Week 1: Preliminary Concepts
Table of Contents
Preliminary Concepts
What is Artificial Intelligence?
What is Thinking?
What is Reasoning?
Reasoning Paradigms
Symbolic Reasoning
Knowledge
Knowledge Representations
Preliminary Concepts
Artificial Intelligence
Knowledge
Representation
Reasoning
What is Artificial Intelligence?
Artificial intelligence is the study of intelligent behavior
achieved through computational means.
Artificial intelligence is not about:
evil killing machines
replicating human intelligence
Artificial intelligence is about:
problem solving by means of computational methods able to derive insight
from data
What is Thinking?
Some authors claim that thinking is a computational process. This idea is
controversial (and very simplistic)
Week 1: Preliminary Concepts 1
, However, this course is rather concerned with the “artificial thinking process”
performed by intelligent systems and agents. This poses the questions can
machines really think? → if we accept the hypothesis that thinking is a
computational process, then the answer would be yes.
What is Reasoning?
Reasoning is the formal manipulation of the symbols
representing a collection of propositions to produce
representations of new ones.
Overall, we need reasoning because the knowledge we have about the problem
domain is rather limited (we often have a simplified representation of the problem
being investigated), and sometimes imprecise.
Reasoning allows us to derive the missing pieces of knowledge from the pieces we
have represented in the knowledge base → we can answer a question that could not
be answered by simply inspecting the knowledge base
Reasoning Paradigms
Symbolic reasoning is about manipulating a collection of
symbols to produce knowledge. It is often transparent but limited.
Symbolic artificial intelligence often attempts to solve a given problem by exploiting
explicit, symbolic knowledge representations. → Also called Good Old-Fashioned
Artificial Intelligence
Sub-symbolic reasoning is about operating numerical
representations to produce knowledge. It is often powerful but
unintelligible.
Sub-symbolic artificial intelligence also attempts to solve the problem while learning
internal knowledge representations from data. → Sometimes, learning those internal
knowledge representations becomes the goal itself.
Symbolic Reasoning
Week 1: Preliminary Concepts 2
, Gottfried von Leibniz was a prominent German logician and mathematician known
for his redefinition of the binary number system. He stated that we do not operate
abstract entities but the symbolic representations of these entities. For example, the
numerical quantity “fourteen” is an abstract entity that can be represented as 14, XIV
or 1110.
The idea may be abstract but the symbols are concrete, so it would be enough to
define rules to manipulate these symbols and generate new ideas in form of
symbols.
The Leibniz idea in a nutshell is:
the rules of arithmetic allow dealing with abstract numbers in terms of
concrete symbols, so the manipulation of those symbols reflects the relations
among the numbers
the rules of logic allow dealing with abstract ideas in terms of concrete
symbols, so the manipulation of those symbols mirror the relations among
the ideas → Although the objects of human though are formless and
abstract, we can still deal with them concretely as a kind of arithmetic, by
representing them symbolically and operating on the symbols.
Symbolic representations are the process of encoding abstract
ideas using tangible symbols.
Overall, we want symbolic representations to be able to:
properly represent the problem domain. This includes both the knowledge
from experts and the data resources.
clearly express the emerging knowledge. This includes both the inner
knowledge that the intelligent system learns to solve the problem and the
knowledge it returns. → In principle, intelligent systems should produce the
same outcome regardless the symbols used to represent the input
knowledge
Granularity is the scale or level of detail in a set of data.
When playing chess, we might say to ourselves something like:
“Hmm. It moved this way because it believed its queen was vulnerable, but
still wanted to attack the rook.”
Week 1: Preliminary Concepts 3
, However an intelligent system can say something like:
“It moved this was because the evaluation procedure P using a static
evaluation function Q returned a value of +7 after an alpha-beta minimax
search depth of 4”
Propositions are ideas that can be expressed through
declarative sentences.
Propositions are considered to hold or to not hold.
People may or may not believe them, fear them, regret them, wish for them, worry
about them, and so on. These various relationships between people and
propositions are what philosophers call propositional attitudes.
Propositions are related to each other in certain ways: a proposition might imply or
provide evidence for, or contradict another proposition → On the other hand,
sentences are symbolic representations of propositions.
Logic is relevant to knowledge representation and reasoning since it allows studying
the entailment of relations, truth conditions and inference rules.
A collection of sentences S1, S2,...Sn logically entails another sentence S if the truth
of S is implicit in the truth of Si. Remark that we do now need to know what the
symbols mean but the truth value of propositions. → This is called logical
entailment and will allow us to reason with propositions.
Knowledge
Knowledge can be roughly defined as the understanding about
a certain domain, which is often expressed in terms of facts.
A knowledge base is a collection of symbolic structures
describing the problem domain with a specific granularity degree.
These knowledge structures are (ideally) believed to be the
ultimate truth about the problem being modeled.
A knowledge-based system is an intelligent reasoning system
that relies on the knowledge base to derive new knowledge to
Week 1: Preliminary Concepts 4
Table of Contents
Preliminary Concepts
What is Artificial Intelligence?
