Revise

advertisement
International Higher Diploma in Computer Studies
Knowledge Based System
The marks given in brackets are indicative of the weight given to each part of the question.
Answer only FOUR questions out of SIX
Time: TWO hours and 10 minutes reading time
Reference materials are NOT allowed.
Question 1
1. a)
With the use of diagram, explain the structure of Expert Systems.
[7 Marks]
It has three common rings, the central shows the knowledge base of facts and inference rules. This
extends the facilities of a conventional database by enabling the storage not only of facts but also of
rules that enable further facts to be derived. The outer ring shows the three classes of operator of
expert systems: Experts, Users and Knowledge Engineers. Lastly, the middle ring shows the expert
system shell that provides computational facilities for applying the knowledge base to user decision
support.
2. b)







List different types of KBS Application.
[8 Marks]
Interpretation - take observations and infer descriptions e.g. natural language understanding
Prediction - recognise situations and infer likely consequences e.g. weather forecasting
Diagnosis - observe symptoms and infer causes of those symptoms e.g. medicine, mechanical
design - given a set of constraints develop configurations which satisfy those constraints e.g.
computer system design
Planning - specifying actions e.g. robot movement, project planning etc.
Monitoring - compare current observations with expected observations and both indicate
discrepancies and suggest corrective action e.g. patient monitoring
Instruction - assist the learning process e.g. recognise errors in student programs and suggest
alternatives
Control - adaptively govern the overall control of various control systems e.g. power plants,
chemical plants.
IHDCS/KBS/0112
Page 1 of 9
3. c)
Name some well known Classical KBS.
[5 Marks]
MYCIN
This is a medical diagnostic system developed at Stanford University, USA. It assists
internists in the diagnosis and treatment of some infectious diseases
DENDRAL
It is also developed at Stanford University, USA. It aids organic chemists to determine the structural
formula of a chemical compound in a given sample.
XCON/R1
It is developed at DEC in conjunction with Carnegie-Mellon (CMU) University, USA. It configures
VAX computers - wiring, connectors, cables, back planes, racks etc for each customer's order.
HEARSAY II
It is a speech recognition and understanding system developed at CMU. It is able to recognize and
understand speech formed from a 1000 word vocabulary. It is a milestone in the development of
distributed AI.
4. d)
List three weak points in KBS Systems technology.
[5 Marks]
1. Lack of a theory of the nature of knowledge
This can partly be attributed to the phenomenon that current methodologies and tools lack
an underlying theory that defines the nature of knowledge.
2. Lack of an adequate representation formalism
A second weak point in current KBS-technology can be spotted at the level of knowledge
representation languages. Business practice reveals that industrial clients impose high
demands on knowledge representation formalisms.
3. Lack of adequate tools
A third weak point is the lack of tools that provide advanced management facilities centred
around a representation formalism that can cope with all the requirements above
IHDCS/KBS/0112
Page 2 of 9
Question 2
5. a)
How do you understand by the term ‘ontology’?
[5 Marks]
Ontology provides an explicit conceptualization (i.e., meta information) that describes the semantics
of the data. They have a function similar to a database schema. Some differences are:
A language for defining ontologies is syntactically and semantically richer than common approaches
for databases.
An ontology must be a shared and consensual terminology because it is used for information sharing
and exchange.
An ontology provides a domain theory and not the structure of a data container.
6. b)
Describe the term ‘rules’. Discuss the advantages and disadvantages of using
‘rules’ as a knowledge representation method.
[12 Marks]
Currently, the most popular method of knowledge representation is in the form of rules (also known
as production rules or rule-based systems).
Advantages
Additions and deletions of rules can be made easily Care must be taken when adding or deleting
rules because the logic of the decision-making has now been potentially changed.
ModularityEach rule in the rule base stands apart from the other rules.
It has uniform structureAll rules in a rule base have the same form.
It provides a natural mode of knowledge representation The time required to learn how to
develop rules bases (knowledge bases that contain rules) can be kept to a minimum. In addition, many
experts solve problems based on combining pieces of evidence (known facts) and the combination of
those facts lead to other “newly inferred” facts (i.e., the consequent).
There exist many E/KBS development packages—known as shells—which use rules as the primary
method of knowledge representation.
Disadvantages
A knowledge base of rules can quickly become unwieldy and unmanageable if not properly
implemented.
Rules can be hard to maintain for the same reasons previously stated.
Rules can be inefficient in processing because the inference engine is performing a search over the
rules to find the rules that could be fired given the current state of knowledge in the system.
Rules cannot handle all types of knowledge.
7. c)
Knowledge can be represented in the following ways. Define each of them:
i. Declarative
Knowledge is stored as a set of statements about the world. These statements are static but can be
added to, deleted or modified.
ii.
Procedural
Knowledge is stored as a set of procedures which can themselves determine when they should be
executed. Their execution is the intelligent behavior that was expected in the situation.
iii.
Symbolic
The storage of the knowledge utilizes symbols in order to represent objects of the outside world or
sets of perceptions about the outside world.
iv.
Uniform Representation
[4 x 2 Marks]
The knowledge base chooses one method for representing the knowledge (e.g. frames, semantic nets
etc) and uses it exclusively.
IHDCS/KBS/0112
Page 3 of 9
Question 3
8. a)
Explain with a block diagram the basic components of building expert system
tools.
[8 Marks]
Basic Components of Expert System Tools
1. Knowledge base: A store of factual and heuristic knowledge.
2. Reasoning engine: Inference mechanisms for manipulating the symbolic information and
knowledge in the knowledge base to form a line of reasoning in solving a problem.
3. Knowledge acquisition subsystem: A subsystem to help experts build knowledge bases.
4. Explanation subsystem: A subsystem that explains the system's actions.
5. User interface: The means of communication with the user.
b) Compare the process of knowledge base verification, validation & evaluation.
[9 Marks]
Verification of an expert system, or any computer system for that matter, is the task of
determining that the system is built according to its specifications. Issues raise includes:




