to as business intelligence (BI).
Utilizing a range of software tools, it entails gathering
information from internal and external sources and turning
it into insights that can be used to inform strategic and
tactical business choices. Data mining, data analysis,
querying, and reporting are just a few of the tasks included
in business intelligence (BI), which frequently uses metrics
and key performance indicators (KPIs) to assess an
organization's or business's success. By offering historical,
current, and predictive insights of corporate operations, BI's
primary goal is to enable better business decision-making
[117 – 120].
- Cybersecurity
The practice of defending computer systems, networks,
programs, and data from online attacks, harm, or illegal
access is known as cybersecurity. In order to protect
information technology (IT) systems and networks from
theft, damage, or disruption while maintaining the integrity,
confidentiality, and availability of data, security policies
and safeguards must be put in place. Malware, ransomware,
phishing scams, hacking, and other types of cybercrime are
just a few of the risks that cybersecurity seeks to protect
against. It includes a wide range of tactics, tools, and best
practices for preventing, spotting, and dealing with digital
threats and security breaches [121 - 135].
- Anomaly detection
Finding patterns or occurrences within a dataset that
drastically depart from expected behavior is known as
anomaly detection. Finding anomalies, outliers, or
abnormalities that deviate from the predicted patterns or
trends in the data entails using statistical analysis, machine
learning methods, and data mining approaches. The
identification of anomalous or suspicious activity is critical
for assuring the integrity, security, and effectiveness of
operations in a number of different sectors, including
cybersecurity, fraud detection, system health monitoring,
and industrial quality control [140 – 154].
- Travel
AI efficiently organizes and processes large datasets
produced by both clients and service providers in the tourist
and hospitality industries. Notably, a sizeable portion of
pertinent data used in the tourism sector comes from GPS
applications and is frequently linked to social media,
Internet of Things (IoT), and site traffic statistics. These
huge databases are managed within the framework of the
"smart" tourism industry, with the goal of providing
passengers with knowledgeable and personalized services.
Applying sophisticated intelligent approaches for analysis
is necessary when working with datasets that are so diverse,
detailed, and dispersed [157 – 176].
- Remote medical diagnosis
The use of wireless capsule endoscopy (WCE) has
significantly increased over the past two years. These
methods provide a level of internal human visibility for
diagnostic purposes that is comparable to standard
endoscopy. These technologies were initially presented in
2000, and after going through clinical studies, they were
given the go-ahead by the Food and Drug Administration in
2001. Notably, these technologies provide improved
portability and have numerous uses in the delivery of
systemic biologics and healthcare services [177 – 190].
- Mobile-cloud-assisted applications
A developing technology with a wide range of possible
applications, the idea of mobile cloud-assisted applications.
The ability to transfer tasks to cloud servers, hence
prolonging the system's operational lifespan, is the main
goal of mobile cloud computing (MCC). The computing
strain on mobile devices like smartphones, tablets, and
iPads is also lessened by MCC. To ensure the effectiveness
of the job offloading inference engine and to ease resource
restrictions on smartphones, which often have much less
processing power compared to older approaches, an
extensive evaluation based on traced data is carried out [195
– 197].
Before using Deep Neural Networks (DNNs), it is urgently
necessary to remove unnecessary and redundant frames
during the Wireless Capsule Endoscopy (WCE) operation
in order to prioritize the video content. However, there are
some significant difficulties with video prioritizing in
WCE, especially when there are limited resources and
processing capabilities. Because of this, the integration of
MCC helps to offer affordable storage, robust
computational power, and software services [198 – 199].
III. CHALLENGES AND FUTURE
DIRECTIONS
Cloud-based Deep Neural Networks (DNNs) have been
constructed for a variety of applications requiring extensive
big data analysis and high-performance computing, making
use of the significant computational power and data storage
capacities provided by cloud platforms. Despite the fact that
DNNs are effective tools for pattern recognition, they pose
a number of research obstacles, including issues with
energy consumption, the length of training and execution,
data security, and cloud compatibility [203 – 275].

- Energy efficiency
- Training cost
- Scalability
- Data security
- Privacy awareness
- Cloud interoperability
- Learning from non-stationary data: retraining
efficiency and adaptation
- Elastic implementations of deep learning models
and flexible resource allocation
- Deep reinforcement learning
- 3D vision applications
- Optimization of DNNs
IV. CONCLUSIONS
DNNs have a wide range of applications in numerous real-
world fields. Deploying DNNs on solitary stand-alone PCs
or mobile devices, however, is frequently problematic for
huge data storage and analysis applications due to their
computational complexity and the enormous number of
parameters necessary in training. As a result, the use of
DNNs in cloud computing systems has attracted a lot of
interest. In the beginning, this review article lays out the
justification for using and training DNNs in cloud-based
systems. It then digs into the intricate computations of
popular DNNs, such as MLP, CNN, RNN, and GNN,
highlighting their high parameter and FLOPs needs.
The research also provides a thorough analysis of volunteer
and public cloud computing platforms that have effectively
included and applied DNNs. Researchers and software
developers can use this information to choose the best cloud
computing platform for their DNN-focused apps. The paper
also provides information on a number of application fields,
including NLP, BI, cybersecurity, anomaly detection, and
travel, which have recently reaped significant advantages
from the integration of DNN in cloud computing. It outlines
the key difficulties involved in this approach while
highlighting the benefits and efficiency of installing DNNs
in cloud-based applications.
The research also suggests potential possibilities to improve
current deployments using cloud computing platforms and
DNNs. This thorough overview study is projected to be a
useful resource for researchers and programmers interested
in successfully implementing DNNs on cloud computing
platforms.
REFERENCES
[1] E. Strubell, A. Ganesh, and A. McCallum, Energy and policy
considerations for deep learning in nlp, in: Annual Meeting
of the Association for Computational Linguistics, 2019.
[2] Q. Yan, F.R. Yu, Q.X. Gong, J.Q. Li, Software-defined
networking (SDN) and distributed denial of service (DDoS)
attacks in cloud computing environments: a survey, some
research issues, and challenges, IEEE Communications
Surveys and Tutorials 18 (1) (2016) 602–622.
[3] S.L. Nita and M.I. Mihailescu, On artificial neural network
used in cloud computing security – a survey, in Proceedings
of the International Conference of Electronics, Computers
and Artificial Intelligence, 2018.
[4] P.J. Sun, Privacy protection and data security in cloud
computing: a survey, challenges, and solutions, IEEE
Access, vol. 7, pp. 147 420–147 452, 2019.
[5] K. Gai, J. Guo, L.H. Zhu, S. Yu, Blockchain meets cloud
computing: a survey, IEEE Transactions on Neural
Networks and Learning Systems 22 (3) (2022) 2009–2030.
[6] F. Xu, F.M. Liu, H. Jin, and A.V. Vasilakos, Managing
performance overhead of virtual machines in cloud
computing: A survey, state of the art, and future directions,
Proceedings of the IEEE, vol. 102, no. 1, pp. 11–31, 2014.
[7] A. Pupykina and G. Agosta, Survey of memory management
techniques for HPC and cloud computing, IEEE Access,
vol. 7, pp. 167 351–167 373, 2019.
[8] B. Wang, C.C. Wang, W.W. Huang, Y. Song, X.Y. Qin, A
survey and taxonomy on task offloading for edge-cloud
computing, IEEE Access 8 (2020) 186 080–186 101.
[9] G. Zhou, W. Tian, R. Buyya, Deep reinforcement learning-
based methods for resource scheduling in cloud computing:
A review and future directions, Journal of Cloud Computing
11 (2022), paper number 3.
[10] Y. Feng and F. Liu, Resource management in cloud
computing using deep reinforcement learning: A survey, in
Proceedings of the 10th Chinese Society of Aeronautics and
Astronautics Youth Forum, 2023, pp. 635–643.
[11] A.R. Khan, M. Othman, S.A. Madani, S.U. Khan, A survey
of mobile cloud computing application models, IEEE
Communications Surveys and Tutorials 16 (1) (2014) 393–
413.
[12] S. Bera, S. Misra, J.J.P.C. Rodrigues, Cloud computing
applications for smart grid: a survey, IEEE Transactions on
Parallel and Distributed Systems 26 (5) (2015) 1477–1494.
[13] K. Cao, S.Y. Hu, Y. Shi, A.W. Colombo, S. Karnouskos, X.
Li, A survey on edge and edge-cloud computing assisted
cyber-physical systems, IEEE Transactions on Industrial
Informatics 17 (11) (2021) 7806–7819.
[14] D. Soni, N. Kumar, Machine learning techniques in emerging
cloud computing integrated paradigms: A survey and
taxonomy, Journal of Network and Computer Applications
205 (2022).
[15] T. Khana, W.H. Tiana, R. Buyya, Machine learning (ml)-
centric resource management in cloud computing: A review
and future directions, Journal of Network and Computer
Applications 204 (2022).
[16] A. Saiyeda, M.A. Mir, Cloud computing for deep learning
analytics: a survey of current trends and challenges,
International Journal of Advanced Research in Computer
Science 8 (2) (2017).
[17] P.S. Priya, P. Malik, A. Mehbodniya, V. Chaudhary, A.
Sharma, and S. Ray, The relationship between cloud