What is Thinking?
What is Reasoning?
Reasoning Paradigms
Symbolic Reasoning
Knowledge
Knowledge Representations
Preliminary Concepts
Artificial Intelligence
Knowledge
Representation
Reasoning
What is Artificial Intelligence?
Artificial intelligence is the study of intelligent behavior
achieved through computational means.
Artificial intelligence is not about:
evil killing machines
replicating human intelligence
Artificial intelligence is about:
problem solving by means of computational methods able to derive insight
from data
What is Thinking?
Some authors claim that thinking is a computational process. This idea is
controversial (and very simplistic)
Week 1: Preliminary Concepts 1
, However, this course is rather concerned with the “artificial thinking process”
performed by intelligent systems and agents. This poses the questions can
machines really think? → if we accept the hypothesis that thinking is a
computational process, then the answer would be yes.
What is Reasoning?
Reasoning is the formal manipulation of the symbols
representing a collection of propositions to produce
representations of new ones.
Overall, we need reasoning because the knowledge we have about the problem
domain is rather limited (we often have a simplified representation of the problem
being investigated), and sometimes imprecise.
Reasoning allows us to derive the missing pieces of knowledge from the pieces we
have represented in the knowledge base → we can answer a question that could not
be answered by simply inspecting the knowledge base
Reasoning Paradigms
Symbolic reasoning is about manipulating a collection of
symbols to produce knowledge. It is often transparent but limited.
Symbolic artificial intelligence often attempts to solve a given problem by exploiting
explicit, symbolic knowledge representations. → Also called Good Old-Fashioned
Artificial Intelligence
Sub-symbolic reasoning is about operating numerical
representations to produce knowledge. It is often powerful but
unintelligible.
Sub-symbolic artificial intelligence also attempts to solve the problem while learning
internal knowledge representations from data. → Sometimes, learning those internal
knowledge representations becomes the goal itself.
Symbolic Reasoning
Week 1: Preliminary Concepts 2
, Gottfried von Leibniz was a prominent German logician and mathematician known
for his redefinition of the binary number system. He stated that we do not operate
abstract entities but the symbolic representations of these entities. For example, the
numerical quantity “fourteen” is an abstract entity that can be represented as 14, XIV
or 1110.
The idea may be abstract but the symbols are concrete, so it would be enough to
define rules to manipulate these symbols and generate new ideas in form of
symbols.
The Leibniz idea in a nutshell is:
the rules of arithmetic allow dealing with abstract numbers in terms of
concrete symbols, so the manipulation of those symbols reflects the relations
among the numbers
the rules of logic allow dealing with abstract ideas in terms of concrete
symbols, so the manipulation of those symbols mirror the relations among
the ideas → Although the objects of human though are formless and
abstract, we can still deal with them concretely as a kind of arithmetic, by
representing them symbolically and operating on the symbols.
Symbolic representations are the process of encoding abstract
ideas using tangible symbols.
Overall, we want symbolic representations to be able to:
properly represent the problem domain. This includes both the knowledge
from experts and the data resources.
clearly express the emerging knowledge. This includes both the inner
knowledge that the intelligent system learns to solve the problem and the
knowledge it returns. → In principle, intelligent systems should produce the
same outcome regardless the symbols used to represent the input
knowledge
Granularity is the scale or level of detail in a set of data.
When playing chess, we might say to ourselves something like:
“Hmm. It moved this way because it believed its queen was vulnerable, but
still wanted to attack the rook.”
Week 1: Preliminary Concepts 3
, However an intelligent system can say something like:
“It moved this was because the evaluation procedure P using a static
evaluation function Q returned a value of +7 after an alpha-beta minimax
search depth of 4”
Propositions are ideas that can be expressed through
declarative sentences.
Propositions are considered to hold or to not hold.
People may or may not believe them, fear them, regret them, wish for them, worry
about them, and so on. These various relationships between people and
propositions are what philosophers call propositional attitudes.
Propositions are related to each other in certain ways: a proposition might imply or
provide evidence for, or contradict another proposition → On the other hand,
sentences are symbolic representations of propositions.
Logic is relevant to knowledge representation and reasoning since it allows studying
the entailment of relations, truth conditions and inference rules.
A collection of sentences S1, S2,...Sn logically entails another sentence S if the truth
of S is implicit in the truth of Si. Remark that we do now need to know what the
symbols mean but the truth value of propositions. → This is called logical
entailment and will allow us to reason with propositions.
Knowledge
Knowledge can be roughly defined as the understanding about
a certain domain, which is often expressed in terms of facts.
A knowledge base is a collection of symbolic structures
describing the problem domain with a specific granularity degree.
These knowledge structures are (ideally) believed to be the
ultimate truth about the problem being modeled.
A knowledge-based system is an intelligent reasoning system
that relies on the knowledge base to derive new knowledge to
Week 1: Preliminary Concepts 4
