Intro Intelligence

Question 1

Intelligence

Answer 1

There is no commonly accepted definition of intelligence, but there are considered different forms, i.e. social,emotional, senso-motoric etc.
● Intelligence is the ability to deal with
complexity
● Intelligence is Goal-Directed adaptive
behavior
● Intelligence is a force, F, that acts to
maximize future freedom of action

philosophical aspect of intelligence.
Intelligence is the thing that makes us who we are

Indirect goal of AI: computational precision of the everyday notion of intelligence.

Question 2

Turing Test

Answer 2

● Is an attempt at a better definition of
intelligence, invented by Alan Turing.
● Turing’s definition is an “operational” one
(as opposed to a conceptual one): He
defined intelligence on the basis of behavior
that is indistinguishable from that of a
human’s.
● Turing’s idea was this: A computer program
is intelligent if it answers (written) questions
asked by a person in such a way that it
seems like the answers could only be
generated by another human being. If the
interviewer is fooled, the test is passed.

eg: Google Duplex (calls)

Question 3

Motivations behind and requirements on AI:

Answer 3

• Visionary: build artefacts that produce intelligent behaviour in manner of humans
• Pragmatic: build artefacts that show behaviour comparable to human intelligent behaviour

eg: AlphaZero (Go AI)

There are also considered 2 types of AI:

• Weak: acts as if intelligent
• Strong: actually thinks

Question 4

Knowledge Based AI

Answer 4

(1956 -)
Guiding Model: A human being
Guiding Assumptions:
● The belief that Intelligence is (or eventually
boils down to) knowledge representation
● A von Neumann computer is a perfect
model of the human “cognitive apparat”
Key terms:
Symbol System Hypothesis: The ability to produce
and manipulate symbols is a necessary and
sufficient condition for intelligence (Essentially,
intelligence is a symbol system, with symbols
representing the state of all our current knowledge
and beliefs)
Top-down design: To build AI, we start with
high-level concepts at the knowledge level and
break them down into smaller, programmable units.
Major Problem: Symbol Grounding - How do we
ensure our symbols relate to the real world?

Question 5

Behavior based AI

Answer 5

(1985 -)
Guiding Model: Human or Animal
Guiding Assumptions:
● Intelligence is built upon elementary
behavioral activities (e.g., moving along a
wall, grasping an object)
● Behavior-based implies acting in the real
world - thus a coupling between the sensory
system and the motor system is essential
● There can be no intelligence unless it is
grounded in an ability to act in the real world
● This leads to the Physical Grounding
hypothesis
Physical Grounding Hypothesis: Rooting of
symbols in the real world is a necessary condition
for intelligence
(No rooting → No meaning → No intelligent
behavior)

Question 6

Connectionism

Answer 6

Guiding model: (human) brain
Guiding Assumptions:
● Processing of information through very
simple but many interconnected units
(neurons) that interact at a low (in terms of
signal-processing) level
● A connectionist system will be massively
parallel, distributed, and process information
below the level of symbols

Question 7

Distributed AI

Answer 7

Distributed AI (1980 -)
Guiding Assumptions:
● An intelligent system could actually be a
group of interacting intelligent beings that
interact on the knowledge level

Key issues: Communication, coordination,
cooperation, negotiation, organizational structure

Why
would we deal with distributed intelligence?
· Some problems can only be solved on the basis of high-level interaction among intelligent
entities.
· Parallelism, scalability, robustness
· Close relationship among intelligence and interaction
· Intuitively clear approach to complex applications