Introduction to Artificial Intelligence

nosesarchaeologistAI and Robotics

Jul 17, 2012 (4 years and 10 months ago)


Introduction to Artificial Intelligence
Ming- Hwa Wang, Ph.D.
COEN 166/366 Artificial Intelligence
Department of Computer Engineering
Santa Clara University
Human Intelligence

human int elligence: use of int uit ion, common sense, j udgment,
creativity, goal- directedness, plausible reasoning, knowledge and beliefs

limitation: intellectually fallible, limited knowledge bases, information
processing of serial nature proceeds very slowly

people demonstrate their intelligence by communication and learning

effective communicat ion requires skills bot h in analysis of messages
(reception) and in synthesis of message (transmission)

learning is t o adapt behavior t o new sit uat ion, t o find out what t he
basic descript ive unit s are in a sit uat ion, and t o acquire t he rules for
combining primitive descriptors

organizing knowledge: relat e pieces of informat ion t o one anot her
and arrange them in memory
Artificial Intelligence

art ificial int elligence: a field of st udy t hat encompasses comput at ional
t echniques for performing t asks t hat apparent ly require int elligence
when performed by human

fundament al issues: knowledge represent at ion, search, percept ion and

knowledge can be available in collect ions of logical assert ions,
heuristic rules, procedures, and statistical correlations

search by avoiding the “combinatorial explosion”

inference: t he process of creat ing explicit represent at ions of
knowledge from implicit ones

deduct ive inference proceeds from a set of assumpt ions
( axioms) t o new st at ement s t hat are logically implied by t he

induct ive inference st art s wit h a set of fact s, feat ures or
observations, and it produce generalizations

artificial intelligence is both an art and a science

a science consist s of a body of proved principles t hat have been
abst ract ed from nat ure t hrough processes of empirical inquiry and
logical deduction

an art is a collect ion of t echniques, developed pragmat ically t o a
sophisticated level, but not necessary in a logic way

purpose: t o increase human’s underst anding of reasoning, learning, and
percept ual processes; t o build useful t ools; and t o achieve a more
mature view of human intelligence

Lady Lovelace’s Obj ect ion: comput er can do as t hey are t old and
consequently cannot perform original (hence, intelligent) actions

an ult imat e int ellect ual and t echnical challenge: what could be a more
intense intellectual experience than creating an intelligence?

early failures due to not produce the promised systems

the Turing’s test

machine learning: concept learning and general- to- specific ordering,
decision t ree learning, neural net work, Bayesian learning, inst ance-
based learning, genet ic algorit hms, analyt ical learning, reinforcement
I ndustrialization of Artificial Intelligent and Expert Systems

theorem proving and automated reasoning system

Intelligent agents

nat ural language processing, speech recognit ion: German/English

data mining

machine vision

planning and robotics


expert systems
The Future of Artificial Intelligence

as t he cost of comput ing syst ems cont inues t o fall, AI t echnology
becomes more widespread

AI pushes comput er t echnology: symbolic programming, parallel
processing, special- purpose IC

int egrat ion of AI t echniques wit h ot her informat ion processing

performing t ask t hat ot herwise requires experienced professionals, labor
cost my be greatly reduced

AI soft ware can be copied and permit a wider dist ribut ion of operat ional
expert ise, and can remain for arbit rarily long t o gain a permanence of

take over various mental or labor activities from human

machine might posses intelligence can be frightening

psychological t hreat t o human ident it y crisis, cont roversy about t he
concept of person ( mind) and t he concept of machine ( art ificial