Network planning for next generation networks based on deep

learning techniques

A SYNOPSIS SUBMITTED TO LAHORE COLLEGE FOR WOMEN UNIVERSITY IN

PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY IN COMPUTER SCIENCE

By
________________________

SUPERVISOR
____________________________
Professor
Department of Computer Science, LCWU.

CO-SUPERVISOR
____________________________
Assistant Professor
Department of Computer Science, LCWU

___________________________________________________
DEPARTMENT OF COMPUTER SCIENCE
LAHORE COLLEGE FOR WOMEN UNIVERSITY, LAHORE
2022
Contents
Abstract............................................................................................................................................................................................ 1
I. INTRODUCTION.................................................................................................................................................................. 1
II. Literature review................................................................................................................................................................ 3
III. Research methodology................................................................................................................................................ 6
IV. Conclusion............................................................................................................................................................................. 12
References..................................................................................................................................................................................... 12

Abstract
Machine learning is being incorporated and integrated within every sector of business and life. As
wireless networks are complex, and moving towards 5 and 6 generations could bring more
complications. So, there is a need of an entity or automation which could handle this, machine
learning is the best fit solution for this. The synopsis research highlights the network planning for
next generations network based on deep learning through pondering deeper insight of literature.
deep learning has proven its remarkable success in multiple fields due to its agile representation
ability, and computational ease. That is why, researches and studies are now focusing on
incorporating the potential of deep learning in accomplishing the lacking of model based methods.
Deep learning could be very fascinating for 6 generation application. Result methodology adopted
for this synopsis is qualitative and secondary data collection techniques were opted as well. The
paper has been discussed and analyzed in the discussion part. In the end the conclusion was made.

Keywords—next generation network, networking, generation networking, deep learning deep

learning and network generation

I. INTRODUCTION
The next-generation networks (NGNs) are designed for the provisioning of use cases defined by
3GPP, from ultra-reliable low latency communication (URLLC) services to enhanced mobile
broadband (eMBB) and massive machine type communications (mMTC) [1]. These use cases are
offered as services that should be capable to sustain the tight requirements needed by applications
like virtual reality, vehicle-to-all (V2X) and mission critical communications. Inevitably, new
deployments and corresponding infrastructures supporting 5G services are being planned. In this
context, network planning assisted by network data is extremely significant as the process of
identifying the highest traffic region at cluster level, determining the appropriate number of clusters
and to provide lower cost per bit. The network planning not only provides the intended coverage
and capacity to the subscribers [2], [3] but it is also an effective way to reduce capital and
operational expenditure (CAPEX and OPEX, respectively) for the mobile network operators
(MNOs). Network generation is painstaking to be a cellular network which is referred by its
number that starts with 1 G, 2 G, 3 G, and so on. [1]. The most recent generation is 5 G. they are
called as generation because this is how they were defined institutionally. On the other hand, there
is deep learning, which is considered to be a type of machine learning along with artificial
intelligence which emulates the way that human thinking and attaining knowledge. This includes
predictive modeling and data science. Applications like virtual reality, vehicle-to-all (V2X), and
mission-critical communications rely on stringent requirements, so these use cases are provided as
a service by Next Generation Networks (NGNs). These use cases include Ultra Reliable Low
Latency Communications (URLLC) services, Enhanced Mobile Broadband (eMBB), and Massive
Machine Type Communications (mMTC) [1]. Plans for new infrastructure and 5G service
deployments are inevitable. As the process of identifying the regions with the highest traffic at the
cluster level, determining the appropriate number of clusters, and reducing the cost per bit, network
planning, supported by network data, is very important in this context. Network planning helps
mobile network operators (MNOs) reduce their capital and operating expenses (CAPEX and
OPEX, respectively) in addition to providing subscribers with the coverage and capacity they
desire. There is a need of flexible and efficient infrastructure in service driven next generation
network, such as low latency and ultra-reliable communication, enhanced model broadband, and
massive machine type communications. It should be capable of supporting existing access of
standard generations like long term evolution, fifth generation and Wi-Fi. It should be capable of
coordinating a heterogenous network with base stations, which encompass micro, macro, Pico,
femto, BSs and applications. Foundation of Future connected society tend to evolve with the advent
of 5 and 6 generation wireless network and its convergence with vertical applications. Along with
complexity of these network, the user equipment, and base stations are the primary challenges
which generate from management stages, planning to development stages, and situation dependent
operations. 5 generation and beyond are heterogenous and multitier. It consists of dense small cells
which assist in achieving 1000-fold capacity, hence increasing cost effectiveness. It should be kept
in mind that, for network planning, manual configuration, optimization, and control will make
things more complicated. Human machine interaction can ne time consuming to be aligned, and
there are errors in it as well. Network operators are mainly dependent on KPIs that conglomerate
from different parts of network in order to derive a decision. The decision is analyzed and then
employed as well. The rapid proliferation of smart smartphones and the growing demand for
connectivity have led to the improved popularity of the 5G technology in North America and
Europe. With the growing popularity of 3G and 4G network in the regions, the infrastructure for 5G
technology is also expanding. The growing mobile device demand is also anticipated to drive 5G
adoption. Wireless communication has proved to be highly effective in improving the quality of
services by increasing the speed and reducing the latency of the services, which is a key factor to
support the enlargement of the 5G technology. The growth in the availability of high-speed
connections is expected to drive the market for 5G connectivity. However, the availability of 4G
and 4.9G network is going to hinder the growth of the 5G connectivity market. In terms of type of
technology, 4G LTE has dominated the 5G connectivity market in terms of revenue, whereas it is
expected to be challenged by 5G MIMO technology during the forecast period. Over the forecast
period, it is anticipated that the market for 5G connectivity in APAC will expand at the highest
CAGR. The growth of the market for 5G connectivity is anticipated to be driven by emerging
economies like China and India. The increasing demand for high-speed data and improved network
coverage, along with the implementation of automation, are driving the growth of the 5G
connectivity market. The developed countries are also implementing 5G technology in their
networks to enhance the connectivity and to achieve a digital transformation. In June 2018, O2
announced the launch of the first 5G smartphone in the U.K. Through the implementation of 5G,
the cellular operators will be able to allow mobile data transfer rate up to 1GB per second. Similar
to other segments of the connectivity market, the 5G solution provider segment is expected to hold
a large share of the global 5G connectivity market during the forecast period. Rising demand for
enhanced connectivity solutions and services is leading to the growth of the 5G solution provider
segment in the 5G connectivity market. 5G technology is currently being deployed for massive
integration in the military and aerospace sectors. The use of 5G technology will help create high
quality data and communications for the development of advanced UAVs. [14]. [2]

