Trustworthiness Model Representation
a directed tree. The root node represents
the system, and the intermediate nodes in
the tree represent the constituent
components of the system. In the example,
the system is comprised of components 1
through m. Each component has a series of
individual observed attributes (attributes 1
through j in the case of component 1) that
are used to compute the Trust Score of the
component. The observed attribute 1 is
weighted by the factor w1, and so on,
through attribute j which is weighted by the
factor wj to compute the Trust Score of
component 1. The computed component
Trust Scores are then fed into the overall
function to compute the Trust Score of the
overall system, using the component
weightings w1 through wm.
each component or attribute influences
trustworthiness. Computing a Trust Score of
each component helps identify factors
affecting the overall Trust Score. The
directed acyclic graph allows components to
be grouped based on the high level
categories such as safety, security, privacy,
reliability, and resilience.
Benefits of the Trust Model
The benefits of a Trust Model constructed in
this manner are that it flexibly addresses the
goals of a variety of solutions. The Trust
Model supports basic operations such as
fusion and discounting as specified in the
Wikipedia definition of a trust metric. 11 Its
output is simple -- the component Trust
Scores and the overall Trust Score provide a
measure that is straightforward to interpret
and enables the user to readily take action to
address the root cause of issues that
diminish trust.
Organizing components and attributes of a
system or solution in a directed acyclic graph
enables the user to control and filter how
Figure 4: Computation of a Trust Score
11
Wikipedia, https://en.wikipedia.org/wiki/Trust_metric, 2018.
- 35 -
IIC Journal of Innovation