computing and deep learning towards organizational
commitment, in: Proceedings of the 2nd International
Conference on Innovative Practices in Technology and
Management, 2022.
[18] F. Benedetto and A. Tedeschi, Big data sentiment analysis for
brand monitoring in social media streams by cloud
computing, in: Sentiment Analysis and Ontology
Engineering. Springer, 2016, pp. 341–377.
[19] S. Mohan, S. Mullapudi, S. Sammeta, P. Vijayvergia, and
D.C. Anastasiu, Stock price prediction using news sentiment
analysis, in: 2019 IEEE Fifth International Conference on
Big Data Computing Service and Applications
(BigDataService). IEEE, 2019, pp. 205–208.
[20] S. Prasomphan, Improvement of chatbot in trading system for
smes by using deep neural network, in: 2019 IEEE 4th
International Conference on Cloud Computing and Big Data
Analysis (ICCCBDA). IEEE, 2019, pp. 517–522.
[21] D. Scheinert, A. Acker, L. Thamsen, M.K. Geldenhuys, and
O. Kao, Learning dependencies in distributed cloud
applications to identify and localize anomalies, in:
Proceedings of IEEE/ACM International Workshop on
Cloud Intelligence, 2021, pp. 7–12.
[22] M.A. Elsayed, M. Zulkernine, Predictdeep: Security analytics
as a service for anomaly detection and prediction, IEEE
Access 8 (2020) 45 184–45 197.
[23] F. Jauro, H. Chiroma, A.Y. Gital, M. Almutairi, S.M.
Abdulhamid, J.H. Abawajy, Deep learning architectures in
emerging cloud computing architectures: Recent
development, challenges and next research trend, Applied
Soft Computing Journal 96 (2020), paper number 106582.
[24] J.-F. Chen, Q.H. Do, H.-N. Hsieh, Training artificial neural
networks by a hybrid pso-cs algorithm, Algorithms 8 (2)
(2015) 292–308.
[25] A.A. Heidari, H. Faris, S. Mirjalili, I. Aljarah, M. Mafarja,
Ant lion optimizer: theory, literature review, and application
in multi-layer perceptron neural networks, Nature-Inspired
Optimizers (2020) 23–46.
[26] Y.-S. Park, S. Lek, Artificial neural networks: multilayer
perceptron for ecological modeling, Developments in
environmental modelling, vol. 28, ch. 7, Elsevier, 2016, pp.
123–140.
[27] R. Qaddoura, M. Al-Zoubi, H. Faris, I. Almomani, et al., A
multi-layer classification approach for intrusion detection in
iot networks based on deep learning, Sensors 21 (9) (2021)
2987.
[28] S. Zahara, S. Sugianto, Prediksi indeks harga konsumen
komoditas makanan berbasis cloud computing
menggunakan multilayer perceptron, JOINTECS (Journal
of Information Technology and Computer Science) 6 (1)
(2021) 21–28.
[29] G.S.B. Jahangeer, T.D. Rajkumar, Cloud storage based
diagnosis of breast cancer using novel transfer learning with
multi-layer perceptron, International Journal of System
Assurance Engineering and Management (2022) 1–13.
[30] D. Mandic, J. Chambers, Recurrent neural networks for
prediction: learning algorithms, architectures and stability,
Wiley, 2001.
[31] Z.C. Lipton, J. Berkowitz, and C. Elkan, A critical review of
recurrent neural
networks for sequence learning, arXiv preprint arXiv:1506.00019,
2015.
[32] R. Pascanu, T. Mikolov, and Y. Bengio, On the difficulty of
training recurrent neural networks, in: Proceedings of
International Conference on Machine Learning, 2013, pp.
1310–1318.
[33] P. Yazdanian, S. Sharifian, E2lg: a multiscale ensemble of
lstm/gan deep learning architecture for multistep-ahead
cloud workload prediction, The Journal of Supercomputing
77 (2021) 11 052–11 082.
[34] Y.S. Patel, J. Bedi, MAG-D: A multivariate attention network
based approach for cloud workload forecasting, Future
Generation Computer Systems (2023).
[35] S. Ouhame, Y. Hadi, A. Ullah, An efficient forecasting
approach for resource utilization in cloud data center using
cnn-lstm model, Neural Computing and Applications 33
(2021) 10 043–10 055.
[36] N. Tran, T. Nguyen, B.M. Nguyen, G. Nguyen, A multivariate
fuzzy time series resource forecast model for clouds using
lstm and data correlation analysis, Procedia Computer
Science 126 (2018) 636–645.
[37] H.L. Leka, Z. Fengli, A.T. Kenea, A.T. Tegene, P. Atandoh,
and N.W. Hundera, A hybrid cnn-lstm model for virtual
machine workload forecasting in cloud data center, in: 2021
18th International Computer Conference on Wavelet Active
Media Technology and Information Processing
(ICCWAMTIP). IEEE, 2021, pp. 474–478.
[38] Z. Ding, J. Wang, Y. Cheng, and C. He, Alice: A lstm neural
network based short-term power load forecasting approach
in distributed cloud-edge environment, in: Journal of
Physics: Conference Series, vol. 1624, no. 5. IOP
Publishing, 2020, p. 052017.
[39] T. Hussain, K. Muhammad, W. Ding, J. Lloret, S.W. Baik,
V.H.C. de Albuquerque, A comprehensive survey of multi-
view video summarization, Pattern Recognition 109 (2021).
[40] K. Muhammad, T. Hussain, J. Del Ser, V. Palade, V.H.C. De
Albuquerque, Deepres: A deep learning-based video
summarization strategy for resource constrained industrial
surveillance scenarios, IEEE Transactions on Industrial
Informatics 16 (9) (2019) 5938–5947.
[41] K. Fukushima, S. Miyake, T. Ito, Neocognitron: A neural
network model for a mechanism of visual pattern
recognition, IEEE Transactions on Systems, Man and
Cybernetics 13 (5) (1983) 826–834.
[42] A. Krizhevsky, I. Sutskever, and G. Hinton, Imagenet
classification with deep convolutional neural network, in:
Proceedings of the Conference of Neural Information
Processing Systems, 2012, pp. 1106–1114.
[43] X. Zhang, X. Zhou, M. Lin, and J. Sun, Shufflenet: An
extremely efficient convolutional neural network for mobile
devices, in: Proceedings of The IEEE Computer Vision and
Pattern Recognition Conference, 2018.
[44] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D.
Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich,
Going deeper with convolutions, in: Proceedings of The