II. Literature review

From literature research I have gotten the insight that smart environments in our daily life have
become common and its importance lies on considering design needs of multiple
telecommunication networks. These networks are capable of allowing transmission of data and
information between components of different environment, such as middleware, internet of things,
and applications. Data network is considered to be significantly dynamic thing which opt behavior
as per the environment, without data network there will be less optimal communication and
functions will be limited as well. [13]. In the contemporary world, the communication networks are
limited by layered protocol architecture, that result in unconsciousness of network state
experienced by different factors. Moreover, the network response to condition has limited and
isolated scope as well, sometime outcomes as suboptimal performance. There is a need of
redesigning next generation networks and transforming them into cognitive network which are
painstaking to be intelligent. Network should be simple and easy to manage. Recently, reactive
mechanism is developed in order to improve the connectivity of network, and addressing different
problem that occur during operation. However, it brings some changes to the network, such as they
are incapable of collecting data and adapt new changes etc. from this point deep learning leads the
way and enable every desired protocols so that every network condition can be observed and by
using previous comprehensive information to respond efficiently. [4]. Deep learning is adopted for
many applications regarding next generation of cognitive network. Deep learning is also a type of
machine learning in which algorithms are run with logical reasoning. Machine learning is a subfield
of artificial intelligence. Machine learning was adopted for Artificial intelligence after 1980s
because it enables the learning ability without programming. It is also capable of enabling
computer to derive data and carry out different tasks. However, it is important to know that
machine learning involves probability, computer science, and statistics that enable the stochastic
algorithm. It assists them in self-learning when they are engaging with new data. from literature I
have gotten insight regarding application of deep learning in next generation network in order to
attain meta picture of highlights and performances. Also, for analyzing tools, that assist in solving
problems of networks. There are also fundamentals of deep learning which should be incorporated.
The requirements and use should be addressed as well. [5].
Network planning has received a lot of attention for 3G/4G networks, but new 5G deployments
face a number of challenges in terms of cost-efficiency. Authors of, has proposed a service-based
network dimensioning framework for new 5G deployments. In this framework, relevant problems
for network planning have been discussed like network data acquisition to identify highest traffic
region, service-based network dimensioning to find required number of ratio sites and capacity
model to guarantee peak data rates to fulfill MNOs’ requirements. The same authors have discussed
the unsupervised clustering framework to minimize the cost per MB given that the MNOs’
technical and financial requirements have been fulfilled. The results show significant cost savings
can be achieved while minimizing the cost per MB and maximizing the potential network
utilization by identifying the highest traffic density area. However, the proposed framework lacks
the self-planning capabilities to adapt corresponding deep-learning techniques to enhance the
performance and cost-efficiency in the network planning of the NGNs. In literature there have been
researches made on artificial intelligence based wireless networks. Demand of intelligence wireless
network make incorporation of artificial intelligence more important. Its employing capabilities are
data driven which network nodes can easily determine by employ based policy. It also reduces the
dependency on other mathematical models which are reliance based as far as network operation and
design is concerned. However, it does not mean that traditional mathematical models can be
neglected. In fact, it will be more efficient to use both models based and artificial intelligence-
based networking generations to improve productivity, and efficiency. [6].

