Page 10 - Unlocking the potential of trust-based AI for city science and smarter cities
P. 10
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.
The ever-growing volume of data and devices in a smart city poses open problems for intelligent
services, trust and privacy. Inside-attackers exploit human intelligence and have access to big
data such that the privacy of data owners may be inferred and violated; even the traditional
cryptographic schemes have been applied to big data.
An alternative to detect these inside attackers is to enhance the traceability and allow a trusted
third party to monitor and audit. Meanwhile, collaborative efforts among municipalities, regulation
departments, industry, academia, and business companies are necessary to set up privacy policies
and regulations.
In addition, to improve the data privacy, availability, and management of the city network, a
distributed computing architecture which delegates AI based processing of data towards the
edge of the network must be considered. Further a smart city is vulnerable to false data injection
in both sensing and control phases. Digital signature techniques cannot prevent the data from
being tampered from the origination. An insight into detecting false data injection is to leverage
machine learning and data mining along with trust-based concepts to come up with a boundary of
reasonable sensing data.
The proposed approach intends to instill a trusted environment for various City Science applications
in the smart city context. It proposes a distributed computing architecture which is conducive to
enhancing trust while enabling innovation for City Science applications.
Important Note: This case study is an example of an R&D project related to city science, rather
than an actual city example. City Science is a relatively novel field and will require substantial R&D
(Research & Development) for developing future urban solutions. The proposed approach is an
actual research project currently being conducted by the author.
2. Trust-based AI Data Management Solution
2.1 Vision and content
The proliferation of computing, networked systems and end-node processing power, has made
Internet a highly dynamic system. Maintaining trust across a large-scale heterogenous distributed
system is a formidable task. It requires preservation of data processing security policies in a
distributed system which can be substantially challenging. Existing security mechanisms (e.g.
authentication, authorization) are not sufficiently scalable for today’s large-scale networks. Hence,
the trust-based approach to distributed systems is developed to address the inadequacy of
traditional mechanisms.
4 Unlocking the potential of trust-based AI for city science and smarter cities - October 2019