IEEE Computer Vision and Pattern Recognition
Conference, 2015.
[45] G. Huang, Z. Liu, L. van der Maaten, and K.Q. Weinberger,
Densely connected convolutional networks, in: Proceedings
of The IEEE Computer Vision and Pattern Recognition
Conference, 2018.
[46] K. He, X. Zhang, S. Ren, J. Sun, Deep residual learning for
image recognition, Microsoft Research, Technical Report
(2015) [Online].
Available:https://arxiv.org/pdf/1512.03385.pdf.
[47] A. Krizhevsky, I. Sutskever, and G.E. Hinton, Imagenet
classification with deep convolutional neural networks, in:
Proceedings of Conference on Neural Information
Processing Systems. IEEE, 2012.
[48] K. Simonyan and A. Zisserman, Very deep convolutional
networks for largescale image recognition, in: Proceedings
of the International Conference on Learning
Representations, 2015.
[49] A.D. Torres, H. Yan, A.H. Aboutalebi, A. Das, L. Duan, and
P. Rad, Patient facial emotion recognition and sentiment
analysis using secure cloud with hardware acceleration, in:
Computational Intelligence for Multimedia Big Data on the
Cloud with Engineering Applications. Elsevier, 2018, pp.
61–89.
[50] A. Makkar, U. Ghosh, P.K. Sharma, Artificial intelligence
and edge computing enabled web spam detection for next
generation iot applications, IEEE Sensors Journal 21 (22)
(2021) 25 352–25 361.
[51] D. Selvapandian, R. Santhosh, Deep learning approach for
intrusion detection in iot-multi cloud environment,
Automated Software Engineering 28 (2) (2021) 1–17.
[52] D. Scheinert and A. Acker, Telesto: A graph neural network
model for anomaly classification in cloud services, in:
Proceedings of International Conference on Service-
Oriented Computing. Springer, 2020, pp. 214–227.
[53] W. Ullah, A. Ullah, I.U. Haq, K. Muhammad, M. Sajjad, S.W.
Baik, CNN features with bi-directional LSTM for real-time
anomaly detection in surveillance networks, Multimedia
Tools and Applications 80 (11) (2021) 16 979–16 995.
[54] H. Saito, T. Aoki, K. Aoyama, Y. Kato, A. Tsuboi, A.
Yamada, M. Fujishiro, S. Oka, S. Ishihara, T. Matsuda, et
al., Automatic detection and classification of protruding
lesions in wireless capsule endoscopy images based on a
deep convolutional neural network, Gastrointestinal
endoscopy 92 (1) (2020) 144–151.
[55] G. Shu, W. Liu, X. Zheng, J. Li, If-cnn: Image-aware
inference framework for cnn with the collaboration of
mobile devices and cloud, IEEE Access 6 (2018) 68 621–68
633.
[56] C.-H. Chen, C.-R. Lee, W.C.-H. Lu, Smart in-car camera
system using mobile cloud computing framework for deep
learning, Vehicular Communications 10 (2017) 84–90.
[57] B. Abu-Salih, M. Al-Tawil, I. Aljarah, H. Faris, P.
Wongthongtham, K.Y. Chan, A. Beheshti, Relational
learning analysis of social politics using knowledge graph
embedding, Data Mining and Knowledge Discovery (2021)
1–40.
[58] P. Cui, X. Wang, J. Pei, W. Zhu, A survey on network
embedding, IEEE Transactions on Knowledge and Data
Engineering 31 (5) (2018) 833–852.
[59] W.L. Hamilton, R. Ying, and J. Leskovec, Representation
learning on graphs: Methods and applications, in:
Proceedings of IEEE Data Engineering Bulletin, 2017.
[60] M.M. Bronstein, J. Bruna, Y. LeCun, A. Szlam, P.
Vandergheynst, Geometric deep learning: going beyond
euclidean data, IEEE Signal Processing Magazine 34 (4)
(2017) 18–42.
[61] J. Zhou, G. Cui, S. Hu, Z. Zhang, C. Yang, Z. Liu, L. Wang,
C. Li, M. Sun, Graph neural networks: A review of methods
and applications, AI Open 1 (2020) 57–81.
[62] F. Frasca, E. Rossi, D. Eynard, B. Chamberlain, M. Bronstein,
and F. Monti, Sign: Scalable inception graph neural
networks, in: International Conference on Machine
Learning, 2020.
[63] W. Hu, M. Shuaibi, A. Das, S. Goyal, A. Sriram, J. Leskovec,
D. Parikh, and C.L. Zitnick, Forcenet: A graph neural
network for large-scale quantum calculations, in:
Proceedings of The Conference on Neural Information
Processing Systems, 2021.
[64] J. Klicpera, S. Giri, J.T. Margraf, and S. Günnemann, Fast and
uncertainty aware directional message passing for non-
equilibrium molecules, in: Proceedings of The Conference
on Neural Information Processing Systems, 2020.
[65] M. Shuaibi, A. Kolluru, A. Das, A. Grover, A. Sriram, Z.
Ulissi, and C.L. Zitnick, Rotation invariant graph neural
networks using spin convolutions, in: Proceedings of The
Conference on Neural Information Processing Systems,
2021.
[66] J. Klicpera, F. Becker, and S. Günnemann, Gemnet: Universal
directional graph neural networks for molecules, in:
Proceedings of The Conference on Neural Information
[67] Available:
https://blog.twitter.com/engineering/en_us/topics/insights/2
020/graph-ml-at-twitter.
[68] A. Sriram, A. Das, B.M. Wood, S. Goyal, and C.L. Zitnick,
Towards training billion parameter graph neural networks
for atomic simulations, in: Proceedings of The International
Conference on Learning Representations, 2022.
[69] J. Gasteiger, F. Becker, and S. Gunnemann, Gemnet:
Universal directional graph neural networks for molecules,
in: Proceedings of The Conference on Neural Information
[70] Z. Wu, S. Pan, F. Chen, G. Long, C. Zhang, S.Y. Philip, A
comprehensive survey on graph neural networks, IEEE
Transactions on Neural Networks and Learning Systems 32
(1) (2020) 4–24.
[71] R. Ying, R. He, K. Chen, P. Eksombatchai, W.L. Hamilton,
and J. Leskovec, Graph convolutional neural networks for
web-scale recommender systems, in: Proceedings of the
24th ACM SIGKDD international conference on knowledge
discovery & data mining, 2018, pp. 974–983.
[72] L. Ge, Y. Jia, Q. Li, X. Ye, Dynamic multi-graph convolution
recurrent neural network for traffic speed prediction, Journal
of Intelligent & Fuzzy Systems no. Preprint (2023) 1–14.

[73] R. Yumlembam, B. Issac, S.M. Jacob, and L. Yang, Iot-based
android malware detection using graph neural network with
adversarial defense, IEEE Internet of Things Journal, 2022.
[74] Y. Li, S. Xie, Z. Wan, H. Lv, H. Song, Z. Lv, Graph-powered
learning methods in the internet of things: A survey,
Machine Learning with Applications 11 (2023).
[75] S. Wang, Y. Zheng, X. Jia, Secgnn: Privacy-preserving graph
neural network training and inference as a cloud service,
IEEE Transactions on Services Computing (2023).
[76] R. v. d. Berg, T.N. Kipf, and M. Welling, Graph convolutional
matrix completion, arXiv preprint arXiv:1706.02263, 2017.
[77] S. Pratiher, S. Chattoraj, D. Nawn, M. Pal, R.R. Paul, H.
Konik, and J. Chatterjee, A multi-scale context aggregation
enriched mlp-mixer model for oral cancer screening from
oral sub-epithelial connective tissues, in: 2022 30th
European Signal Processing Conference (EUSIPCO). IEEE,
2022, pp. 1323–1327.
[78] Y. Li, K. Li, X. Liu, Y. Wang, L. Zhang, Lithium-ion battery
capacity estimation–a pruned convolutional neural network
approach assisted with transfer learning, Applied Energy
285 (2021).
[79] P.N. Suganthan, R. Katuwal, On the origins of randomization-
based feedforward neural networks, Applied Soft
Computing 105 (2021).
[80] T. Le-Cong, H.J. Kang, T.G. Nguyen, S.A. Haryono, D. Lo,
X.-B.D. Le, and Q.T. Huynh, Autopruner: transformer-
based call graph pruning, in: Proceedings of the 30th ACM
Joint European Software Engineering Conference and
Symposium on the Foundations of Software Engineering,
2022, pp. 520–532.
[81] J. Poyatos, D. Molina, A.D. Martinez, J. Del Ser, F. Herrera,
Evoprunedeeptl: An evolutionary pruning model for transfer
learning based deep neural networks, Neural Networks 158
(2023) 59–82.
[82] D.T. Hoang, C. Lee, D. Niyato, P. Wang, A survey of mobile
cloud computing: architecture, applications, and
approaches, Wireless Communications and Mobile
Computing 13 (2013) 1587–1611.
[83] Q. Zhang, L. Cheng, R. Boutaba, Cloud computing: State-of-
the-art and research challenges, Journal of Internet Services
and Applications 1 (1) (2010) 7–18.
[84] T. Dillon, C. Wu, and E. Chang, Cloud computing: Issues and
challenges, in: Proceedings of the IEEE International
Conference on Advanced Information Networking and
Applications, 2010.
[85] J. Novet. (2021) How amazon’s cloud business generates
billions in profit. [Online]. Available:
https://www.cnbc.com/2021/09/05/how-amazon-
webservices-makes-money-estimated-margins-by-
service.html.
[86] What is distributed cloud? 2022.
https://www.windriver.com/solutions/learning/distributed-
cloud.
[87] A. Smola and S. Narayanamurthy, An architecture for parallel
topic models, Proceedings of the VLDB Endowment, vol. 3,
no. 1–2, pp. 703–710, 2010.
[88] M. Li, D.G. Andersen, J.W. Park, A.J. Smola, A. Ahmed, V.
Josifovski, J. Long, E.J. Shekita, and B.-Y. Su, Scaling
distributed machine learning with the parameter server, in:
Proceedings of Symposium on Operating Systems Design
and Implementation, 2014, pp. 583–598.
[89] A. Harlap, A. Tumanov, A. Chung, G.R. Ganger, and P.B.
Gibbons, Proteus: agile ML elasticity through tiered
reliability in dynamic resource markets, in: Proceedings of
the Twelfth European Conference on Computer Systems,
2017, pp. 589–604.
[90] A. Qiao, A. Aghayev, W. Yu, H. Chen, Q. Ho, G.A. Gibson,
and E.P. Xing, Litz: Elastic framework for high-
performance distributed machine learning, in: Proceedings
of Annual Technical Conference, 2018, pp. 631–644.
[91] Y. Peng, Y. Bao, Y. Chen, C. Wu, C. Meng, W. Lin, Dl2: A
deep learning-driven scheduler for deep learning clusters,
IEEE Transactions on Parallel and Distributed Systems 32
(8) (2021) 1947–1960.
[92] Y. Chen, Y. Peng, Y. Bao, C. Wu, Y. Zhu, and C. Guo, Elastic
parameter server load distribution in deep learning clusters,
in: Proceedings of the 11th ACM Symposium on Cloud
Computing, 2020, pp. 507–521.
[93] S. Wang, A. Pi, X. Zhou, Elastic parameter server:
Accelerating ML training with scalable resource scheduling,
IEEE Transactions on Parallel and Distributed Systems 33
(5) (2021) 1128–1143.
[94] L. Hu, J. Zhu, Z. Zhou, R. Cheng, X. Bai, and Y. Zhang, An
optimal resource allocator of elastic training for deep
learning jobs on cloud, arXiv preprint arXiv:2109.03389,
2021.
[95] T. Menouer, KCSS: kubernetes container scheduling strategy,
The Journal of Supercomputing 77 (5) (2021) 4267–4293.
[96] N. Nakata, J.P. Chang, J.F. Lawrence, P. Boue, Body wave
extraction and tomography at long beach california with
ambient noise interferometry, Journal of Geophysical
Research: Solid Earth 120 (2) (2015) 1159–1173.
[97] D.A. Philips, C.P. Santillan, L.M. Wang, R.W. King, W.M.
Szeliga, T. Melbourne, M. Floyd, T.A. Herring, Plate
boundary observatory and related networks: GPS data
analysis methods and geodetic products, Reviews of
Geophysics 54 (2016) 759–808.
[98] T. Cowles, J. Delaney, J. Orcutt, R. Weller, The ocean
observatories initiative: sustained ocean observing across a
range of spatial scales, Marine Technology Society Journal
44 (6) (2010) 54–64.
[99] I. Foster, D.B. Gannon, Cloud Computing For Science And
Engineering, The MIT Press, 2017.
[100] T. Akidau, R. Bradshaw, C. Chambers, S. Chernyak, R.J.
Fernandez-Moctezuma, R. Lax, S. McVeety, D. Mills, F.
Perry, E. Schmidt, and S. Whittle, The dataflow model: a
practical approach to balancing correctness, latency, and
cost in massive scale, unbounded, out-of-order data
processing, Proceedings of the VLDB ENdowment, vol. 8,
no. 12, pp. 1792–1803, 2015.
[101] W. Rang, D.L. Yang, D.Z. Cheng, Y. Wang, Data life aware
model updating strategy for stream-based online deep
learning, IEEE Transactions on Parallel and Distributed
Systems 32 (10) (2021) 2571–2581.
[102] A. Ashfahani, Autonomous deep learning: Incremental
learning of deep neural networks for evolving data streams,

