Page 9 - Unlocking the potential of trust-based AI for city science and smarter cities
P. 9
data. Moreover, the highly dynamic nature of smart cities calls for a new generation of approaches
that are flexible and adaptable to cope with the dynamicity of data to perform analytics and learn
from real-time data. Development of smart city applications supported by big data analytics is
subject to several challenges that need to be addressed to achieve a reliable and accurate system.
Some of the major challenges include the following.
Integrating Big and Fast/Streaming Data Analytics:
In a smart city context, there are many time-sensitive applications (e.g., smart and connected
vehicles) that need real-time or near-real-time analytics of the stream of data. Such applications call
for new analytic frameworks that support big data analytics in conjunction with fast/streaming data
analytics.
Preserving Trust, Security and Privacy:
Data-driven approaches (e.g., deep learning) can be attacked by false data injection (FDI), which
compromises the validity and trustworthiness of the system. Resilience against such attacks is a
must for such inference algorithms. In general, entities must be capable of building up an opinion
about every other device/service they interact with and eventually more authoritative and reliable
communication can be built up with the same pair of hosts. Privacy preservation is another
important factor since a large part of smart city data comes from individuals who may not prefer
their data to be publicly available. Data modelling algorithms should address these concerns to
enable the wide acceptance of smart city systems by organizations and citizens.
On-Device Intelligence:
Smart city applications also call for lightweight AI algorithms deployable on resource constrained
devices for hard real-time intelligence. This is also in line with the trust, security and privacy
preservation requirement since data is not transferred to the fog or cloud.
Big Dataset Shortage:
Development and evaluation of smart city applications need real-world datasets, which are not
readily available for many application domains. It is necessary to confirm results based on simulated
big data.
Context Awareness:
Integrating contextual information with raw data is crucial to get more value from the data, and
perform faster and more accurate reasoning and actuation. 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.
Unlocking the potential of trust-based AI for city science and smarter cities - October 2019 3