Does the design reflect the requirements? Are all of the issues contained in the
requirements addressed in the design?
Does the detailed design reflect the design goals?
Does the code accurately reflect the detailed design?
Is the code correct with respect to the language syntax?
Validation is the process of determining that the system actually fulfills the purpose for which
it was intended. Issues raise include:





How well do inferences made compare with knowledge and heuristics of experts in the
field?
How well do inferences made compare with historic (known) data?
What fraction of pertinent empirical observations can be simulated by the
system?
What fraction of model predictions are empirically correct?
What fraction of the system parameters does the model attempt to mimic?
IHDCS/KBS/0112
Page 4 of 9
Evaluation reflects the acceptance of the system by the end users and its performance in
the field. In other words:



Verify to show the system is built right.
Validate to show the right system was built.
Evaluate to show the usefulness of the system.
c) Describe boundary value analysis & white-box testing
[8 Marks]
Boundary value analysis
Boundary value analysis (checking that the expert system produces the correct results when given
quantitative numerical input), which describes numbers in an allowable range, numbers exactly on the
numerical boundaries of the range, and numbers just outside of the range, is a useful method of blackbox testing to eliminate a great deal of test cases. These eliminated test cases would in the tester's
opinion not reveal unexpected output errors (if they were tested).
White-box testing
Another method which can be used instead (or perhaps even done in conjunction with black-box
testing methods, like correctness testing) is to use white-box testing. The tester examines the internal
logic of all parts of the expert system, and tries to justify the existence of each part, and to understand
the processing/reasoning involved, changing the reasoning in cases where it seems to be wrong or
inefficient.
Question 4
a) Discuss the difference between forward and backward chaining expert system.
[8 Marks]
Backward chaining inference engine is 'goal-orientated' in the sense that it tries to prove a
goal or rule conclusion by confirming the truth of all of it's premises. A chain of inference
steps will lead to a value for the goal being found. Backward chaining is best suited for
applications in which the possible conclusions are limited in number and well defined.
Forward chaining inference engine starts from the other end. It examines the current state
of the knowledge base and, finds those rules whose premises can be satisfied from known
given data, and adds the conclusions of those rules to the knowledge base. It then reexamines the complete knowledge base and repeats the process, which can now progress
further since new information has been added. Both the backward and forward inference
process will consist of a chain of steps that can be traced by the expert system. This
enables expert systems to explain their reasoning processes. Forward-chaining systems are
commonly used to solve more open-ended problems of a design or planning nature, such
as, for example, establishing the configuration of a complex product.
b) Write short notes on the following:
i. Bayesian Network
These are probabilistic models based on directed graphs capturing causal
relationships between a number of variables being modeled. These can
provide very accurate tools for predication and diagnosis. Microsoft is a big
supporter basing various diagnostic tools on the technology.
IHDCS/KBS/0112
Page 5 of 9
ii.
Machine Learning
These are techniques for deriving useful knowledge from data or examples. Data mining is a
key approach that looks for using useful patterns in databases. By using machine learning,
some of the problems of the knowledge engineering bottleneck can be obviated.
iii.
Argumentation Systems
[9 Marks]
These are models based on capturing arguments for and against inferences. For
professionals needing decision-support, seeing the inter-relationships between evidence and
viewpoints can be much more valuable than just being given a computer-generated decision.
c)
What are the limitations of an expert system?
[8 Marks]
Expert Knowledge Is Expensive.
A commitment of considerable time from knowledgeable people is essential to the development of
expert systems. Usually, the people who make the best experts are the ones most highly valued by
their own organizations and therefore in some respect the least accessible.
Expert Knowledge Has to Be Acquired Incrementally and Tested.
Expert knowledge is not acquired all at once: The process of building an expert system spans
several months and sometimes several years. In the course of this development, it is typical to expand
and reformulate the knowledge base many times.
Even so, expert systems take time to build.
Toy programs for a small demonstration can be built quickly-often in just a few months using current
technology. However, for large-scale systems with knowledge spanning a wide domain, the time
needed to develop a system that can be put in the field can be measured in years, not months, and in
tens of worker-years, not worker-months.
Question 5
a)
Certain basic conditions must be met if a Decision Support Systems is to have the
desired impact on the decision-making process of an organization. Describe some
of these conditions.
[8 Marks]
(a) the right problems must be addressed; (b) the right people must participate in the
development of the decision support system; (c) the right tools must be used; and (d) the
process must evolve as decision situations and technology change.



