UNIT-II

Expert System: Components of Expert System: Knowledge Base, Inference Engine, User
Interface, Features of Expert System, Expert System Life Cycle, Categories of Expert System,
Rule Based vs. Model Based Expert Systems, Advantages/Limitations of Expert System,
Developing an Expert System: Identification, Conceptualization, Formalization,
Implementation, Testing, Using an Expert System, Application Areas of Expert System

Expert Systems are interactive, trustworthy computer-based decision-making tools that utilize
data and heuristics to address challenging decision-making issues. It is regarded as
representing the pinnacle of human knowledge and wisdom. An expert system's job is to
resolve the trickiest problems in a specific field.

An expert system (ES) is a knowledge-based system that employs knowledge about its
application domain and uses an inferencing (reason) procedure to solve problems that would
otherwise require human competence or expertise. The power of expert systems stems
primarily from the specific knowledge about a narrow domain stored in the expert
system's knowledge base.

Expert systems do not have human capabilities. They use a knowledge base of a particular
domain and bring that knowledge to bear on the facts of the particular situation at hand. The
knowledge base of an ES also contains heuristic knowledge - rules of thumb used by human
experts who work in the domain.

Components of Expert System

There are five components of the expert system in ai:

Components of an expert system

The expert system in artificial intelligence consists of the following components:

 User interface

 Inference engine

 Knowledge base

 Knowledge base acquisition facility

 Explanation facility

The architecture of an expert system

User interface (UI)

Users interact with interactive UI. The UI will take all the parameters and pass the
information to the inference engine. After computing the result from the user's input, UI then
shows the result on the screen.

Inference engine

The mind of the expert system has all the predefined rules to use the information from the
knowledge base. It takes user input queries from the user interface and applies knowledge
facts and rules. It can also deduce the problem to find a solution. Two strategies are used to
find the answer to the problem:

 Forward chaining: It uses rules of inference and data to deduce more rules until it
comes to a conclusion. And adds new information to its database for future problems.
 It moves its reasoning toward the goal. Its approach can be called data-driven
reasoning.

 Backward chaining: It works its way back from the goal. So basically, it does the
backtracking from the goal state to the present state. It is goal-driven reasoning. As
the goal is already known, it becomes easy to use rules of inference to find the
solution.

Knowledge base

It is the database where experts' knowledge is stored. Knowledge engineers gather knowledge
from the experts and keep it in the knowledge base. Knowledge can be heuristic or factual:

 Heuristic knowledge is disordered information in nature.

 Factual knowledge is based on facts that are proven and widely accepted.

Knowledge acquisition and learning module: With the help of this component, expert

systems can gather more information from numerous sources. After that, the knowledge is

stored in the knowledge base.

Explanation base

It is the optional component that shows the user the explanation of the solution. It shows the
reason why it suggests the proposed solution.

Characteristics of Expert System

Follow are the characteristics of an expert system.

 A human expert can change, but an expert system can last forever.
 It facilitates the distribution of human expertise.
 The expert system might incorporate knowledge from multiple human experts, which
would increase the effectiveness of the answers.
 It lowers the expense of seeking advice from a specialist in various fields, including
medical diagnosis.
 Instead of using standard procedural code, expert systems can handle complex issues by
inferring new facts from known facts of knowledge, which are typically represented as if-
then rules.

Examples of Expert Systems

 The following are some examples of expert systems:

 MYCIN: It could recognize different bacteria that might cause acute infections and was
based on backward chaining. Additionally, it might suggest medications based on the
weight of the patient. It is among the top examples of an expert system.
 DENDRAL: A molecular structure prediction tool for chemical analysis.
 CaDet: It's one of the best examples of an expert system that can detect cancer in its
earliest stages.
 PXDES: The kind and stage of lung cancer are identified using the PXDES expert system.
It takes a photo of the upper body, which resembles the shadow, to identify the condition.
This shadow determines the kind and severity.
Phases in Building Expert Systems
Building an ES initially requires extracting the relevant knowledge from a human domain
expert; this knowledge is often based on useful thumb rules and experience rather than
absolute certainties. Usually experts find it difficult to express concretely the knowledge and
rules used by them while solving a problem. This is because their knowledge is almost
subconscious or appears so obvious that they do not bother mentioning it.
After extracting knowledge from domain experts, the next step is to represent this
knowledge in the system. Representation of knowledge in a computer is not straight
forward and requires special expertise. A knowledge engineer handles the responsibility of
extracting this knowledge and building the ES’s knowledge base.
This process of gathering knowledge from a domain expert and codifying it according to the
formalism is called knowledge engineering. This phase is known as knowledge acquisition,
which is a big area of research. A wide variety of techniques have been developed for this
purpose.
Generally, an initial prototype based on the information extracted by interviewing the expert
is developed. This prototype is then iteratively refined on the basis of the feedback received
from the experts and potential users of the ES.
Refinement of a system is possible only if the system is scalable and modifiable and does not
require rewriting of major code. The developed system should be able to explain its reasoning
to its users and answer questions about the solution process. Moreover, updating the system
should just involve adding or deleting localized regions of knowledge [Waterman 1956).
A simple ES primarily consists of a knowledge base and an inference engine, while features
such as reasoning with uncertainty and explanation of the line of reasoning enhance the
capabilities of ES. Since an ES uses uncertain or heuristic knowledge just like humans, its
credibility is often questionable. In real-life situations whenever we obtain an answer to a
problem which seems questionable, we want to know the rationale behind it, and we believe
the answer only if the rationale seems plausible to us. Similar is the case with expert systems;
most expert systems han the ability to answer questions of the form 'Why the answer X?
explanations can be tracing the line of reasoning used by the inference engine.
To be more precise, the different interdependent and overlapping phases involved in building
ES may be categorized as follows:
  Identification Phase In this phase, the knowledge engineer determines important
features of the problem with the help of the human domain expert. The parameters
that are determined in this phase include the type and scope of the problem, the kind
of resources required, and the goal and objective of the Expert System.

 Conceptualization Phase :In this phase, knowledge engineer and domain expert
decide the concepts, relations, and control mechanism needed to describe the
problem-solving method. At this stage, the issue of granularity is also addressed,
which refers to the level of details required in the knowledge.

 Formalization Phase :This phase involves expressing the key concepts and relations
in some framework supported by Expert system building tools. Formalized
knowledge consists of data structures, inference rules, control strategies, and
languages required for implementation.

 Implementation Phase : During this phase, formalized knowledge is converted to a

working computer program, initially called prototype of the whole system.

 Testing Phase : This phase involves evaluating the performance and utility of
prototype system and revising the system, if required. The domain expert evaluates
the prototype system and provides feedback, which helps the knowledge engineer to
revise it.

Benefits of Expert Systems

Using expert systems instead of human experts has several advantages:

 Accuracy: Expert systems are immune to emotional or human inaccuracy. They base their
choices on facts and rules.
 Permanent: When human specialists leave their positions, the technical information could
follow. Knowledge-based systems offer an everlasting reservoir of information and
knowledge.
 Logical deduction: Expert systems use a variety of principles, such as if-then rules, to
derive conclusions from data already known.
 Cost-control: When compared to the cost of hiring human specialists, expert systems are
comparatively cheap. They can assist in making judgments more quickly and
inexpensively.
 Several experts: The knowledge base of an expert system is augmented by multiple
experts. This gives more information to draw upon and stops one expert from influencing
the decision-making process.
Expert System Applications
 It has a wide range of applications in designing and manufacturing tangible objects,
including designing and producing automobiles and camera lenses.
 These systems are typically used in the knowledge domain to disseminate pertinent
information to users. Tax advisors and advisors are the two standard ES used for this
subject.
 It is used in the financial sector to identify any potential fraud and suspicious conduct.
 Expert system in ai is used in the evaluation and repair of equipment. The ES system is
utilized in medical diagnosis and was the first application of these systems.

Conventional System vs. Expert System

The crucial difference between expert systems and conventional problem-solving software is
the method used to encode the problem-related expertise. In traditional applications, data
structures and programs are both used to encode problem expertise. All problem-related
expertise is represented in data structures alone, not in programs, in the expert method.

Conventional systems are unable to explain a specific solution to a problem. These systems
aim to provide straightforward solutions. However, expert systems can justify why certain
information is required during a process and how a specific result was achieved.

Typically, an expert system executes its inference by symbolic computations using the
symbols for knowledge, such as the rules, various forms of networks, frames, scripts, etc.
Nevertheless, these terms cannot be expressed by traditional systems. They oversimplify the
issues and cannot articulate the "how, why" questions.

Human Expert vs. Expert System

The significant distinction between expert systems in artificial intelligence and human
experts is that expert systems process knowledge represented in the form of rules and use
representational reasoning in a limited area, whereas human experts use knowledge in the
form of heuristics of rules of thumb to solve problems in a limited domain.

Expert systems are always accessible, unlike human experts, who are only available during
certain hours of the day. Expert systems handle any problem in a concise amount of time, but
human experts can take their time.