Over the past few years, wireless communication systems have made significant revolutionary
progress. The challenges posed by the rapid increase in traffic and the immediate requirement for
services cannot be fully met by the commercial deployment of 5G applications that are currently
being deployed worldwide. The next generation (NG) of communication systems and networks are
already being developed in this regard by academia and industry. Data rates, bandwidth, quality of
experience, quality of service, and operating frequency, among other factors, determine the
proportion of each generation among technologies ranging from 1G to 5G. Efficient allocation of
network resources while ensuring end-user QOE can also be used to maximize profits, but this is a
major obstacle in wireless networks. of workers. There are many obstacles to overcome when
providing a particular service. B. Strict QoS, human-centric features, large numbers of devices,
heterogeneous networks and devices, and heterogeneous environments. Now that the term Quality
of Experience (QoE) has become more popular, much research is being done to create or provide
services that are reliable, high quality, and have a great user experience.Quantification of Quality of
Service (QoS) and Quality of Experience (QoE) for 5G technologies: The author of the review
claims that his 5G QoS parameters such as packet loss, loss rate, network delay and travel time
become less effective, especially for video-based multimedia communications. QoS parameters
such as buffers start times, and acceptable bitrates do not produce high QoE values because there
are many unpredictable service-related user satisfaction parameters. Another way to match QoS
and QoE scores is to combine objective and subjective parameters. We collected subjective
parameters using the Mean Opinion Score (MOS) method of user satisfaction and objective
parameters using packet loss rate and latency information about communication traffic. The
information for quantifying QoE in the various pre-5G technologies. QoE is generally measured as
real-world tests for specific scenarios such as video. [7].

Based on our testing, we determined that video appears to experience worse QoE in 5G compared
to 3G and 4G. For 4G to match a 5G network, the QoE scores would have to be much higher (30).
The author also studied 5G wireless routers with more than one 3G or 4G radio, suggesting that 5G
routers will have a shorter QoE. Some potential reasons include network congestion, local and
remote control of the 5G router. Cannibalization is discussed ahead. Another factor is the fact that
many current mobile operators are committed to 4G technology and in many places, including the
U.S., they are making plans to complete 5G upgrades by 2020. But some experts are saying that
even the early deployments of 5G may add too much capacity. "It's going to hurt 4G even though
we've deployed much more spectrum for it," said John Gartner, a telecommunications consultant.
"The answer will be congestion management and there may be more capacity, but it may not be for
mobile services." However, the 5G process may be faster because of 4G LTE network capabilities
and 5G's ability to leverage those 4G assets. The 5G Mobile Trial Network, the first U.S.-based test
bed for 5G, will be available in Seattle, Indianapolis and two other cities to help mobile operators
and device makers begin to understand 5G's impact on the business and consumer worlds. It is said
that our consumer devices will become a 'single wire virtual network terminal' with the 5G mobile
network. This would enable consumers to use both 5G fixed wireless access and mobile services
with devices like laptops and smartphones. They will be able to connect to fixed wireless and
mobile services with the same device. Service quality underpins 5G: The authors of "A recent study
on 4G network performance in Thailand" studied current QoS parameters in Thailand, such as
latency, jitter, loss, download speed, and speed upload. The system called MIQ (Mobile Internet
QoS Measurement Tool) was created.It was built with Ookla's API, which is also used to create a
very popular speed test app. The increase in end-to-end network bandwidth by using the 5G
technology was previously questioned. Mentioned study authors state that 5G networks use large
bandwidth (2-5 Gbps) but limited data rate, i.e., very low latency (less than 100ms), which allows
for high data rates with low latency while providing satisfactory QoS. However, it is not possible to
make concrete conclusions from such data in 5G in the short term. The measurement of quality of
service is done with purpose of decision-making in order to increase efficiency, optimize operation
of business and improve customer satisfaction. It is performed by the engineers of
telecommunications and networks in the network layer as well as the access layer. [15].