in: Proceedings of the International Conference on Data
Mining Workshops, 2019, pp. 83–90.
[103] Y.L. Li, M. Zhang, W. Wang, Online real-time analysis of
data streams based on an incremental high-order deep
learning model, IEEE Access 6 (2018) 77 615–77 623.
[104] M. Pratama, W. Pedrycz, G.I. Webb, Online real-time
analysis of data streams based on an incremental high-order
deep learning model, IEEE Transactions on Fuzzy Systems
28 (7) (2020) 1315–1328.
[105] D. Nguyen, R. Vadaine, G. Hajduch, R. Garello, and R.
Fablet, A multi-task deep learning architecture for maritime
surveillance using ais data streams, in: Proceedings of IEEE
5th International Conference on Data Science and Advanced
Analytics, 2018, pp. 331–340.
[106] S.S. Zhang, J.W. Liu, and X. Zuo, Adaptive online
incremental learning for evolving data streams, 2022.
[Online]. Available: https://arxiv.org/abs/1805.04754.
[107] D.W. Otter, J.R. Medina, J.K. Kalita, A survey of the usages
of deep learning for natural language processing, IEEE
Transactions on Neural Networks and Learning Systems 32
(2) (2021) 604–624.
[108] W. Medhat, A. Hassan, H. Korashy, Sentiment analysis
algorithms and applications: A survey, Ain Shams
Engineering Journal 5 (4) (2014) 1093–1113.
[109] R. Feldman, Techniques and applications for sentiment
analysis, Communications of the ACM 56 (4) (2013) 82–89.
[110] R. Obiedat, L. Al-Qaisi, R. Qaddoura, O. Harfoushi, A. Al-
Zoubi, An intelligent hybrid sentiment analyzer for personal
protective medical equipments based on word embedding
technique: The COVID-19 era, Symmetry 13 (12) (2021)
2287.
[111] D.A. Al-Qudah, A.-Z. Ala’M, P.A. Castillo-Valdivieso, H.
Faris, Sentiment analysis for e-payment service providers
using evolutionary extreme gradient boosting, IEEE Access
8 (2020) 189 930–189 944.
[112] R. Obiedat, O. Harfoushi, R. Qaddoura, L. Al-Qaisi, A. Al-
Zoubi, An evolutionary-based sentiment analysis approach
for enhancing government decisions during COVID-19
pandemic: The case of jordan, Applied Sciences 11 (19)
(2021) 9080.
[113] R.O. Sinnott and S. Cui, Benchmarking sentiment analysis
approaches on the cloud, in: Proceedings of 2016 IEEE 22nd
International Conference on Parallel and Distributed
Systems, 2016, pp. 695–704.
[114] M. Ghorbani, M. Bahaghighat, Q. Xin, F. Özen,
Convlstmconv network: a deep learning approach for
sentiment analysis in cloud computing, Journal of Cloud
Computing 9 (1) (2020) 1–12.
[115] M.R. Raza, W. Hussain, and J.M. Merigó, Long short-term
memory-based sentiment classification of cloud dataset, in:
Proceedings of IEEE Innovations in Intelligent Systems and
Applications Conference, 2021, pp. 1–6.
[116] G. Preethi, P.V. Krishna, M.S. Obaidat, V. Saritha, and S.
Yenduri, Application of deep learning to sentiment analysis
for recommender system on cloud, in: Proceedings of
International conference on computer, information and
telecommunication systems. IEEE, 2017, pp. 93–97.
[117] M.A. Khan, S. Saqib, T. Alyas, A.U. Rehman, Y. Saeed, A.
Zeb, M. Zareei, E.M. Mohamed, Effective demand
forecasting model using business intelligence empowered
with machine learning, IEEE Access 8 (2020) 116 013–116
023.
[118] B.M. Balachandran, S. Prasad, Challenges and benefits of
deploying big data analytics in the cloud for business
intelligence, Procedia Computer Science 112 (2017) 1112–
1122.
[119] C. Moreno, R.A.C. González, E.H. Viedma, Data and
artificial intelligence strategy: A conceptual enterprise big
data cloud architecture to enable market-oriented
organisations, International Journal of Interactive
Multimedia and Artificial Intelligence 5 (6) (2019) 7–14.
[120] C. Juarez and H. Afli, Online news analysis on cloud
computing platform for market prediction. in Collaborative
European Research Conference, 2020, pp. 125–140.
[121] P. Dixit and S. Silakari, ‘‘Deep learning algorithms for
cybersecurity applications: A technological and status
review,” Computer Science Review, vol. 39, paper number
100317, 2021
[122] M.M. Alani, Big data in cybersecurity: a survey of
applications and future trends, Journal of Reliable
Intelligent Environments (2021) 1–30.
[123] P. Podder, S. Bharati, M. Mondal, P.K. Paul, and U. Kose,
Artificial neural network for cybersecurity: a
comprehensive review, arXiv preprint arXiv:2107.01185,
2021.
[124] L. Gupta, T. Salman, A. Ghubaish, D. Unal, A.K. Al-Ali, R.
Jain, Cybersecurity of multi-cloud healthcare systems: A
hierarchical deep learning approach, Applied Soft
Computing 118 (2022), paper number 108439.
[125] Y. Chai, Y. Zhou, W. Li, Y. Jiang, An explainable multi-
modal hierarchical attention model for developing phishing
threat intelligence, IEEE Transactions on Dependable and
Secure Computing 19 (2) (2022) 790–803.
[126] B. Abu-Salih, D.A. Qudah, M. Al-Hassan, S.M. Ghafari, T.
Issa, I. Aljarah, A. Beheshti, and S. Alqahtan, An intelligent
system for multi-topic social spam detection in
microblogging, arXiv preprint arXiv:2201.05203, 2022.
[127] F.J. Abdullayeva, Advanced persistent threat attack
detection method in cloud computing based on autoencoder
and softmax regression algorithm, Array 10 (2021), paper
number 100067.
[128] A. Vadariya and N.K. Jadav, A survey on phishing URL
detection using artificial intelligence, in: Proceedings of
International Conference on Recent Trends in Machine
Learning, IoT, Smart Cities and Applications, 2021, pp. 9–
20.
[129] L. Tang, Q.H. Mahmoud, A survey of machine learning-
based solutions for phishing website detection, Machine
Learning and Knowledge Extraction 3 (3) (2021) 672–694.
[130] G. Li, P. Sharma, L. Pan, S. Rajasegarar, C. Karmakar, N.
Patterson, Deep learning algorithms for cyber security
applications: A survey, Journal of Computer Security 29 (5)
(2021) 447–471.
[131] K. Sethi, R. Kumar, N. Prajapati, and P. Bera, Deep
reinforcement learning based intrusion detection system for

