Assessment 3 Milestone 1: Machine Learning and the Internet of Things

Desklib provides past papers and solved assignments for students. This report analyzes machine learning's role in IoT.

Assessment-3 Milestone-1
Contents
Introduction............................................................1
Purpose...................................................................1
Problem Statement.................................................2
Conclusion..............................................................3
Reference................................................................3
Introduction
IOT has immense application in today’s word. With
the increasing use of IOT in real world applications,
the data is increasing day by day. Hence in order to
manage the immense amount of data generated,
machine learning algorithms have to be used and
implemented in the IOT devices. Hence this report is
aimed to understand the importance of machine
learning in IOT. Machine learning when embedded
into IOT may have some disadvantages also so the
report aims to study and analyse the advantages and
disadvantages of machine learning. Machine learning
for IOT is beneficial when the algorithms are in
accordance with the goal to be achieved. Machine
Learning can incorporate data into its algorithms but
lack of human intervention is affecting accuracy of
data association. IOT with its automated appliances is
a revolution but doesn't solve much when it comes to
complex phenomena and finding a substantial cause
and effect
Purpose
The purpose of this report is to critically analyse the
importance of Internet of things and Machine
learning. IOT is life transforming and will always
continue to be. IOT is making machines smarter and
in upcoming years almost every device that we can
even think of will be connected to the internet and
with each other to communicate (Yang et al., 2011).
IOT is present everywhere from household to
vehicles, machinery to buildings to transmit the data.
IOT generates huge amount of data which can be
analysed in real time and effective approaches can be
taken for example, IOT made possible to connect the
Location tracking system of Vehicles, with each other
(Liu and Wu, 2012). In case an accident occurs all the
nearby vehicles will automatically get a notification.
This is a single application of IOT but there are many
more applications of IOT. Machine learning is also a
popular field of AI which when embedded with IOT
can be very beneficial in making some real world
applications. Machine learning makes it possible to
analyse and monitor the huge amount of data that is
produced by IOT devices. The data generated will be
too large in upcoming years that it will be impossible
to analyse that with human efforts hence machine
learning will not only help in analysing the data but
also the algorithms will effectively allow the devices
to take decision on their own (Doshi-Velez and Kim,
2017). Therefore, it is important to study and analyse
the importance if machine learning field into IOT.
Hence this report is aimed to understand the
importance of IOT in today’s world, how IOT will
affect the future, what are the components of IOT
which are being used to build real world applications,
1

to understand how IOT evaluated. This report further
aims to identify the importance of machine learning
in IOT. Machine learning when embedded into IOT
may have some disadvantages also so the report aims
to study and analyse the advantages and
disadvantages of machine learning.
The topic chosen for the analysis is “Machine
learning in IOT”. As machine learning is making the
machines smart enough as humans and making
machines learn from the experiences and make
decisions based on the learning and experiences. IOT
can be defined as an ecosystem which makes several
devices to connect and communicate with each other
through internet. There are numerous benefits of
implementing machine learning with IOT. For
example, machine learning can collect and analyse
the data and learn from the previous experiences,
which can be further send to a IOT device to perform
operations. One such application can be Nest
thermostat which is smart device, works on machine
learning algorithm. It has a function of cooling and
heating the atmosphere, this device learns what
temperature needs to be set in someone’s house when
the person wakes up or when he enters into the house.
All these learning comes from observing the data
collected of temperatures, that person prefers. There
are infinite possibilities of making such real world
applications. IOT industry is growing at faster pace
hence the data generated is also going to be increased
in upcoming years. The increased data will give more
opportunities to introduce machine learning
algorithms into IOT devices (Li, Ota and Dong,
2018). The following are the applications of machine
learning into IOT:
 Machine learning with IOT will make the
energy saving machines.
 To indicate the fastest routes on the basis of
real time traffic congestion.
 Several devices like heart rate sensors etc. can
be built.
 Smart meters for gas, electricity and water can
be built and information regarding the usage
will be stored.
 IOT devices used in factories can be used
with machine learning algorithms to prevent
the abnormal conditions beforehand (Wuest et
al., 2016).
 In healthcare sector the personalised health
monitoring system can track the information
and inform the doctor immediately if anything
goes wrong.
 Insurance companies can analyse the data
gathered from the personal devices and assign
a proper insurance plan for the customer.
 The transportation services can use it for real
time transport management and real time
tracking of vehicles.
 Real time tracking of vehicles to prevent
accidents and to alert the driver.
Machine learning for IOT is beneficial when the
algorithms are in accordance with the goal to be
achieved. Because machine learning uses supervised
techniques to take effective decisions and use historic
data available. The more historic data is collected
better are the chances of correct prediction. Although
the IOT devices produces huge amount of data which
can only be analysed with the help of machine
leaning but the data which is needed it be analysed
depends upon the human effort to make the machine
in such a way that it keeps on leaning on the basis of
data collected and the machine can learn from the
environment then only the data analysis will be
effective. All these supervision requires human
efforts. (Shanthamallu et al., 2017)
Expected format of the report is A4 size paper with
two columns and justified text.
Problem Statement
2

