CI-490: 10.27.99
Instructor: Ed Quigley 724.774.2088

Chapter 13: Artificial Intelligence and Expert Systems

(pg. 387)

Resistance Is Futile: See sidebar, pg. 388
See Cat Images, www.

Artificial Intelligence
Artificial: Man Made
Intelligence: Adaptability. Learning. 

Robotics. (p. 389) AI in physical world. Things that move, and affect other 
physical objects
Artificial Vision, Object Recognition.
NLP: Natural Language Processing

Expert Systems
Sortware that solves difficult multi-factor problems in tightly defined 
domains utilizing the distilled reasoning, resources, and processes of human experts.

Domain Expert (SME)
Knowledge Engineer makes tacit Knowledge into system knowledge
Rule-Based Knowledge Representation
Other techniques of Knowledge Representation:
	Semantic Frames
	Semantic Networks

Neural Networks
Machine Learning
Learns through trial and error

Fuzzy Logic
Rather than representing Knowledge through if/then conditions, permits definition 
by shades of gray rather than 0/1. Describes situations through linguistic variables. 
Can be said to be a representation of analog (non-binary) info. Permits human-like modeling.
See: Turing test

Genetic Algorithms (p. 389)
mathematical functions that use random mutation and Darwinian (evolution of the fittest)
principles to improve an application or procedure. Given a definition of a successul 
operation, the system proposes and analyzes myriad small mutations, kills those that work 
worse than original procedure, promoted those that work better than baseline, then uses all 
the improvements to cross-breed the next generation of a simulation, for many generations.

Intelligent Agents (p. 399)
Most agents in use on web today are not Intelligent. 
Biggest use of agents: web-bots (aka spiders), buy-bots, news bots.

What is Expertise? (p.400)
skill and knowledge, gained from experience, producing significantly better results.
hueristics : dont spit into the wind
metaphors: don't count your chickens until they're hatched

Expert Systems Components:
Dialog Management (interface)
Knowledge Base
Inference Engine (where the rules and procedures are executed)

Some Issues:
What is Legitimate Compensation for a Knowledge dump

p.404 Expert System Shells (4GL ES)
Forward Chaining
Backward Chaining
	Example: Landing on the moon

See pg. 406, Fig 13.12, Justification of an ES
See Page 407, Why Should Managers Know about AI?

p.411, Limitations of ES