cloud infrastructure, in: Proceedings of the IEEE
International Conference on Communication Systems and
Networks, 2020, pp. 1–6.
[132] M.N. Hossain, J. Wang, O. Weisse, R. Sekar, D. Genkin, B.
He, S.D. Stoller, G. Fang, F. Piessens, and E. Downing,
Dependence-preserving data compaction for scalable
forensic analysis, in: Proceedings of the 27th Security
Symposium, 2018, pp. 1723–1740.
[133] M. Du, F. Li, G. Zheng, and V. Srikumar, Deeplog: Anomaly
detection and diagnosis from system logs through deep
learning, in: Proceedings of the 2017 ACM SIGSAC
Conference on Computer and Communications Security,
2017, pp. 1285–1298.
[134] T. Thilagam, R. Aruna, Intrusion detection for network
based cloud computing by custom rc-nn and optimization,
ICT Express 7 (4) (2021) 512–520.
[135] Y. Shen, E. Mariconti, P.A. Vervier, and G. Stringhini,
Tiresias: Predicting security events through deep learning,
in: Proceedings of the 2018 ACM SIGSAC Conference on
Computer and Communications Security, 2018, pp.
592–605.
[136] M. Ramilli, Malware training sets: a machine learning
dataset for everyone, Marco Ramilli Web Corner, 2016.
[137] N. Koroniotis, N. Moustafa, E. Sitnikova, B. Turnbull,
Towards the development of realistic botnet dataset in the
internet of things for network forensic analytics: Bot-iot
dataset, Future Generation Computer Systems 100 (2019)
779–796.
[138] Z. Yang, D. Yang, C. Dyer, X. He, A. Smola, and E. Hovy,
Hierarchical attention networks for document classification,
in: Proceedings of the 2016 conference of the North
American chapter of the association for computational
linguistics: human language technologies, 2016, pp. 1480–
1489.
[139] N. Moustafa and J. Slay, Unsw-nb15: a comprehensive data
set for network intrusion detection systems (unsw-nb15
network data set), in: 2015 military communications and
information systems conference (MilCIS). IEEE, 2015, pp.
1–6.
[140] L. Bergman and Y. Hoshen, Classification based anomaly
detection for general data, in: Proceedings of International
Conference on Learning Representations, 2020.
[141] Y. Wu, H.-N. Dai, and H. Tang, Graph neural networks for
anomaly detection in industrial internet of things, IEEE
Internet of Things Journal, 2021.
[142] H. Wang, Z. Wu, H. Jiang, Y. Huang, J. Wang, S. Kopru,
and T. Xie, Groot: An event-graph-based approach for root
cause analysis in industrial settings, in: Proceedings of
IEEE/ACM International Conference on Automated
Software Engineering, 2021.
[143] A. Protogerou, S. Papadopoulos, A. Drosou, D. Tzovaras, I.
Refanidis, A graph neural network method for distributed
anomaly detection in iot, Evolving Systems 12 (1) (2021)
19–36.
[144] F. Fusco, B. Eck, R. Gormally, M. Purcell, and S. Tirupathi,
Knowledge-and data-driven services for energy systems
using graph neural networks, in: Proceedings of IEEE
International Conference on Big Data. IEEE, 2020, pp.
1301–1308.
[145] M. Lee, S. Hosseinalipour, C.G. Brinton, G. Yu, and H. Dai,
A fast graph neural network-based method for winner
determination in multi-unit combinatorial auctions, IEEE
Transactions on Cloud Computing, 2020.
[146] M. Gao, Y. Li, and J. Yu, Workload prediction of cloud
workflow based on graph neural network, in: Proceedings of
International Conference on Web Information Systems and
Applications, 2021, pp. 169–189.
[147] A. Rafiq, T.A. Khan, M. Afaq, and W.-C. Song, Service
function chaining and traffic steering in sdn using graph
neural network, in: Proceedings of IEEE International
Conference on Information and Communication
Technology Convergence, 2020, pp. 500–505.
[148] Z. Yu, W. Liu, X. Liu, and G. Wang, Drag-jdec: A deep
reinforcement learning and graph neural network-based job
dispatching model in edge computing, in: Proceedings of
IEEE/ACM 29th International Symposium on Quality of
Service. IEEE, 2021, pp. 1–10.
[149] T. Liang, X. Sheng, L. Zhou, Y. Li, H. Gao, Y. Yin, L. Chen,
Mobile app recommendation via heterogeneous graph
neural network in edge computing, Applied Soft Computing
vol. 103, paper number 107162 (2021).
[150] A. Said, S.-U. Hassan, S. Tuarob, R. Nawaz, M. Shabbir,
DGSD: Distributed graph representation via graph statistical
properties, Future Generation Computer Systems 119
(2021) 166–175.
[151] S. Chen, J. Dong, P. Ha, Y. Li, S. Labi, Graph neural
network and reinforcement learning for multi-agent
cooperative control of connected autonomous vehicles,
Computer-Aided Civil and Infrastructure Engineering 36 (7)
(2021) 838–857.
[152] Y. Gao, X. Gu, H. Zhang, H. Lin, and M. Yang, ‘‘Runtime
performance prediction for deep learning models with graph
neural network,” Technical Report MSR-TR-2021-3.
Microsoft, Tech. Rep., 2021.
[153] W. Ullah, A. Ullah, T. Hussain, Z.A. Khan, S.W. Baik, An
efficient anomaly recognition framework using an attention
residual lstm in surveillance videos, Sensors 21 (8) (2021)
2811.
[154] W. Ullah, A. Ullah, T. Hussain, K. Muhammad, A.A.
Heidari, J. Del Ser, S.W. Baik, V.H.C. De Albuquerque,
Artificial intelligence of things-assisted two stream neural
network for anomaly detection in surveillance big video
data, Future Generation Computer Systems 129 (2022) 286–
297.
[155] W. Xu, L. Huang, A. Fox, D. Patterson, and M. Jordan,
‘‘Online system problem detection by mining patterns of
console logs,” in: 2009 ninth IEEE international conference
on data mining. IEEE, 2009, pp. 588–597.
[156] D. Duplyakin, R. Ricci, A. Maricq, G. Wong, J. Duerig, E.
Eide, L. Stoller, M. Hibler, D. Johnson, K. Webb et al., The
design and operation of fCloudLabg”, in: 2019 USENIX
annual technical conference (USENIX ATC 19), 2019, pp.
1–14.
[157] X. Zhou, C. Xu, B. Kimmons, Detecting tourism
destinations using scalable geospatial analysis based on