Can machine learning be embedded into IOT without
human intervention?
According to various organizations Machine
Learning in tandem with IOT can store lots of data
and variables but without any human intervention a
pure relation between the two is difficult to achieve.
Machine Learning can incorporate data into its
algorithms but lack of human intervention is affecting
accuracy of data association. IOT with its automated
appliances is a revolution but doesn't solve much
when it comes to complex phenomena and finding a
substantial cause and effect when comes to
identifying associations in large data when it comes
to aiding complex human phenomena in the real
world (Marjani et al., 2017) To have practical
applications we need smart IOT that understands
complex human phenomena and that has smart IOT
devices that connects and optimizing devices which
will be the next phase of automation and actual
solutions of real world problems and phenomena.
Technologies like IOT, Machine Learning and AI are
a revolution when talked about individually. But we
need to optimize and integrate them together to
leverage their Maximum potential and that's where
we need human intervention to have a pure relation
and synergy between them so that they can have a
substantial effect on our civilization and true power
of these technologies in tandem cannot be imagined.
Human intervention is necessary which gives us
another roadblock i.e. Lack of Skilled engineers in
IOT and Machine Learning and when there's need of
human intervention to have a pure relation between
the two the group of people who possess such a skill
is a rare breed. Another roadblock is that such pure
relation between machine learning and internet of
things is a space consisting really complex
calculations that are required to facilitate all the
connections, automations and data to solve the real
world phenomena and to cross the barrier and
welcome the next generation synthesis of automation
and machine learning (Bandyopadhyay and Sen
2011). Billions of devices over the world are
connected to IOT and this number will escalate even
more in the coming years. The devices connected to
the IOT environment generate a deluge of
information which needs human intervention and
needs to be further monitored and analysed so that
the machine can have better understanding of the data
sets and learn continuously to improve at any manual
Intervention which will require processed data which
is a mammoth task ahead of all the organizations
working on these revolutionising technologies
(Canedo & Skjellum 2016). And with machine
learning we also have huge amount of data coming in
and we need to identify patterns in such data coming
from diverse sources which is required for efficient
decisions and appropriate analysis which are the
challenges faced in embedded systems working on
Internet of things and machine learning due to lack
human intervention (Chen, 2012).
Human Intervention is necessary to have a pure
relation between the world Internet of things and
Machine Learning to identify patterns and efficient
analysis of data to solve problems and complex
phenomenas in the real world.
Conclusion
As machine learning is making the machines smart
enough as humans and making machines learn from
the experiences and make decisions based on the
learning and experiences. IOT can be defined as an
ecosystem which makes several devices to connect
and communicate with each other through internet.
There are numerous benefits of implementing
machine learning with IOT. Machine learning for
IOT is beneficial when the algorithms are in
accordance with the goal to be achieved. Because
machine learning uses supervised techniques to take
effective decisions and use historic data available.
The more historic data is collected better are the
chances of correct prediction.
3

Reference
Bandyopadhyay, D., & Sen, J. (2011). Internet of
things: Applications and challenges in technology
and standardization. Wireless Personal
Communications, 58(1), 49-69.
Canedo, J., & Skjellum, A. (2016). Using machine
learning to secure IoT systems. In 2016 14th Annual
Conference on Privacy, Security and Trust (PST),
219-222.
Chen, Y. K. (2012). Challenges and opportunities of
internet of things. In 17th Asia and South Pacific
design automation conference, 383-388.
Doshi-Velez, F., & Kim, B. (2017). Towards a
rigorous science of interpretable machine
learning. arXiv preprint arXiv:1702.08608.
Li, H., Ota, K., & Dong, M. (2018). Learning IoT in
edge: deep learning for the internet of things with
edge computing. IEEE Network, 32(1), 96-101.
Liu, X. Y., & Wu, M. Y. (2012). Vehicular CPS: an
application of IoT in vehicular networks. Jisuanji
Yingyong/ Journal of Computer Applications, 32(4),
900-904.
Marjani, M., Nasaruddin, F., Gani, A., Karim, A.,
Hashem, I. A. T., Siddiqa, A., & Yaqoob, I. (2017).
Big IoT data analytics: architecture, opportunities,
and open research challenges. IEEE Access, 5, 5247-
5261.
Shanthamallu, U. S., Spanias, A., Tepedelenlioglu,
C., & Stanley, M. (2017). A brief survey of machine
learning methods and their sensor and IoT
applications. In 2017 8th International Conference
on Information, Intelligence, Systems & Applications
(IISA), 1-8.
Wuest, T., Weimer, D., Irgens, C., & Thoben, K. D.
(2016). Machine learning in manufacturing:
advantages, challenges, and applications. Production
& Manufacturing Research, 4(1), 23-45.
Yang, Z., Yue, Y., Yang, Y., Peng, Y., Wang, X., &
Liu, W. (2011). Study and application on the
architecture and key technologies for IOT.
International Conference on Multimedia
Technology ,747-751.
4