2D Video and Next Generation Virtual and Augmented Reality: Enabling efficient streaming over
QoE-enabled mobile networks provided a model that used basic IP network performance metrics to
predict video streaming quality. A lightweight application QoS and QoE predictor, LEAP, has also
been proposed. It can give you a complete picture of how network performance affects video
streaming applications and the user experience they provide. The following four network
performance indicators can be used by the model:Throughput, jitter, packet loss, and latency are the
four main issues. The model predicts three video playback performance metrics that can be used to
evaluate performance during video playback: Improving Quality of Experience (QoE) for future
wireless ecosystems: i) startup time, duration of each booth and number of booths A machine
learning approach) used big data and machine learning (ML) to create an overall QoS management
model (QoXphere) for the next generation of wireless ecosystem. To identify user-related KQIs
(key quality indicators) as well as network performance anomalies (NPs), the methodological
approach uses supervised and unsupervised machine learning techniques monitor. The next
generation of mobile networks will rely heavily on the performance and quality of experience
(QoE) of video traffic. “Enable Controllable and Premium Video Streaming QoE with 5G Network
Orchestration” describes a method to improve end-user experience for multiple DASH (Dynamic
Adaptive Streaming over HTTP) clients by combining Advanced network orchestration with
MPEG-DASH SAND protocol. The authors of the "Survey on QoE Evaluation of Multimedia
Services and Machine Learning-Based Prediction" investigated application-driven, ML-based QoE
prediction models for quality management of experiences for multimedia services. The HEVC
(High Efficiency Video Coding) standard can now provide multimedia services such as video
streaming over 4G/LTE. Due to its 50% compression advantage over the Advanced Video Coding
(AVC) standard, the HEVC standard partially solves the problem of limited bandwidth. The
authors of "Next Generation Emergency Service Video Transmission" conducted extensive and
objective research on the effects of HEVC encoding and LTE transmission on video quality. This is
necessary to avoid wasting scarce bandwidth and to provide the required original video quality. For
service providers to improve services and deliver quality services to end users, it is essential that
users have a positive quality experience (QoE) with video content. The authors of "Measuring
Effect of Packet Reordering on Quality of Experience (QoE) in Video Streaming") have suggested
several tests that use artificial packet reordering to measure the degree of user stratification in video
quality. and network acceptability to end users of the video. diffusive. These experiences negatively
affect end-user QoE and video quality. The NetEm engine is used to record and play back video so
participants can see how they feel, create artificial rearrangements of packets in a video stream, and
study the quality of experience (QoE) of the videos. Different rearrangement percentages of some
packages. Tests have shown that when video quality is reordered due to network traffic, user QoE
will be affected. Wireless channel models must be able to incorporate QoS-sensitive evaluation
metrics such as latency, data rates, and latency breach probability for latency-sensitive real-time
applications. QoS metrics are not considered in the currently used physical layer channel models.
Effective Capacity (EC) is the link layer wireless channel model proposed in Efficient Capacity in
Using QoS parameters, a QoE model is developed for QoE-aware access (QoE-aware automatically
adjusts the service priority factor to optimize resource allocation in LTE networks). To select the
most suitable parameter for the service priority factors, the authors propose to use the particle
genetic algorithm (PGA). The proposed algorithm is efficient because it simultaneously increases
the average QoE of the network while maintaining the QoE threshold for each multiservice. It is
believed that customer satisfaction or dissatisfaction with a service, such as calling or watching TV,
can be measured by the quality of experience (QoE) in the network. Similar to the field of user
experience, which originates from telecommunications, the quality of the experience is completely
dependent on the service experience. Researchers have proposed a WiFi6-specific QoS predictor
for next-generation data-based WiFi6 campus networks. This tool uses a spatial-temporal graphical
approach to predict throughput and other QoS parameters related to user location in a WiFi6
network. At the mobile level, the quality of the experience has improved. Researchers have created
a real-time interactive visualization system that allows network planners or administrators to
effectively plan and monitor next-generation communication networks by displaying network data.
as large-scale as router configuration and port-level QoS metrics. Provide low-latency services
across a wide range of virtual resources: Current research on supporting low-latency services
during their runtime. [15].