cloud computing platform, Computers, Environment and
Urban Systems 54 (2015) 144–153.
[158] K. Vassakis, E. Petrakis, I. Kopanakis, J. Makridis, and G.
Mastorakis, Locationbased social network data for tourism
destinations, in: Proceedings of Big data and innovation in
tourism, travel, and hospitality, 2019, pp. 105–114.
[159] W. Wang, N. Kumar, J. Chen, Z. Gong, X. Kong, W. Wei,
H. Gao, Realizing the potential of the internet of things for
smart tourism with 5g and ai, IEEE Network 34 (6) (2020)
295–301.
[160] S.Y. Park, B. Pan, Identifying the next non-stop flying
market with a big data approach, Tourism Management 66
(2018) 411–421.
[161] A. Kontogianni, E. Alepis, C. Patsakis, Promoting smart
tourism personalized services via a combination of deep
learning techniques, Expert Systems with Applications vol.
187, paper number 115964 (2022).
[162] M. Mariani, R. Baggio, Big data and analytics in hospitality
and tourism: a systematic literature review, International
Journal of Contemporary Hospitality Management no. 1,
paper number 232 (2021).
[163] J.C. Cepeda-Pacheco, M.C. Domingo, Deep learning and
internet of things for tourist attraction recommendations in
smart cities, Neural Computing and Applications (2022) 1–
19.
[164] J. Guerra-Montenegro, J. Sanchez-Medina, I. La na, D.
Sanchez-Rodriguez, I. Alonso-Gonzalez, and J. Del Ser,
‘‘Computational intelligence in the hospitality industry: A
systematic literature review and a prospect of challenges,”
Applied Soft Computing, vol. 102, p. 107082, 2021.
[165] F. Piccialli, F. Giampaolo, G. Casolla, V.S. Di Cola, K. Li,
A deep learning approach for path prediction in a location-
based iot system, Pervasive and Mobile Computing vol. 66,
paper number 101210 (2020).
[166] Y.-C. Chang, C.-H. Ku, C.-H. Chen, Using deep learning
and visual analytics to explore hotel reviews and responses,
Tourism Management 80 (2020).
[167] G. Díaz, H. Macià, V. Valero, J. Boubeta-Puig, F. Cuartero,
An intelligent transportation system to control air pollution
and road traffic in cities integrating cep and colored petri
nets, Neural Computing and Applications 32 (2) (2020)
405–426.
[168] P. Arthurs, L. Gillam, P. Krause, N. Wang, K. Halder, A.
Mouzakitis, A taxonomy and survey of edge cloud
computing for intelligent transportation systems and
connected vehicles, IEEE Transactions on Intelligent
Transportation Systems (2021).
[169] Q. Cai, M. Abdel-Aty, Y. Sun, J. Lee, J. Yuan, Applying a
deep learning approach for transportation safety planning by
using high-resolution transportation and land use data,
Transportation Research Part A: Policy and Practice 127
(2019) 71–85.
[170] J. Wang, R. Chen, Z. He, Traffic speed prediction for urban
transportation network: A path based deep learning
approach, Transportation Research Part C: Emerging
Technologies 100 (2019) 372–385.
[171] X. Wang, J. Liu, T. Qiu, C. Mu, C. Chen, P. Zhou, A real-
time collision prediction mechanism with deep learning for
intelligent transportation system, IEEE Transactions on
Vehicular Technology 69 (9) (2020) 9497–9508.
[172] S. Yang, W. Ma, X. Pi, S. Qian, A deep learning approach
to real-time parking occupancy prediction in transportation
networks incorporating multiple spatio-temporal data
sources, Transportation Research Part C: Emerging
Technologies 107 (2019) 248–265.
[173] A. Paranjothi, M.S. Khan, S. Zeadally, A survey on
congestion detection and control in connected vehicles, Ad
Hoc Networks 108 (2020).
[174] C. Chen, B. Liu, S. Wan, P. Qiao, Q. Pei, An edge traffic
flow detection scheme based on deep learning in an
intelligent transportation system, IEEE Transactions on
Intelligent Transportation Systems 22 (3) (2020) 1840–
1852.
[175] M.M. Iqbal, M.T. Mehmood, S. Jabbar, S. Khalid, A.
Ahmad, G. Jeon, An enhanced framework for multimedia
data: Green transmission and portrayal for smart traffic
system, Computers & Electrical Engineering 67 (2018)
291–308.
[176] R.S. Thejaswini and S. Rajaraajeswari, A real-time traffic
congestion avoidance framework for smarter cities, in:
Proceedings of AIP Conference Proceedings, vol. 2039, no.
1, paper number 020009, 2018.
[177] S. Sarker, B. Wankum, T. Perey, M.M. Mau, J. Shimizu, R.
Jones, B.S. Terry, A novel capsule-delivered enteric drug-
injection device for delivery of systemic biologics: A pilot
study in a porcine model, IEEE Transactions on Biomedical
Engineering 69 (6) (2022) 1870–1879.
[178] S. Sangodoyin, E.M. Ugurlu, M. Dey, M. Prvulovic, A.
Zajic´, Leveraging on-chip transistor switching for
communication and sensing in neural implants and
gastrointestinal devices, IEEE Transactions on Biomedical
Engineering 69 (1) (2021) 377–389.
[179] S. Li, P. Si, Z. Zhang, J. Zhu, X. He, N. Zhang, Dfca-net:
Dual feature context aggregation network for bleeding areas
segmentation in wireless capsule endoscopy images, Journal
of Medical and Biological Engineering 42 (2) (2022) 179–
188.
[180] M.R. Basar, F. Malek, K.M. Juni, M.S. Idris, M.I.M. Saleh,
Ingestible wireless capsule technology: A review of
development and future indication, International Journal of
Antennas and Propagation vol. 2012, paper number 807165
(2012).
[181] I. Mehmood, M. Sajjad, S.W. Baik, Mobile-cloud assisted
video summarization framework for efficient management
of remote sensing data generated by wireless capsule
sensors, Sensors 14 (9) (2014) 17 112–17 145.
[182] R. Hamza, K. Muhammad, Z. Lv, F. Titouna, Secure video
summarization framework for personalized wireless capsule
endoscopy, Pervasive and Mobile Computing 41 (2017)
436–450.
[183] K. Muhammad, S. Khan, N. Kumar, J. Del Ser, S. Mirjalili,
Vision-based personalized wireless capsule endoscopy for
smart healthcare: Taxonomy, literature review,
opportunities and challenges, Future Generation Computer
Systems 113 (2020) 266–280.

[184] K. Muhammad, M. Sajjad, M.Y. Lee, S.W. Baik, Efficient
visual attention driven framework for key frames extraction
from hysteroscopy videos, Biomedical Signal Processing
and Control 33 (2017) 161–168.
[185] K. Muhammad, J. Ahmad, M. Sajjad, S.W. Baik, Visual
saliency models for summarization of diagnostic
hysteroscopy videos in healthcare systems, SpringerPlus
vol. 5, paper number 1495 (2016).
[186] R. Shrestha, S.K. Mohammed, M.M. Hasan, X. Zhang, K.A.
Wahid, Automated adaptive brightness in wireless capsule
endoscopy using image segmentation and sigmoid function,
IEEE Transactions on Biomedical Circuits and Systems 10
(4) (2016) 884–892.
[187] J.-Y. He, X. Wu, Y.-G. Jiang, Q. Peng, R. Jain, Hookworm
detection in wireless capsule endoscopy images with deep
learning, IEEE Transactions on Image Processing 27 (5)
(2018) 2379–2392.
[188] X. Wu, H. Chen, T. Gan, J. Chen, C.-W. Ngo, Q. Peng,
Automatic hookworm detection in wireless capsule
endoscopy images, IEEE Transactions on Medical Imaging
35 (7) (2016) 1741–1752.
[189] Y. Yuan, J. Wang, B. Li, M.Q.-H. Meng, Saliency based
ulcer detection for wireless capsule endoscopy diagnosis,
IEEE Transactions on Medical Imaging 34 (10) (2015)
2046–2057.
[190] I. Mehmood, M. Sajjad, S.W. Baik, Video summarization
based tele endoscopy: a service to efficiently manage visual
data generated during wireless capsule endoscopy
procedure, Journal of medical systems 38 (9) (2014) 1–9.
[191] Y. Yuan, M.Q.-H. Meng, Deep learning for polyp
recognition in wireless capsule endoscopy images, Medical
physics 44 (4) (2017) 1379–1389.
[192] X. Jia, M.Q.-H. Meng, A deep convolutional neural network
for bleeding detection in wireless capsule endoscopy
images, 38th annual international conference of the IEEE
engineering in medicine and biology society (EMBC), IEEE
2016 (2016) 639–642.
[193] S. Jain, A. Seal, A. Ojha, A. Yazidi, J. Bures, I. Tacheci, O.
Krejcar, A deep cnn model for anomaly detection and
localization in wireless capsule endoscopy images,
Computers in Biology and Medicine 137 (2021).
[194] T. Aoki, A. Yamada, K. Aoyama, H. Saito, A. Tsuboi, A.
Nakada, R. Niikura, M. Fujishiro, S. Oka, S. Ishihara, et al.,
Automatic detection of erosions and ulcerations in wireless
capsule endoscopy images based on a deep convolutional
neural network, Gastrointestinal endoscopy 89 (2) (2019)
357–363.
[195] X. Gu, K. Nahrstedt, A. Messer, I. Greenberg, D. Milojicic,
Adaptive offloading for pervasive computing, IEEE
Pervasive Computing 3 (3) (2004) 66–73.
[196] K. Yang, S. Ou, H.-H. Chen, On effective offloading
services for resource constrained mobile devices running
heavier mobile internet applications, IEEE Communications
Magazine 46 (1) (2008) 56–63.
[197] A.P. Miettinen and J.K. Nurminen, Energy efficiency of
mobile clients in cloud computing, in: Proceedings of 2nd
USENIX Workshop on Hot Topics in Cloud Computing,
2010.
[198] J.-C. Hsieh, A.-H. Li, C.-C. Yang, Mobile, cloud, and big
data computing: contributions, challenges, and new
directions in tele cardiology, International Journal of
Environmental Research and Public Health 10 (11) (2013)
6131–6153.
[199] G. Fortino, D. Parisi, V. Pirrone, G. Di Fatta, Bodycloud: A
SaaS approach for community body sensor networks, Future
Generation Computer Systems 35 (2014) 62–79.
[200] C. Xia, J. Zhao, H. Cui, X. Feng, J. Xue, Dnntune:
Automatic benchmarking dnn models for mobile-cloud
computing, ACM Transactions on Architecture and Code
Optimization (TACO) 16 (4) (2019) 1–26.
[201] B. Kumar, G. Pandey, B. Lohani, S.C. Misra, A multi-
faceted cnn architecture for automatic classification of
mobile lidar data and an algorithm to reproduce point cloud
samples for enhanced training, ISPRS journal of
photogrammetry and remote sensing 147 (2019) 80–89.
[202] A.E. Eshratifar, M.S. Abrishami, M. Pedram, JointDNN: An
efficient training and inference engine for intelligent mobile
cloud computing services, IEEE Transactions on Mobile
Computing 20 (2) (2019) 565–576.
[203] A. Katal, S. Dahiya, T. Choudhury, A survey on cloud
computing in energy management of the smart grids, Cluster
Computing (2021).
[204] W. Pedrycz, Towards green machine learning: challenges,
opportunities, and developments, Journal of Smart
Environments and Green Computing 2 (4) (2022) 163–174.
[205] R. Schwartz, J. Dodge, N.A. Smith, O. Etzioni, Green ai,
Communications of the ACM 63 (12) (2020) 54–63.
[206] H. Jin, Q. Song, and X. Hu, Auto-keras: An efficient neural
architecture search system, in: ACM SIGKDD Conference
on Knowledge Discovery and Data Mining, 2019, pp. 1946–
1956.
[207] A. Li, O. Spyra, S. Perel, V. Dalibard, M. Jaderberg, C.J.
Gu, D. Budden, T. Harley, and P. Gupta, A generalized
framework for population based training, in: Proceedings of
ACM SIGKDD Conference on Knowledge Discovery and
Data Mining, 2019.
[208] T. Elsken, J.H. Metzen, F. Hutter, Neural architecture
search: A survey, The Journal of Machine Learning
Research 20 (1) (2019) 1997–2017.
[209] P. Ren, Y. Xiao, X. Chang, P.-Y. Huang, Z. Li, X. Chen, X.
Wang, A comprehensive survey of neural architecture
search: Challenges and solutions, ACM Computing Surveys
(CSUR) 54 (4) (2021) 1–34.
[210] H. Li, P. Wu, Z. Wang, J. f. Mao, F.E. Alsaadi, and Z.N. Y,
‘‘A generalized framework of feature learning enhanced
convolutional neural network for pathology-image-oriented
cancer,” Computers in Biology and Medicine, vol. 151A,
2022, 106265.
[211] H. Li, N.Y. Zeng, P.S. Wu, K. Clawson, Cov-net: A
computer-aided diagnosis method for recognizing covid-19
from chest x-ray images via machine vision, Expert Systems
with Applications 207 (2022).
[212] P. Wu, Z. Wang, B. Zheng, H. Li, F.E. Alsaadi, N. Zeng,
Aggn: Attention-based glioma grading network with multi-
scale feature extraction and multi-modal information fusion,
Computers in Biology and Medicine 152 (2023).