The DSS should be demand-driven rather than supply-generated--the demand for
decision support must come from top-level management rather than being "force-fed"
on the basis of available data.
A DSS should be capable of continual evolution, often in unanticipated ways as
problems evolve.
The DSS should be provided as a relatively stable product since users do not want to
have to learn new commands every few weeks in order to access the system.
IHDCS/KBS/0112
Page 6 of 9
b)
Explain the role of Memory Organization Packages used in Case Based
Reasoning.
[5 Marks]
MOPs are used to represent knowledge about events. An event can either be a
representation of a full event or more typically some significant portion of a complete case.
These events within a MOP are useful in partial matching, in that the required information
from a specific case can be used to further a solution state without utilizing the entire case,
which may not be a perfect match. The information contained within a MOP is organized and
classified by norms. Norms contain basic features of a MOP and are used to structure it. A
MOP containing the event `going on a trip', would have norms for planning and what is
required to plan, getting tickets and where to go to purchase them, etc. Norms are a listing or
classification of expected information or action associated with the event, stored within the
MOP.
c)
Write down six key attributes of groupware systems and discuss each briefly.
[5 Marks]
Deferred communication
At the end of the day, you could end up with a tool that enhances communication between employees,
streamlines the idea flow, and maybe even helps spark some new ideas.
Real-time communication
It supports chatting, text, video and voice communication. People can communicate in an
actual time and the response from the other side is real-time.
Workflow management
They appear to be fairly similar in functionality; allowing the user to check other people's schedules
and post his or her own.
Secure transactions
It has Security encryption and it allows client to send encrypted mail to each other
Authentication of participants
Exchange's mail client has menu options for logging in using Secure Password Authentication.
Distributed or Centralized server(s)
The server component of Communicator, Netscape Suitespot, is particularly flexible, supporting
multiple administrators and servers, all of which can be set up through the Netscape browser.
d)
Discuss the problems faced in Knowledge elicitation. State the ways to correct it?
[7 Marks]
Problems
1. Inability of an expert to express his knowledge,
2. Difficulty in transferring this knowledge (or rather the elicited part of it) into appropriate
computerised form, and
3. Difficulty in maintaining consistency, when knowledge has to be updated, due to changes in the
domain, or when new knowledge is revealed.
There a lot of ways to correct but it depends on which kind of Expert system has to be
created. It also depends on the environment from which Knowledge Engineer collects or
acquires the knowledge from a domain expert.
IHDCS/KBS/0112
Page 7 of 9
Question 6
a)
What are the benefits of Knowledge Representation Techniques in Decision
Tables?
[7 Marks]
Completeness
Knowledge bases often suffer from missing attribute values or combinations of attribute values,
unreachable conclusions, etc.
Consistency
Inconsistency occurs when rules with the same premises but different conclusions exist. When these
conclusions are contradictory, the rules are in conflict.
Non-redundancy
By definition, a single hit decision table eliminates redundant rules and premises, as a combination of
condition states will be included in only one column.
Correctness
After the decision tables have been designed, the knowledge engineer may want to check the
(semantic) correctness of the decision specification, verifying that for each possible case the
right action(s) will be executed.
b)
With the help of a diagram, describe the CBR Cycle.
[10 Marks]
Retain
Incorporate what is useful to learn from the problem solving experience into the existing knowledge
IHDCS/KBS/0112
Page 8 of 9
Retrieve
Retrieving a case means to start with a (partial) new case, and retrieve the best matching previous
case.
Reuse
The focus of reuse is to find the difference between the new and the old case, and find what part of the
old case that can be used in the new case.
Revise
If the solution generated by the last phase is not correct, the system can learn from its failures.
c)
What do you mean by “inference engine”? Differentiate its two alternative
strategies.
[8 Marks]
The inference engine:
1. Combines the facts of a specific case with the knowledge contained in the knowledge
base to come up with a recommendation. In a rule-based expert system, the inference
engine controls the order in which production rules are applied
2. Directs the user interface to query the user for any information it needs for further
inferencing.
Two Alternatives are:
Backward chaining inference engine is 'goal-orientated' in the sense that it tries to prove a
goal or rule conclusion by confirming the truth of all of it's premises. A chain of inference
steps will lead to a value for the goal being found. Backward chaining is best suited for
applications in which the possible conclusions are limited in number and well defined.
Forward chaining inference engine starts from the other end. It examines the current state
of the knowledge base and, finds those rules whose premises can be satisfied from known
given data, and adds the conclusions of those rules to the knowledge base. It then reexamines the complete knowledge base and repeats the process, which can now progress
further since new information has been added. Both the backward and forward inference
process will consist of a chain of steps that can be traced by the expert system. This
enables expert systems to explain their reasoning processes. Forward-chaining systems are
commonly used to solve more open-ended problems of a design or planning nature, such
as, for example, establishing the configuration of a complex product.
IHDCS/KBS/0112
Page 9 of 9
Download