Page 33 - Artificial Intelligence for the Internet of Everything

20 Artificial Intelligence for the Internet of Everything

In this chapter we first provide a background and motivation scenario
in Section 2.2.In Section 2.4, we discuss how to be able to quantify and
minimize the uncertainty with respect to training an ML algorithm. This
leadsustothe fieldofstochasticoptimization, which is covered broadly
in this section. In Section 2.4, we discuss UQ in ML. Specifically, we
study how to develop ways for a model to know what it doesn’t know.
In other words we study how to enable the model to be especially cau-
tious for data that is different from that on which it was trained.
Section 2.5 explores the recent emerging trends on adversarial learning,
which is a new application of UQ in ML in Internet of Battlefield Things
(IoBT) in an offensive and defensive capacity. Section 2.6 concludes the
chapter.

2.2 BACKGROUND AND MOTIVATING IOBT SCENARIO

In recent years the Internet of Things (IoT) technologies have seen signif-
icant commercial adoption. For IoT technology, a key objective is to deliver
intelligent services capable of performing both analytics and reasoning over
data streams from heterogeneous device collections. In commercial settings
IoT data processing has commonly been handled through cloud-based ser-
vices, managed through centralized servers and high-reliability network
infrastructures.
Recent advances in IoT technology have motivated the defense com-
munity to research IoT architecture development for tactical environments,
advancing the development of the IoBT for use in C4ISR applications
(Kott, Swami, & West, 2016). Toward advancing IoBT adoption, differ-
ences in military versus commercial network infrastructures become an
important consideration. For many commercial IoT architectures, cloud-
based services are used to perform needed data processing, which rely upon
both stable network coverage and connectivity. As observed in Zheng and
Carter (2015), IoT adoption in the tactical environment faces several tech-
nical challenges: (i) limitations on tactical network connectivity and reliabil-
ity, which impact the amount of data that can be obtained from IoT sensor
collections in real time; (ii) limitations on interoperability between IoT
infrastructure components, resulting in reduced infrastructure functionality;
and (iii) availability of data analytics components accessible over tactical
network connections, capable of real-time data ingest over potentially sparse
IoT data collections.
Challenges such as these limit the viability of cloud-based service usage in
IoBT infrastructures. Hence, significant changes to existing commercial IoT