Download Introduction to Artificial Intelligence

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Document related concepts

Human–computer interaction wikipedia , lookup

Wizard of Oz experiment wikipedia , lookup

Knowledge representation and reasoning wikipedia , lookup

Artificial intelligence in video games wikipedia , lookup

Technological singularity wikipedia , lookup

Computer Go wikipedia , lookup

Embodied cognitive science wikipedia , lookup

AI winter wikipedia , lookup

Intelligence explosion wikipedia , lookup

Ethics of artificial intelligence wikipedia , lookup

Existential risk from artificial general intelligence wikipedia , lookup

Philosophy of artificial intelligence wikipedia , lookup

History of artificial intelligence wikipedia , lookup

Transcript
Chapter 1
Introduction to Artificial Intelligence
1.1.1 What is intelligence, as presumably evidenced by humans?
This is a very tough question for anybody to answer, so the first response to this question is to
ask why we, as computer scientists, should care about the number of possible answers to this
question. The reason that we should care is that a number of models of intelligence lead
immediately to distinct approaches to software design. We shall begin with one model that is
not very helpful and move on to those we can use.
The Human Computer
There are known cases of humans with remarkable computing abilities. Many of these are
called “idiot savants” in that they have this ability at the cost of normal functionality. A
calendar calculator is a typical idiot savant, he can almost immediately tell what day of the
week any given date fell on (such as July 27, 1782), but probably cannot make change for a
twenty dollar bill. More perplexing are fully-functional people with such abilities; there is a
well-known scientist who used to work at Los Alamos; he could take the square root of a 51
digit number in his head in about five seconds. Apart from this ability, which he considered
on the same level as a well-executed magic trick, the man was a remarkable scientist who
interacted well with others (a trait absent in idiot savants) and did much productive work.
Consider a simpler problem – addition of two numbers with large (20 or more) numbers of
digits. The inability of most of us to do this in our heads evidences one fact about our mental
processes, that they are not “register oriented” – we cannot set up variables mentally, store
values in them, and instantly recall the values as needed.
The reader should note one of the consequences of the last examples. People who appear to
have a “register oriented” memory in that they can memorize and recall large numbers are rare,
while most models of computing machines are register (or variable) oriented. This indicates
that a modern digital computer may not be a natural model of the human mind, leading to the
conclusion that human intelligence may be hard to mimic in such a computer.
Intelligence as the accumulation of knowledge
One model of intelligence focuses on the acquisition of large amounts of knowledge and the
ability to reason using such knowledge. In an informal sense, this picture represents one of the
more common ideas of intelligence; how well does one score on an IQ test. Put simply this
view equates intelligence with the ability to do symbolic reasoning.
One significant feature of programs that model this behavior is the facility to explain the line
of reasoning that lead to a specific answer. This feature tends to enhance acceptance by
human experts, such as medical doctors, who are far more comfortable accepting results for
which consistent explanations are offered.
The computational model that reflects this approach is that of expert systems. When we study
expert systems, we shall find them to be quite powerful, but only in narrow ranges of
specialization. Specifically, expert systems do not adapt well to unexpected inputs.
Page 4 of 15
CPSC 5185
Revised June 26, 2017
Copyright © 2005 by Edward L. Bosworth, Ph.D. All rights reserved.