III. Research methodology

Research methodology encompass the designing of algorithm based on artificial intelligence and
deep learning. Model prepared through multiple parameter values according to Mos requirements
ad put in the algorithm. Then we will show how much qoe enhance.
nterrelated aspects, as telecommunication service is not instantaneous. QoE is a signal of a great
change in user experience in telecommunication sector that transforms users to comfortable and
secure environments that utilize mobile technologies. M2M communication field has effective
method of implementation of QoE in this technology, because it is developing its scope globally
and have high reliability.

QoE has three aspects; infrastructure, service quality, and end user device support. As infrastructure
is the major part of QoE, it is a major part of a very advanced M2M communication field. Wireless
network consists of numerous layers, application layer is the fundamental layer in a wireless
network whereas protocol layer is used to exchange information between computing devices.
Application layer is majorly responsible for exchanging information within the network. In the past
decade, these layers have seen major developments and advancement and the market is expecting
the same. Research on new networking technologies and their practical applications is a key factor
driving the growth of the wireless network segment. The market for wireless networking
technologies is growing on the backdrop of rising demand for low-cost data, continuous increase in
the adoption of internet of things and rising demand for smarter network solutions for various
applications across vertical industries such as BFSI, retail and manufacturing. Another factor
driving the growth of the market is the growing popularity of wireless high-speed communication
among businesses which is fuelling demand for wireless networking technology in the coming
years. A big advantage of using deep learning for network planning is that it can help identify
hidden patterns and correlations in the data. This can help us to more accurately predict traffic
demands and plan our networks accordingly. Deep learning can also be used to improve the
accuracy of simulations. By feeding simulations with data that has been pre-processed by a deep
neural network, we can produce more realistic results that are closer to what we would expect to
see in the real world. Overall, deep learning offers some exciting possibilities for the future of
network planning. With its ability to identify hidden patterns and correlations, as well as its
improved accuracy in simulations, it has the potential to create next generation networks that are
smarter and more efficient than ever before. When it comes to creating next generation networks,
deep learning can help you reap a variety of benefits. First, you can have better network planning.
With deep learning technologies, you can generate a more efficient network architecture and plan
for future development. This way, you will be able to deploy services more quickly and reduce
operational costs as well. [8]. Furthermore, you can use deep learning to optimize network
performance in terms of speed, bandwidth and latency. You will be able to better manage traffic
load, decrease latency and improve the end user experience. Additionally, deep learning can also
help reduce energy costs by enabling the efficient use of resources such as servers or switches.
Finally, leveraging deep learning for next generation networks can help enhance security protocols
by accurately predicting threats and taking proactive measures before they become a problem. This
way, you will be able to ensure a secure environment for all users on the network. Creating next-
generation networks based on deep learning techniques comes with its own set of challenges. For
example, the need for huge datasets and computational resources can make it difficult to develop
these networks in a cost-efficient manner. Additionally, a lack of interpretability for deep learning
models can make it difficult to understand the impact of specific decisions made by the model.
Security is also a concern when creating next generation networks with deep learning techniques.
Deep learning models are often trained on private data sets which can be vulnerable to malicious
attacks. As such, these networks need to be built using security measures such as secure
authentication and encryption to protect against any potential threats. Furthermore, the network
may need to be monitored continuously to ensure that any malicious activities are identified quickly
and appropriately mitigated. [10].
Table1:CapacityModel