[213] I. Makdoom, M. Abolhasan, H. Abbas, W. No, Blockchains
adoption in iot: the challenges, and a way forward, Journal
of Network and Computer Applications 125 (2019) 251–
279.
[214] H. Hu, Z. Salcic, L. Sun, G. Dobbie, P.S. Yu, X. Zhang,
Membership inference attacks on machine learning: A
survey, ACM Computing Surveys (CSUR) 54 (11s) (2022)
1–37.
[215] M. Jegorova, C. Kaul, C. Mayor, A.Q. O’Neil, A. Weir, R.
Murray-Smith, S.A. Tsaftaris, Survey: Leakage and privacy
at inference time, IEEE Transactions on Pattern Analysis
and Machine Intelligence (2022).
[216] M. Abadi, A. Chu, I. Goodfellow, H.B. McMahan, I.
Mironov, K. Talwar, and L. Zhang, Deep learning with
differential privacy, in: Proceedings of the 2016 ACM
SIGSAC conference on computer and communications
security, 2016, pp. 308–318.
[217] C. Dwork, ‘‘Differential privacy: A survey of results,” in
Theory and Applications of Models of Computation: 5th
International Conference, TAMC 2008, Xi’an, China, April
25-29, 2008. Proceedings 5. Springer, 2008, pp. 1–19.
[218] S. Meftah, B.H.M. Tan, C.F. Mun, K.M.M. Aung, B.
Veeravalli, V. Chandrasekhar, Doren: toward efficient deep
convolutional neural networks with fully homomorphic
encryption, IEEE Transactions on Information Forensics
and Security 16 (2021) 3740–3752.
[219] Z. Zhong, W. Bao, J. Wang, X. Zhu, X. Zhang, Flee: A
hierarchical federated learning framework for distributed
deep neural network over cloud, edge, and end device, ACM
Transactions on Intelligent Systems and Technology (TIST)
13 (5) (2022) 1–24.
[220] R. Gupta, I. Gupta, D. Saxena, A.K. Singh, A differential
approach and deep neural network based data privacy-
preserving model in cloud environment, Journal of Ambient
Intelligence and Humanized Computing (2022) 1–16.
[221] J. Vasa, A. Thakkar, Deep learning: Differential privacy
preservation in the era of big data, Journal of Computer
Information Systems (2022) 1–24.
[222] M.H. Rahman, M.M. Mowla, S. Shanto, Differential privacy
enabled deep neural networks for wireless resource
management, Mobile Networks and Applications 27 (5)
(2022) 2153–2162.
[223] J. Xiong, H. Zhu, Real-time trajectory privacy protection
based on improved differential privacy method and deep
learning model, Journal of Cloud Computing 11 (1) (2022)
1–15.
[224] F. Gava, L.M. Bayati, A scalable algorithm for
homomorphic computing on multi-core clusters, 2022 21st
International Symposium on Parallel and Distributed
Computing (ISPDC), IEEE (2022) 57–64.
[225] S. Meftah, B.H.M. Tan, K.M.M. Aung, L. Yuxiao, L. Jie, B.
Veeravalli, Towards high performance homomorphic
encryption for inference tasks on cpu: An mpi approach,
Future Generation Computer Systems 134 (2022) 13–21.
[226] J.A. Alzubi, O.A. Alzubi, A. Singh, M. Ramachandran,
Cloud-iiot-based electronic health record privacy-
preserving by cnn and blockchain-enabled federated
learning, IEEE Transactions on Industrial Informatics 19 (1)
(2022) 1080–1087.
[227] Z. Liu, Z. Gao, J. Wang, Q. Liu, J. Wei, et al., Ppefl: An edge
federated learning architecture with privacy-preserving
mechanism, Wireless Communications and Mobile
Computing 2022 (2022).
[228] J. Fiosina, ‘‘Interpretable privacy-preserving collaborative
deep learning for taxi trip duration forecasting,” in Smart
Cities, Green Technologies, and Intelligent Transport
Systems: 10th International Conference, SMARTGREENS
2021, and 7th International Conference, VEHITS 2021,
Virtual Event, April 28–30, 2021, Revised Selected Papers.
Springer, 2022, pp. 392–411.
[229] B. Jan, H. Farman, M. Khan, M. Imran, I.U. Islam, A.
Ahmad, S. Ali, G. Jeon, Deep learning in big data analytics:
a comparative study, Computers & Electrical Engineering
75 (2019) 275–287.
[230] J. Dean, G.S. Corrado, R. Monga, K. Chen, M. Devin, Q.V.
Le, M.Z. Mao, M. Ranzato, A. Senior, P. Tucker et al.,
Large scale distributed deep networks, in: Proceedings of
Advances in Neural Information Processing Systems, 2012.
[231] B. Abu-Salih, P. Wongthongtham, D. Zhu, K.Y. Chan, A.
Rudra, Social Big Data Analytics: Practices, Techniques,
and Applications, Springer Nature, 2021.
[232] A.A. Taha, W. Ramo, H.H.K. Alkhaffaf, Impact of external
auditor–cloud specialist engagement on cloud auditing
challenges, Journal of Accounting & Organizational Change
17 (3) (2021) 309–331.
[233] Z. Zhang, C. Wu, D.W. Cheung, A survey on cloud
interoperability: taxonomies, standards, and practice, ACM
SIGMETRICS Performance Evaluation Review 40 (4)
(2013) 13–22.
[234] H. Schulze, Cloud security report, Fortinet, Report, 2021.
[Online]. Available:
https://www.fortinet.com/content/dam/fortinet/assets/analy
st-reports/arcybersecurity-cloud-security.pdf.
[235] C. Ramalingam, P. Mohan, Addressing semantics standards
for cloud portability and interoperability in multi cloud
environment, Symmetry 13 (2) (2021) 317.
[236] Available: https://insightaas.com/cloud-interoperability-
andportability-necessary-or-nice-to-have/.
[237] R. Ranjan, The cloud interoperability challenge, IEEE Cloud
Computing 1 (2) (2014) 20–24.
[238] A. Romasanta and J. Wareham, Fair data through a federated
cloud infrastructure: Exploring the science mesh, in:
Proceedings of European Conference on Information
Systems, 2021.
[239] B. Abu-Salih, Domain-specific knowledge graphs: A
survey, Journal of Network and Computer Applications 185
(2021).
[240] H. Mezni, M. Sellami, S. Aridhi, F.B. Charrada, Towards
big services: a synergy between service computing and
parallel programming, Computing 103 (11) (2021) 2479–
2519.
[241] O. Adedugbe, E. Benkhelifa, R. Campion, F. Al-Obeidat,
A.B. Hani, U. Jayawickrama, Leveraging cloud computing
for the semantic web: review and trends, Soft Computing 24
(8) (2020) 5999–6014.

