Expert Systems Expert Systems Chapter 6 Introduction to Artificial Intelligence MEngg

Document Sample
Expert Systems Expert Systems Chapter 6 Introduction to Artificial Intelligence MEngg Powered By Docstoc
					    Expert Systems
             Chapter 6
Introduction to Artificial Intelligence
    MEngg (Computer Systems)
    Reference Book
   Expert Systems and
Applied Artificial Intelligence

      By: Efraim Turban
The Case of CATS – 1 (DELTA)
   The Problem
   Traditional Solution
   Expert System
      Basic Concepts of
       Expert Systems
Expertise
» Facts about problem Area
» Theories
» Hard-and-fast Rules and Procedures
» Heuristic Rules
» Global Strategies
» Meta Knowledge
      Basic Concepts of
       Expert Systems
Experts (Human Experts)
» Recognizing & Formulating the problem
» Solving the problem quickly & properly
» Explaining the solution
» Learning from Experience
» Breaking rules
» Determining Relevance
» Degrading Gracefully
        Basic Concepts of
         Expert Systems
Transferring Expertise
» Expert -> Computer -> Other humans
Inferencing
» Ability to Reason
Rules
Explanation Capability
» Explanation
» Justification
» What?, How?, Why?
             Expert System

  A computer implemented expert system can
work continually (24hrs a day) can be duplicated
(thus creating many experts), never dies (taking
knowledge with it), learns indefinitely (so long as
new information is added to the system), always
  operates at peak performance, and does not
    suffer form personality incompatibilities.
   A computer expert system, however, is NOT
 INTELLIGENT. They may appear to be working
     intelligently, but that is because they are
   programmed to emulate human intelligence.
Expert System – Definition

 An expert system is a computer system
 which emulates the decision-making ability
            of a human expert.

  A model and associated procedure that
     exhibits, within a specific domain, a
 degree of expertise in problem solving that
  is comparable to that of a human expert.
     Target Problem Areas
Problems that are normally tackled by a
human “Expert”. Real experts in the problem
domain are asked to provide “rules of thumb”.
Some Target Areas
»   Diagnosis
»   Repair
»   Instruction
»   Interpretation
»   Prediction
»   Design and Planning
»   Monitoring and Control
Benefits of Expert Systems
Increased Output & Productivity
Permanence
Reduced downtime and Increased Efficiency
Consistency
Capture of scarce expertise
Reliability
Accessibility to knowledge & help desks
Documentation
Completeness
  Disadvantages of Expert
         Systems
Common sense
Creativity
Learning
Sensory Experience
Expert Systems vs. Traditional
  Problem Solving Systems
 Difference is in the way in which the problem
 related expertise is coded
 In traditional applications:
 » problem expertise is encoded in both program and
   data structures
 » algorithm + data structures = program
 In the expert system approach:
 » All of the problem related expertise is encoded in
   data structures only. None is in programs.
 » inference engine + knowledge = expert system
  Database Systems vs. Expert
           Systems
• Cannot draw conclusions        • Contains expertise and hence
                                 can emulate human reasoning
• User has to draw own
conclusions

                                 Declarative
Declarative                      Knowledge            Conclusions
Knowledge                Facts     AND                based on
                                 Procedural           Reasoning
                                 Knowledge
Conventional Systems Vs Expert Systems
No Explanation Capability     Explanation is a part of ES
Difficult to Change           Easy to Change
Step-by-Step Execution        Execution using Heuristics & Logic

Complete Information needed   Partial/Uncertain Information Sufficient

Data & Databases Involved     Knowledge & Knowledge bases Involved

Quantitative Data             Qualitative Data
        Structure of Expert Systems
    Consultation Environment                Development Environment
              Facts about
              the Specific                    Knowledge Base
 User                                     Facts: What is known about the
              Incident
                         Explanation             Domain Area
                           Facility       Rules: Logical Reference
  User
Interface
                                          • Interpreter
                 Inference Engine         • Scheduler          Knowledge
                 Draws Conclusions        • Consistency         Engineer
                                             Enforcer
                                                                  Knowledge
Recommended
  Action                                                          Acquisition
                 Blackboard (Workspace)                          Expert
                 Plan     Agenda                               Knowledge
                 Solution Problem                 Knowledge
                          Description
                                                  Refinement
Individuals Involved with
     Expert Systems
The End User
The Problem Domain Expert
The Knowledge Engineer
Other Participants
» System Builder
» Tool Builder
» Vendors
» Support Staff
      Individuals Involved with
           Expert Systems
                                                          Support
Tool Builder                    Expert                     Staff
                 Documented                    Tests          Supports
      Builds     Knowledge                                    Tasks