InputParameters OutputParameter

Wth Wma UD P Rx Zeta 30 50 70

x L

10 2 200 0.35 100 0.2616 3.8393 23.8393 43.8393

15 2 200 0.35 100 0.0617 23.8266 43.8266 63.8266

20 2 200 0.35 100 0.0121 28.7882 48.7882 68.7882

25 2 200 0.35 100 0.0021 29.7867 49.7867 69.7867

30 2 200 0.35 100 3.4961e-04 29.9650 49.9650 69.9650

35 2 200 0.35 100 5.4461e-05 29.9946 49.9946 69.9946

40 2 200 0.35 100 8.1695e-06 29.9992 49.9992 69.9992

45 2 200 0.35 100 1.1902e-06 29.9999 49.9999 69.9999

50 2 200 0.35 100 1.6943e-07 30.0000 50.0000 70.0000

Intable1,fixed3parametersvalue,UDL=200,P=0.35&Rx=100.AccordingtoUDLandRXvalue,getth
evalueof Wmax. Wth parameter value increases with difference of 5. Higher the Wth, lower the
zeta value. Both areinversely proportional.
We take another variable, named as (Desired Zeta) and compared it with algorithm Zeta. In table
1, when wehave10Wth,thendifference is3.8393forzeta30,whichiscloser toalgozeta.Same in the
casefor 50and70
%ofzeta,mentionedintable.
Table2:CapacityModel2

InputParameters OutputParameter

Wth Wma UDL P Rx Zeta 30 50 70

x

10 4 200 0.35 50 0.7515 45.1496 25.1496 5.1496

15 4 200 0.35 50 0.3519 5.1943 14.8057 34.8057

20 4 200 0.35 50 0.1182 18.1803 38.1803 58.1803

25 4 200 0.35 50 0.0320 26.7952 46.7952 66.7952

30 4 200 0.35 50 0.0075 29.2482 49.2482 69.2482

35 4 200 0.35 50 0.0016 29.8410 49.8410 69.8410

40 4 200 0.35 50 3.1122e-04 29.9689 49.9689 69.9689

45 4 200 0.35 50 5.7375e-05 29.9943 49.9943 69.9943

50 4 200 0.35 50 1.0085e-05 29.9990 49.9990 69.9990

In table 2, fixed 3 parameters value, UDL =200, P=0.35 & Rx=50. According to UDL and RX
value, get the valueof Wmax. Wth parameter value increases with difference of 5. Higher the
Wth, lower the zeta value. Both areinversely proportional.
We take another variable, named as (Desired Zeta) and compared it with algorithm Zeta. In table
2, when wehave15Wth,then differenceis5.1943for zeta30,whichiscloser toalgozeta.Same in
thecase for 50and70 %ofzeta,mentionedintable.
Once the 5G fixed wireless access portion of the 5G upgrade is in place, it is expected to extend
mobile broadband services to rural areas by utilizing low-band spectrum. In contrast, the same 5G
technology can be deployed for fixed wireless access to residential areas and has the potential to
help deliver mobile broadband service in addition to fixed broadband services. Limited Use Cases
for 5G encompass Even though there are potential use cases for 5G that are very high, there are
many high-value uses for current 4G services such as video, mobile gaming and, home automation.
The author argued that because of the high-volume demands on 4G, it is not enough for wireless
operators to add 4G and 5G to their networks. Instead, they should develop a coordinated use case-
by-use case strategy. The author indicated that currently a business that needs to deliver a network
service in a short time frame may benefit from a short-term investment in 5G. While a consumer
might not need these new capabilities right now, they might need them in the future. "It's a
balancing act between what consumers need now and what is important for the future," said
Edward Leonard, wireless systems engineer at Applied Analysis. "If the 5G networks are deployed
on an uncoordinated basis, it could slow down network speeds and reduce overall network speed."
Technology Has a Long Development and Market Life Cycle Because of its complexity, a 5G
network could not be rolled out immediately. The author said that industry carriers have indicated
that it will take three to five years to roll out 5G. Leonard said that the majority of the industry's 4G
investments are not yet complete. He added that there are opportunities for improvement in 4G
network[9].

IV. Conclusion
The next-generation organizations (NGNs) are intended for the provisioning of purpose cases
characterized by 3GPP, from super solid low idleness correspondence (URLLC) administrations to
improved versatile broadband (eMBB) and gigantic machine type interchanges. These utilization
cases are offered as administrations that ought to be competent to support the tight necessities
required by applications like virtual reality. Next generation like 5 G and 6 G will bring more
complications but machine learning and deep learning will cover the complications. With deep
learning innovations, you can produce a more effective organization engineering and plan for future
turn of events. Along these lines, you will actually want to send benefits all the more rapidly and
lessen functional expenses also. Besides, you can involve deep learning to advance organization
execution with regards to speed, data transmission and idleness. You will actually want to all the
more likely oversee traffic load, decline inertness and further develop the end client experience.
Moreover, deep learning can likewise assist with lessening energy costs by empowering the
effective utilization of assets like servers or switches. At last, utilizing deep learning for next
generation organizations can assist with improving security conventions by precisely anticipating
dangers and going to proactive lengths before they become an issue. Along these lines, you will
actually want to guarantee a solid climate for all clients on the organization. Making next-
generation networks in light of deep learning methods accompanies its own arrangement of
difficulties. For instance, the requirement for immense datasets and computational assets can make
it challenging to foster these organizations in an expense effective way. Moreover, an absence of
interpretability for deep learning models can make it challenging to comprehend the impact of
explicit choices made by the model. Security is likewise a worry while making next generation
networks with deep learning procedures. Deep learning models are often prepared on confidential
informational collections which can be powerless against malevolent assaults. Thusly, these
organizations should be constructed utilizing safety efforts, for example, secure confirmation and
encryption to safeguard against any likely dangers. Moreover, the organization might should be
checked consistently to guarantee that any malignant activities are recognized rapidly and fittingly
alleviated.

