Transcript KBS - teachmath1729

AI & KBS
AI & KBS
Overview
* A brief history of Artificial Intelligence (AI)
- requirements of KBS
* Introduction to Knowledge-Based Systems (KBS)
- definition
- architecture
- development tools
* Examples of some KBS
* Characteristics of KBS
AI & KBS
1. The Evolution of AI
(1) The request of an intelligent machine
- early robots ---- electro-mechanical devices
- later robots ---- use computers, ‘dumb’
limited and pre-specified tasks
- ideal robots ---- intelligent machine combines the memory
accuracy, and speed of computers with the
intelligence and flexibility of humans.
AI - Computer programs that undertake tasks that, if done
by people would be described as requiring intelligence.
* Can computers think?
- Turing Test
Imitation game - machine mistaken for a human
“Computing machinery and intelligence”, Alan Turing, 1950
Wall
AI & KBS
In the Turing test a human communicates with an unseen respondent through
a terminal, not knowing if the respondent is a person or a machine. If the tester
mistakes computer answers for human answers, the computer successfully
passes the Turning test.
AI & KBS
ELIZA
• Developed by MIT 1966
• Called ELIZA after Shaws play - it could be taught to speak
increasing well
• It picked up words from its conversational partner
• Transform this into a canned response
ELIZA’s opening statement (appears on the terminal screen):
Do you have any problems?
Human: Yes. I am unhappy. (types response on the terminal screen)
ELIZA: Why are you unhappy?
Human: My friend is mean to me.
ELIZA: Tell me about your friend. (Rogerian Psychoanalyst)
It could fool people into thinking it was a real person but it contained
no intelligence.
AI & KBS
(2) Game playing - early AI emphasis
- Board games: chess, checkers, & 16-puzzle
- No ambiguity in representation of the board
configuration
- Rules generate large search space: require heuristics
Move
1-X
2-O
3-X
4-O
5-X
Tic-Tac-Toe game
AI & KBS
(3) Theorem proving
- The proving of mathematical theorems by a
computer program
- Theorems automatically proven from a given
set of axioms
- Theorems & axioms expressed in logic and
logical inferences applied
- First theorem prover developed in mid-50s but
breakthrough in 1960s
- Breakthrough came after introduction of
Resolution inference rule
AI & KBS
Theorem proving -Resolution
All Irish are Europeans.
Dave is a Irish.
Therefore, Dave is a European
AI & KBS
(4) Problem solving
- GPS (General Problem Solver)
focus on systems with general capability for solving
different types of problems
- Problem represented in terms of initial state,
wished-for final state (goal) and a set of legal
transitions to transfer states into new states
- Using states & operators, GPS generates sequence of
transitions that transform initial state into final state
AI & KBS
- Problems with GPS:
* efficiency in choosing path to reach the goal
* GPS did not use specific info about problem at hand
in selection of state transition
* GPS examined all states leading to exponential time
complexity
* breakthrough in AI towards more specialised
problem-solving system, i.e.,
Knowledge-based systems
AI & KBS
(5) Other AI fields - a tree representation
AI & KBS
(6) KBS as real-world problem solvers
- Problem-solving power does not lie with smart reasoning
techniques nor clever search algorithms but
domain dependent real-world knowledge
- Real-world problems do not have well-defined
solutions
- Expertise not laid down in algorithms but are domain
dependent rules-of-thumb or heuristics (cause-and-effect)
- KBS allow this knowledge to be represented in
computer & solution explained
AI & KBS
2. Knowledge-based Systems: A definition
- A system that draws upon the knowledge of
human experts captured in a knowledge-base to solve
problems that normally require human expertise.
- Heuristic rather than algorithmic
- Heuristics in search vs. in KBS
general vs. domain-specific
- Highly specific domain knowledge
- Knowledge is separated from how it is used
KBS = knowledge-base + inference engine
AI & KBS
3. KBS Architecture
Facts
Heuristics, etc.