 Tools,                       Knowledge     Builds     Expert System
Languages         Uses         Engineer
                   Uses
      Provides            Cooperate       Builds              Uses
                                               Connects
                               System                     End-User
 Vendor                        Builder
             The End User
Sees an expert system through an interactive dialog
Sample Interaction:
Q. Do you know to which restaurant you want to go?
A. No
Q. Is there any kind of food you would particularly like?
A. Unknown
                      I am trying to determine the type of
Q. Do you like spicy food?
                  restaurant to suggest. So far Chinese is not
A. No
                  a likely choice. It is possible that French is
Q. Do you usually drink wine with meals?
                  a likely choice. I know that if the diner is a
A. Yes
                    wine drinker, and the preferred wine is
                   French, then there wine?
Q. When you drink wine, is it Frenchis strong evidence that
A. Why                the restaurant choice should include
                                     French.
   The Knowledge Engineer
The knowledge engineer is concerned with the
     representation chosen for the expert's
knowledge declarations and with the inference
   engine used to process that knowledge.
Sample Expert Systems
     Expert System to identify
        Norwegian coins
The first step is to identify the variables:
 SIZE
 COLOUR
 DECORATION
Then assign a range of values for each variable:
     SIZE
          diameter is >25mm diameter is <25mm
     COLOUR
          silver               bronze
     DECORATION
          head                 ship
          lion                 crown
         Expert System to identify
            Norwegian coins
   Then rules are constructed that identify the coins by
combinations of attributes

IF SIZE > 25 and COLOUR is bronze and DECORATION
is ship THEN coin is 20K

IF SIZE < 25 and COLOUR is silver and DECORATION is
crown THEN coin is 1K

IF SIZE >25 and COLOUR is silver and DECORATION is
lion THEN coin is 5K
         Whale Watcher

The Whale Watcher will ask you some
questions as it tries to identify the whale
that you have observed.

http://www.aiinc.ca/demos/dlwhale.shtml

           Run Demonstration
            MYCIN
MYCIN was an Expert System developed
at Stanford in the 1970s
To diagnose and recommend treatment
for certain blood infections
Written in LISP
MYCIN is a (primarily) goal-directed
system, using the basic backward
chaining reasoning strategy.
Organization of MYCIN
       System
     How MYCIN works?
MYCIN has a four stage task:
» decide which organisms, if any, are
  causing significant disease.
» determine the likely identity of the
  significant organisms.
» decide which drugs are potentially useful.
» select the best drug, or set of drugs.
The control strategy for doing this is
coded as meta-knowledge.
                      MYCIN
  MYCIN's rules:
IF the infection is pimary-bacteremia
AND the site of the culture is one of the sterile sites
AND the suspected portal of entry is the
 gastrointestinal tract
THEN there is suggestive evidence (0.7) that
 infection is bacteroid


  Sample Interaction with MYCIN
Graduate Admission Screening System
  Scenario 1:
   » Questionnaire
   » Response


  Scenario 2:
   » Questionnaire
   » Response
  How do People Reason?
They create categories
They use specific rules, a priori rules
They Use Heuristics --- "rules of thumb"
They use past experience --- "cases"
They use "Expectations"
How do Computers Reason?
Computer models are based on our
models of human reasoning
» Frames
» They use rules A--->B--->C
» They use cases
» They use pattern recognition/expectations
 Types of Expert Systems
Rule-based Systems
Frame-based Systems
Hybrid Systems
Model-based Systems
Ready-made (Turnkey) Systems
Real-Time Expert Systems
Knowledge Acquisition and
       Validation
Knowledge Engineering
  Knowledge Engineering
Knowledge Acquisition
Knowledge Representation
Knowledge Validation
Inference
Explanation
Sources of Knowledge

Documented
Undocumented
 Levels of Knowledge

Shallow (Surface) knowledge
Deep Knowledge
 Categories of Knowledge
Declarative Knowledge
Procedural Knowledge
Semantic Knowledge
Episodic Knowledge
Meta Knowledge
  Difficulties in Knowledge
          Acquisition
Problems in Transferring Knowledge
» Expressing the knowledge
» Transfer to a machine
» Number of participants
» Structuring the knowledge
» Non-cooperative Experts
» Noisy input
Required Skill of Knowledge
        Engineers
Effective Communication Abilities
Computer Skills
Fast Learning capabilities
Advanced, socially sophisticated verbal
skills
Tolerance and Patience
Logical Thinking
Versatility and Inventiveness
Methods of Knowledge Acquisition

 Manual
 » Interviews
   » Unstructured
   » Structured
 » Tracking Methods
   » Protocol Analysis
   » Observations
 » Other
   » Case Analysis
   » Prototyping
   » Etc…
Methods of Knowledge Acquisition

 Semi Automatic
 » Support the Expert
   » Repertory Grid Analysis
      – ETS
      – KRITON
      – AQUINAS
 » Support the Knowledge Engineer
   » Editors
   » Explanation
   » Documentation
   » Front-end Tools
Methods of Knowledge Acquisition

 Automatic
 » Rule Induction
 » Machine Learning

				
DOCUMENT INFO