IEEE COMMUNICATIONS MAGAZINE, VOL. X, NO. X, XXXXX 201X 7
are not readily available for many application domains.
It is required to confirm results based on simulated big
data.
• Context-awareness: Integrating contextual information
with raw data is crucial to get more value of the data and
perform faster and more accurate reasoning and actuation
[14]. For example, detecting a sleepy face in a human
pose detection system could lead to totally different
actions in the contexts of driving a car and relaxing at
home.
In addition, there are other challenges that affect the design
of a smart city ecosystem such as integration of different
analytic frameworks, distribution of analytic operations, and
lack of comprehensive testbeds.
B. Future Directions
The conventional analytic approach for IoT is to send raw
data to the cloud for processing. However, this scheme is not
effective and scalable for smart city deployments. Decentral-
ization of data analytic computations is a new trend that aims
to bring analytics closer to the fog and IoT devices [6], [8].
Here we list several promising future research directions in
this regard.
• A trained model works well when the same feature set
and distribution model forms the training and test data
sets. By changing the distribution, the trained model
needs to be rebuilt from new training data. For example,
in a radio frequency based localization application (e.g.,
WiFi, BLE), the RSSI values for the same time and posi-
tion in Android and iOS devices are different. The trained
localization model on one platform can be transferred to
the new platform without the need to collect RSSI values
for other devices. Transfer Learning is a field of research
that can help in such scenarios [15].
• Integration with semantic technologies is also a need for
the development of smart city applications. The need
stems from the interaction of those systems with citizens
and the use of social media data.
• Intelligent virtual objects can be used in smart city ser-
vices joint with DRL algorithms, considering each phys-
ical object has a virtual representation in the smart city
and these VOs can learn, decide, and act autonomously.
for the new generation of smart city systems since citizens
are the main players in smart cities. The small size of
mobile devices and wearables nowadays does not allow
space for touch screens or keyboards. Instead, automatic
speech recognition and natural language understanding is
the most convenient way of interaction with these devices.
VII. CONCLUSION
There are many machine learning algorithms
that can be uti-
lized to learn from the big data collected through a smart city’s
infrastructure. However, most traditional machine learning
techniques assume a fixed training model and a static context.
These assumptions do not apply to smart city applications
where the environment and consequently the training data
evolve over the time.
In this article, we addressed challenges and opportunities
that arise when utilizing machine learning to realize new
smart city services. These challenges include: data-recycling,
efficient sampling, and devising scalable models. We reviewed
state-of-the-art methods and approaches that embrace smart
city big data toward future cognitive smart cities. Then, a
hierarchical framework was introduced to incorporate machine
learning techniques in accord to the hierarchy of big data in
smart city. We proposed a semi-supervised deep reinforcement
learning framework to address the presented challenges and
highlighted the position of the framework in various smart city
application domains. Finally, we articulated several challenges
and trending research directions for incorporating machine
learning to realize new smart city services.
REFERENCES
[1] A. Al-Fuqaha, M. Guizani, M. Mohammadi, M. Aledhari, and
M. Ayyash, “Internet of Things: A survey on enabling technologies,
protocols, and applications,” IEEE Communications Surveys & Tutorials,
vol. 17, no. 4, pp. 2347–2376, 2015.
[2] P. Kaltenrieder, E. Portmann, and T. Myrach, “Fuzzy knowledge rep-
resentation in cognitive cities,” in Fuzzy Systems (FUZZ-IEEE), 2015
IEEE International Conference on. IEEE, 2015, pp. 1–8.
[3] P. Vlacheas, R. Giaffreda, V. Stavroulaki, D. Kelaidonis, V. Foteinos,
G. Poulios, P. Demestichas, A. Somov, A. R. Biswas, and K. Moessner,
“Enabling smart cities through a cognitive management framework for
the internet of things,” IEEE Communications Magazine, vol. 51, no. 6,
pp. 102–111, 2013.
[4] Q. Wu, G. Ding, Y. Xu, S. Feng, Z. Du, J. Wang, and K. Long,
“Cognitive Internet of Things: a new paradigm beyond connection,”
IEEE Internet of Things Journal, vol. 1, no. 2, pp. 129–143, 2014.
[5] S. Feng, P. Setoodeh, and S. Haykin, “Smart Home: Cognitive Interactive
People-Centric Internet of Things,” IEEE Communications Magazine,
vol. 55, no. 2, pp. 34–39, 2017.
[6] B. Tang, Z. Chen, G. Hefferman, S. Pei, W. Tao, H. He, and Q. Yang,
“Incorporating Intelligence in Fog Computing for Big Data Analysis
in Smart Cities,” IEEE Transactions on Industrial Informatics, vol. PP,
no. 99, pp. 1–11, 2017.
[7] A. Al-Fuqaha, A. Khreishah, M. Guizani, A. Rayes, and M. Moham-
madi, “Toward better horizontal integration among IoT services,” IEEE
Communications Magazine, vol. 53, no. 9, pp. 72–79, 2015.
[8] S. Han, J. Pool, J. Tran, and W. Dally, “Learning both weights and con-
Processing Systems, 2015, pp. 1135–1143.
[9] V. Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G.
Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski
et al., “Human-level control through deep reinforcement learning,”
Nature, vol. 518, no. 7540, pp. 529–533, 2015.
[10] D. P. Kingma, S. Mohamed, D. J. Rezende, and M. Welling, “Semi-
supervised learning with deep generative models,” in Advances in Neural
Information Processing Systems, 2014, pp. 3581–3589.
[11] M. Mohammadi, A. Al-Fuqaha, M. Guizani, and J. S. Oh, “Semi-
supervised Deep Reinforcement Learning in Support of IoT and Smart
City Services,” IEEE Internet of Things Journal, vol. PP, no. 99, pp.
1–12, 2017.
[12] T. Boyle, D. Giurco, P. Mukheibir, A. Liu, C. Moy, S. White, and
R. Stewart, “Intelligent metering for urban water: a review,” Water,
vol. 5, no. 3, pp. 1052–1081, 2013.
[13] D. A. Talbot and M. Paz-Canales, “Smart Grid Paybacks: The Chat-
tanooga Example,” The Municipal Fiber Project. Berkman Klein Center
for Internet & Society Research Publication, 2017.
[14] C. Perera, A. Zaslavsky, P. Christen, and D. Georgakopoulos, “Context
aware computing for the internet of things: A survey,” IEEE Communi-
cations Surveys & Tutorials, vol. 16, no. 1, pp. 414–454, 2014.
[15] P. Casale, M. Altini, and O. Amft, “Transfer Learning in Body Sensor
Networks Using Ensembles of Randomized Trees,” IEEE Internet of
Things Journal, vol. 2, no. 1, pp. 33–40, 2015.