References

[1]. Bogale, T. E., Wang, X., & Le, L. B. (2018). Machine intelligence techniques for next-
generation context-aware wireless networks. arXiv preprint arXiv:1801.04223.

[2]. Kibria, M. G., Nguyen, K., Villardi, G. P., Zhao, O., Ishizu, K., & Kojima, F. (2018). Big data
analytics, machine learning, and artificial intelligence in next-generation wireless networks. IEEE
access, 6, 32328-32338.

[3]. Tomkos, I., Klonidis, D., Pikasis, E., & Theodoridis, S. (2020). Toward the 6G network era:
Opportunities and challenges. IT Professional, 22(1), 34-38.

[4]. Zappone, A., Di Renzo, M., & Debbah, M. (2019). Wireless networks design in the era of deep
learning: Model-based, AI-based, or both?. IEEE Transactions on Communications, 67(10), 7331-
7376.
[5]. Ozpoyraz, B., Dogukan, A. T., Gevez, Y., Altun, U., & Basar, E. (2022). Deep learning-aided
6G wireless networks: A comprehensive survey of revolutionary PHY architectures. arXiv preprint
arXiv:2201.03866.

[6]. Buenrostro-Mariscal, R., Santana-Mancilla, P. C., Montesinos-López, O. A., Nieto Hipólito, J.

I., & Anido-Rifón, L. E. (2022). A Review of Deep Learning Applications for the Next Generation
of Cognitive Networks. Applied Sciences, 12(12), 6262.

[7]. Ibarrola, E., Davis, M., Voisin, C., Close, C., & Cristobo, L. (2019). QoE enhancement in next
generation wireless ecosystems: a machine learning approach. IEEE Communications Standards
Magazine, 3(3), 63-70.

[8]. Shen, X., Gao, J., Wu, W., Lyu, K., Li, M., Zhuang, W., ... & Rao, J. (2020). AI-assisted
network-slicing based next-generation wireless networks. IEEE Open Journal of Vehicular
Technology, 1, 45-66.

[9]. Yang, F., & Ma, J. (2019). Deep-learning inversion: A next-generation seismic velocity model
building method. Geophysics, 84(4), R583-R599.

[10]. Friederich, P., Häse, F., Proppe, J., & Aspuru-Guzik, A. (2021). Machine-learned potentials
for next-generation matter simulations. Nature Materials, 20(6), 750-761.

[11]. Sarieddeen, H., Saeed, N., Al-Naffouri, T. Y., & Alouini, M. S. (2020). Next generation
terahertz communications: A rendezvous of sensing, imaging, and localization. IEEE
Communications Magazine, 58(5), 69-75.

[12]. Ghorbani, H., Mohammadzadeh, M. S., & Ahmadzadegan, M. H. (2020, April). Modeling for
malicious traffic detection in 6G next generation networks. In 2020 International Conference on
Technology and Entrepreneurship-Virtual (ICTE-V) (pp. 1-6). IEEE.

[13]. Shafique, K., Khawaja, B. A., Sabir, F., Qazi, S., & Mustaqim, M. (2020). Internet of things
(IoT) for next-generation smart systems: A review of current challenges, future trends and
prospects for emerging 5G-IoT scenarios. Ieee Access, 8, 23022-23040.
[14]. Kelleher, J. D. (2019). Deep learning. MIT press.

[15]. Ibarrola, E., Davis, M., Voisin, C., Close, C., & Cristobo, L. (2019). QoE enhancement in
next generation wireless ecosystems: a machine learning approach. IEEE Communications
Standards Magazine, 3(3), 63-70.