[242] I. Grangel-González, F. Lösch, and A. ul Mehdi, Knowledge
graphs for efficient integration and access of manufacturing
data, in: Proceedings of 25th IEEE International Conference
on Emerging Technologies and Factory Automation
(ETFA), vol. 1. IEEE, 2020, pp. 93–100.
[243] G. Ditzler, M. Roveri, C. Alippi, R. Polikar, Learning in
nonstationary environments: A survey, IEEE Computational
Intelligence Magazine 10 (4) (2015) 12–25.
[244] J. Lu, A. Liu, F. Dong, F. Gu, J. Gama, G. Zhang, Learning
under concept drift: A review, IEEE transactions on
knowledge and data engineering 31 (12) (2018) 2346–2363.
[245] M. De Lange, R. Aljundi, M. Masana, S. Parisot, X. Jia, A.
Leonardis, G. Slabaugh, T. Tuytelaars, A continual learning
survey: Defying forgetting in classification tasks, IEEE
transactions on pattern analysis and machine intelligence 44
(7) (2021) 3366–3385.
[246] T. Khan, W. Tian, G. Zhou, S. Ilager, M. Gong, R. Buyya,
Machine learning (ml)–centric resource management in
cloud computing: A review and future directions, Journal of
Network and Computer Applications (2022).
[247] R.S. Sutton, A.G. Barto, Reinforcement learning: An
introduction, MIT press, 2018.
[248] V. François-Lavet, P. Henderson, R. Islam, M.G. Bellemare,
and J. Pineau, An introduction to deep reinforcement
learning, arXiv preprint arXiv:1811.12560, 2018.
[249] Y. Li, Deep reinforcement learning: An overview, arXiv
preprint arXiv:1701.07274, 2017.
[250] B.R. Kiran, I. Sobh, V. Talpaert, P. Mannion, A.A. Al
Sallab, S. Yogamani, P. Pérez, Deep reinforcement learning
for autonomous driving: A survey, IEEE Transactions on
Intelligent Transportation Systems (2021) 4909–4926.
[251] F. Zeng, C. Wang, and S.S. Ge, ‘‘A survey on visual
navigation for artificial agents with deep reinforcement
learning,” IEEE Access, vol. 8, pp. 135 426–135 442, 2020.
[252] K. Arulkumaran, M.P. Deisenroth, M. Brundage, A.A.
Bharath, Deep reinforcement learning: A brief survey, IEEE
Signal Processing Magazine 34 (6) (2017) 26–38.
[253] K. Shao, Z. Tang, Y. Zhu, N. Li, and D. Zhao, A survey of
deep reinforcement learning in video games, arXiv preprint
arXiv:1912.10944, 2019.
[254] A. Gupta, A. Anpalagan, L. Guan, A.S. Khwaja, Deep
learning for object detection and scene perception in self-
driving cars:survey, challenges, and open issues, Array vol.
10, paper number 100057 (2021).
[255] N.Y. Zeng, P.S. Wu, Z.D. Wang, H. Li, W.B. Liu, X.H. Liu,
A small-sized object detection oriented multi-scale feature
fusion approach with application to defect detection, IEEE
Transactions on Instrumentation and Measurement 71
(2022).
[256] N. Rodríguez-Barroso, A.R. Moya, J.A. Fernández, E.
Romero, E. Martínez-Cámara, and F. Herrera, Deep
learning hyper-parameter tuning for sentiment analysis in
twitter based on evolutionary algorithms, in: 2019 Federated
Conference on Computer Science and Information Systems
(FedCSIS). IEEE, 2019, pp. 255–264.
[257] S.R. Young, D.C. Rose, T.P. Karnowski, S.-H. Lim, and
R.M. Patton, Optimizing deep learning hyper-parameters
through an evolutionary algorithm, in: Proceedings of the
Workshop on Machine Learning in High-Performance
Computing Environments, 2015, pp. 1–5.
[258] R. Tanabe and A. Fukunaga, Success-history based
parameter adaptation for differential evolution, in:
Proceedings of IEEE Congress on Evolutionary
Computation, 2013, pp. 71–78.
[259] M. Wu, W. Su, L. Chen, Z. Liu, W. Cao, K. Hirota,
Weightadapted convolution neural network for facial
expression recognition in human-robot interaction, IEEE
Transactions on Systems, Man, and Cybernetics - Systems
51 (3) (2021) 1473–1484.
[260] M. Gong, J. Liu, A. Qin, K. Zhao, K.C. Tan, Evolving deep
neural networks via cooperative coevolution with
backpropagation, IEEE Transactions on Neural Networks
and Learning Systems 32 (1) (2021) 420–434.
[261] M.R. Chen, B.P. Chen, G.Q. Zeng, K. Lu, P. Chu, An
adaptive fractional-order bp neural network based on
extremal optimization for handwritten digits recognition,
Neuro computing 391 (2020) 260–272.
[262] Z. Lu, G. Sreekumar, E. Goodman, W. Banzhaf, K. Deb,
V.N. Boddeti, Neural architecture transfer, IEEE
Transactions on Pattern Analysis and Machine Intelligence
43 (2021).
[263] B. Ma, X. Li, Y. Xia, Y. Zhang, Autonomous deep learning:
A genetic dcnn designer for image classification,
Neurocomputing 379 (2020) 152–161.
[264] F. Mattioli, D. Caetano, A. Cardoso, E. Naves, and E.
Lamounier, An experiment on the use of genetic algorithms
for topology selection in deep learning, Journal of Electrical
and Computer Engineering, no. Article ID 3217542, 2019.
[265] B. Wang, Y. Sun, B. Xue, and M. Zhang, A hybrid ga-pso
method for evolving architecture and short connections of
deep convolutional neural networks, in: Proceedings of the
16th Pacific Rim International Conference on Artificial
Intelligence, 2019.
[266] M. Suganuma, S. Shirakawa, and T. Nagao, A genetic
programming approach to designing convolutional neural
network architectures, in: Proceedings of Twenty-Seventh
International Joint Conference on Artificial Intelligence,
2017, pp. 5369–5373.
[267] A.D. Martinez, J. Del Ser, E. Villar-Rodriguez, E. Osaba, J.
Poyatos, S. Tabik, D. Molina, F. Herrera, Lights and
shadows in evolutionary deep learning: Taxonomy, critical
methodological analysis, cases of study, learned lessons,
recommendations and challenges, Information Fusion 67
(2021) 161–194.
[268] T. Serizawa and H. Fujita. (2021) Optimization of
convolutional neural network using the linearly decreasing
weight particle swarm optimization. [Online]. Available:
https://arxiv.org/abs/2001.05670.
[269] J. Fregoso, C.I. Gonzalez, G.E. Martinez, Optimization of
convolutional neural networks architectures using pso for
sign language recognition, Axioms vol. 10, number paper
139 (2021).
[270] Z. Fouad, M. Alfonse, M. Roushdy, A.-B.M. Salem, Hyper-
parameter optimization of convolutional neural network
based on particle swarm optimization algorithm, Bulletin of
Electrical Engineering and Informatics 10 (6) (2021) pp.

[271] X. Chen, Y. Sun, M. Zhang, D. Peng, Evolving deep
convolutional variational autoencoders for image
classification, IEEE Transactions on Evolutionary
Computation 25 (5) (2021).
[272] T. Zhang, C. Lei, Z. Zhang, X.-B. Meng, C.L.P. Chen, As-
nas: Adaptive scalable neural architecture search with
reinforced evolutionary algorithm for deep learning, IEEE
Transactions on Evolutionary Computation 25 (5) (2021).
[273] L. Xie and A. Yuille, Genetic cnn, in: Proceedings of IEEE
International Conference on Computer Vision, 2017.
[274] S. Pimminger, S. Wagner, W. Kurschl, and J. Heinzelreiter,
Optimization as a service: On the use of cloud computing
for metaheuristic optimization, in: Proceedings of the
International Conference on Computer Aided Systems
Theory, 2013, p. 348–355.
[275] A.D. Martinez, J.D. Ser, E. Villar-Rodriguez, E. Osaba, J.
Poyatos, S. Tabik, D. Molina, F. Herrera, Lights and
shadows in evolutionary deep learning taxonomy, critical
methodological analysis, cases of study, learned lessons,
recommendations and challenges, Information Fusion 67
(2021) 161–194.

Deep Learning Neural Networks in the Cloud

More Related Content

Similar to Deep Learning Neural Networks in the Cloud (20)

Recently uploaded (20)

Deep Learning Neural Networks in the Cloud