Explanation
Queries
End-user
interface
Conclusions
Expertise
Recommendations
for action
Inference
engine
Knowledge
-base
Knowledgerepresentation
schema
AI & KBS
(1) Knowledge-base
Hypothesis
Heuristics
Rules
Objects
Facts
Knowledgebase
Processes
Attributes
Events
Definitions
Relationships
AI & KBS
(2) Knowledge representation formalisms
& Inference
KR
* Logic
* Production rules
* Semantic nets &
Frames
* Case-based
Reasoning
Inference
Resolution principle
backward (top-down, goal directed)
forward (bottom-up, data-driven)
Inheritance & advanced reasoning
Similarity based
AI & KBS
(3) KBS tools - Shells
- Consist of KA Tool, Database &
Development Interface
- Inductive Shells
- simplest
- example cases represented as matrix of known data
(premises) and resulting effects (conclusions)
- matrix converted into decision tree or IF-THEN statements
- examples selected for the tool
- Rule-based shells
- simple to complex
- IF-THEN rules
AI & KBS
- Hybrid shells
- sophisticate & powerful
- support multiple KR paradigms & reasoning schemes
- generic tool applicable to a wide range
- Special purpose shells
- specifically designed for particular types of problems
- restricted to specialised problems
-Scratch
- require more time and effort
- no constraints like shells
- shells should be investigated first
AI & KBS
4. Some example KBSs
(1) DENDRAL (chemical)
- Pioneering work developed in 1965 for NASA at
Stanford University by Buchanan & Feigenbaum
- DENDRAL infers the molecular structure given mass
spectral data
- Traditional method of generate-and-test, difficulty:
millions of possible structures might be generated
to account for data
- Experts used rules-of-thumb to weed-out structures
that are unlikely to account for the data
- Encoded this expertise & produced program which
performed as well as an expert chemist
AI & KBS
(2) MYCIN (medicine)
- Developed in 1970 at Stanford by Shortcliffe
- Assist internists in diagnosis and treatment of
infectious diseases: meningitis & bacterial septicemia
- When patient shows signs of infectious disease, culture
of blood and urine set to lab (>24hrs) to determine
bacterial species
- Given patient data (incomplete & inaccurate) MYCIN
gives interim indication of organisms that are most likely
cause of infection & drugs to control disease
- Drug interactions & already prescribed drugs taken into
account
- Able to provide explanation of diagnosis (limited)
(3) XCON/RI (computer)
AI & KBS
- Configures DEC’s VAX, PDP11 and VAX
- DEC offers the customer a wide choice of components
when purchasing computer equipment, so that
client achieves a custom-made system
- Given the customer’s order, configuration is made,
perhaps involving component replacement or addition
- Problem: information subject to rapid change &
configuring a computer system requires
skills and effort
- Since 1981, XCON with XSEL assists DEC agents
in drawing up orders.
AI & KBS
(4) DRILLING ADVISOR (industry)
- Developed in 1983 by Teknowledge for oil company
to replace human drilling advisor
- Problem:drill bits becoming stuck
- Difficulty: lack of subsurface information on
location & condition on end of drill
- (scarcity) expert examines rock pieces, mud, lubricant
brought up by drilling to determine cause
- Drilling Advisor uses geological site information,
conditions of problem, historical information about
other problems experienced in the past
- Provide recommendation to correct problem & advice
on how to change current practices to avoid problem
AI & KBS
(5) Human Resource Management




HRM facilitates the most effective use of employees to
achieve organisational and individual goals
HRM KBS forms part of overall strategy (includes DSS &
EIS)
KBS helps decision making for HRM managers with
heuristic knowledge in unstructured & semi-structured
problems (job placement & pay rises)
Using semantic nets & Prolog, illustrates use of KBS in
HR planning, recruiting, compensation & labourmanagement relations
(see Human resource management expert systems
technology, Byun & Suh, ES, May 94, 11:2)
AI & KBS
5. Typical tasks of KBS
(1) Diagnosis - To identify a problem given a set of symptoms
or malfunctions.
e.g. diagnose reasons for engine failure
(2) Interpretation - To provide an understanding of a situation
from available information. e.g. DENDRAL
(3) Prediction - To predict a future state from a set of data or
observations. e.g. Drilling Advisor, PLANT
(4) Design - To develop configurations that satisfy constraints of
a design problem. e.g. XCON
(5) Planning - Both short term & long term in areas like project
management, product development or financial planning.
e.g. HRM
AI & KBS
(6) Monitoring - To check performance & flag exceptions.
e.g., KBS monitors radar data and estimates the position of
the space shuttle
(7) Control - To collect and evaluate evidence and form opinions
on that evidence.
e.g. control patient’s treatment
(8) Instruction - To train students and correct their performance.
e.g. give medical students experience diagnosing illness
(9) Debugging - To identify and prescribe remedies for
malfunctions.
e.g. identify errors in an automated teller machine network and
ways to correct the errors
AI & KBS
6. Advantages & Limitations
(1) Advantages
- Increase availability of expert knowledge
expertise not accessible
training future experts
- Efficient and cost effective
- Consistency of answers
- Explanation of solution
- Deal with uncertainty
AI & KBS
(2) Limitations
-Lack of common sense
-Inflexible, Difficult to modify
- Restricted domain of expertise
- Lack of learning ability
- Not always reliable
AI & KBS
Overview
- Traditional AI & its limitations for real-world problem
solving
- KBS emergence in 60’s
emphasis on specific domain-knowledge rather than GPS
separation of knowledge and reasoning
- KBS basic components:
knowledge-base, inference engine & user-interface
- Examples
- Advantages & limitations