Download Symbols and Search (Allen Newell and Herbert A. Simon)

DeGroff 4/29/2017
Classic Paper Study/Discussion Guide
Title: “Computer Science as Empirical Inquiry: Symbols and Search”
Author: Allan Newell and Herbert A. Simon
Knowledge Relating to the Learning Outcomes:
1.
Symbol Systems:
a. Symbols lie at the root of intelligent action, which is, of course, the
primary topic of artificial intelligence. For that matter, it is a primary
question for all of computer science.
2.
Symbol Systems:
a. A physical symbol system consists of a set of entities, called symbols,
which are physical patterns that occur as components of another type of
entity called an expression (or symbol structure).
3.
Big Idea:
a. The roots of the hypothesis go back to the program of Frege and of
Whitehead and Russell for formalizing logic: capturing the basic
conceptual notions of mathematics in logic and putting the notions of
proof and deduction on secure footing.
4.
Interdisciplinary Approach and Psychology:
a. The symbol system hypothesis implies that the symbolic behavior of man
arises because he has the characteristics of a physical symbol system.
Hence, the results of efforts to model human behavior with symbol
systems became an important part of the evidence for the hypothesis, and
DeGroff 4/29/2017
research in artificial intelligence goes on in close collaboration with
research in information processing psychology, as it is usually called.
5.
Psychology:
a. The principle body of evidence for the symbol system hypothesis that we
have not considered is negative evidence: the absence of specific
competing hypotheses as to how intelligent activity might be
accomplished—whether by man or machine. Most attempts to build such
hypotheses have taken place within the field of psychology.
6. Big Idea:
a. To deal with this puzzle, Plato invented his famous theory of recollection:
when you think you are discovering or learning something, you are really
just recalling what you already knew in a previous existence.
7. Consciousness:
a. If you think there is nothing problematic or mysterious about a symbol
system solving problems, then you are a child of today, whose views have
been formed since mid-century. Plato (and, by his account, Socrates)
found difficulty understanding even how problems can be entertained,
much less how they could be solved.
DeGroff 4/29/2017
Top Five Items of Interest:
1.
The focus on the laws of qualitative structure in science. AI is compared to
Pasteur’s theory of germs as the cause of disease. The idea is to assume a
solution and then try to work backward from the solution to the origin of the
problem.
2.
The Physical Symbol System Hypothesis. The entire point of the article, Newell
and Simon claim that Physical Symbol Systems have intelligence due to their
organization and ability to adapt to the demands of a problem. I find it funny that
immediately after the Physical Symbol System Hypothesis is introduced it is then
quickly referred to as a law of qualitative structure.
3.
That data be inert is essential to the reduction of computation to physical
processes. Data must not inherently mean anything, thus a symbol is not
something that designates any particular thing. The modularity of the symbol
system is what allows for machines to use them.
4. Intelligence seems to equal ‘the ability to solve a problem better than random
search’. Newell and Simon claim that the appearance of intelligence is that the
distributions of solutions be not entirely random, that patterns be detectable, and
that solutions occur in response to changes of pattern. This makes sense, but
leads to funny possibilities: Doe the change of course of a river to the lowest and
easiest flowing avenue due to the changes in environment imply that the river is
intelligent? The change of course does not appear random, and the change of the
course is predictable, and the river does change in response to the previously
chosen changes in course. What makes the river not intelligent?
DeGroff 4/29/2017
5.
“The task of intelligence, then, is to avert the ever-present threat of the
exponential explosion of search.” I like this approach, especially in terms of
artificial intelligence. Human beings seem to subconsciously decrease their
branching of problem solving (as seen in language acquisition of children). The
goal of AI is to reproduce human-level thinking, and thus the reduction of search
is the correct approach.