Page 150 - DCAP310_INTRODUCTION_TO_ARTIFICIAL_INTELLIGENCE_AND_EXPERT_SYSTEMS

Introduction to Artificial Intelligence & Expert Systems

Notes true uncertainty or indeterminacy. Risk assessment has become a central guiding principle at
the U.S. Environmental Protection Agency (EPA) and other environmental management agencies,
but true uncertainty has yet to be adequately incorporated into environmental protection
strategies. Scientists treat uncertainty as a given, a characteristic of all information that must be
honestly acknowledged and communicated.
Over the years, scientists have developed increasingly sophisticated methods to measure and
communicate uncertainty arising from various causes. In general, however, scientists have
uncovered more uncertainty rather than the absolute precision that the lay public often mistakenly
associates with scientific results. Scientific inquiry can only set boundaries on the limits of
knowledge. It can define the edges of the envelope of known possibilities, but often the envelope
is very large, and the probabilities of what’s inside (the known possibilities) actually occurring
can be a complete mystery. For instance, scientists can describe the range of uncertainty about
global warming and toxic chemicals and maybe say something about the relative probabilities
of different outcomes, but, in most important cases, they cannot say which of the possible
outcomes will occur with any degree of accuracy.

!
Caution Deduct uncertain information for knowledge.

Task List five cases of uncertainty of information.

Self Assessment

State whether the following statements are true or false:
13. Cognitive maps are external representations of our physical environment, particularly
associated with spacial relationships.

14. Heuristics were also found to be used in the manipulation and creation of cognitive maps.
15. Visual intake is a key part of mapping.
16. Heuristic refers to experience-based techniques for problem solving.

7.5 Summary

The reasoning supporting these networks, based on two simplifying assumptions (that
reasoning could not be cyclic and that the causality supporting a child state would be
expressed in the links between it and its parent states) made BBN reasoning quite
manageable computationally.
Bayesian belief networks (BBNs) can dramatically reduce the number of parameters of the
full Bayesian model and show how the data of a domain (or even the absence of data) can
partition and focus reasoning.

Bayes’s theorem requires masses of statistical data in addition to the degrees of belief in
preconditions.
D – S assigns the part of the probability to the entire universe to make the aggregate of all
events to 1. This part is called ignorance level.
Dumpster – Shafer theory allows updating of beliefs based on evidence gathered.

144 LOVELY PROFESSIONAL UNIVERSITY