PHIL1-3

Question 1

What is philosophy

Answer 1

Search for conceptual clarity and method of careful and rigorous justification

Question 2

What is philosophy of AI

Answer 2

Aims to clarify the concept of intelligence (Along with interrelated concepts)

Question 3

What are descriptive claims

Answer 3

What is and what is not. Describing what or how

Question 4

What are normative claims

Answer 4

Claims about what should and should not be. According to norms

Question 5

What is a necessary condition?

Answer 5

Everything that is intelligent is P. S is intelligent -> S possesses property P.

Given this, we are able to determine that certain things are not P.
Anything that is not P is not intelligent.. S does not possess P -> S is not intelligent

Question 6

What is a sufficient criteria?

Answer 6

Everything that is Q is intelligent. S possesses Q -> S is intelligent. We can allow for the possibility that some S may be intelligent whether or not we recognise it.

Question 7

What are biological criteria and their related concepts?

Answer 7

Properties relevant are properties of biological organisms.

Overly restrictive chauvinism: Non biological entities are excluded from the start. Is that justified?

Excessive liberalism: Many organisms possess biological properties, are these intelligent? (Plants)

Question 8

What is computational criteria?

Answer 8

The properties that define certain classes of computational systems. E.g. Turing Machines, PSS etc.

Question 9

What is a consequence and therefore challenge of computational limitations?

Answer 9

Features of intelligence that cannot be replicated by either one or all of those kinds of computational systems like creativity consciousness

Question 10

What is the challenge of cognitive scientific uncertainty?

Answer 10

We are unsure whether human beings are computational systems, so why suppose any kind of computation is sufficient for intelligence?

Question 11

What are behavioral criteria?

Answer 11

Relevant properties for the attribution of intelligence are behavioral. S possess an IG test, possesses capacities like walking talking etc.

Question 12

What are the different types of criteria? Are they necessary or sufficient?

Answer 12

Biological, computational and behavioral. These are sufficient.

Question 13

What is the Turing Test Criterion?

Answer 13

S does well at the Imitiation Game -> S is intelligent

Question 14

What type of criterion is the Turing Test?

Answer 14

Behavioral. S needs to behave in a certain way that fools an interrogator. It is sufficient (octopi, preverbal children)

Question 15

What is the heads in the sand objection?

Answer 15

Instead of an argument against the claim that machines can think: It is an expression of various fears about what might follow if there were thinking machines. (Turing Test)

Question 16

What is the argument of various disabilities?

Answer 16

Imitation game only considers verbal behavior. Analogous games could be designed for any other behaviors.

Question 17

What is the Total Turing Test and the criterion?

Answer 17

It is perhaps not sufficient for S to possess just any of the behavioral capacities but must possess all. S “fools humans in any conceivable context” -> S is intelligent

Question 18

What is Lady Lovelace’s objection?

Answer 18

Machines only do what we know how to order them to do. — becomes less compelling as our ability to control and predict the machine’s behavior decreases (ML and NN)

Question 19

What is the threat of pretense?

Answer 19

“Can’t something behave as if intelligent without actually being intelligent? Does intelligence depend on capabilities or also how you do it?

Question 20

What defines a PSS implementing a TM ?

Answer 20

Finite alphabet of symbols. Finite set of states and table of transition rules from one pairing (symbol,state) to another.

Question 21

What is the relationship between problems and TMs?

Answer 21

Every computable problem can be solved by a TM. (Computable = solvable through rule-based manipulation of symbols)

A universal TM can solve all such problems.

All programmable computers that we use today are universal TM’s

Question 22

What is a problem and what makes it interesting?

Answer 22

A problem is a transformation of one symbol string to another. Interesting: Can be interpreted by us - can think of them as representing something meaningful. They reflect norms: Rules of mathematics, principles of x..

Question 23

What is the argumentation to say AI is possible?

Answer 23

P1. We can design a TM to solve any computable problem
P2. All ‘interesting’ problems are computable
C1. We can design a TM to solve any interesting problem
P3. Being intelligent is the ability to solve interesting problems
C2. We can design a TM that is intelligent
P4. A digital computer can implement any TM
C3. We can build a digital computer that is intelligent —> AI is possible

Question 24

What is a PSS and what does it entail?

Answer 24

Implemented TM with interpretable symbols.

Possesses symbols, states and transition rules.
Implemented in some physical system
Can be interpreted with respect to some environment, context or domain (e.g. solves interesting problems)

Question 25

Which one is correct? 1) All TMs are PSSs, but not all PSSs are TM’s 2) All Psss are TMs, but not all TMs are PSSs

Answer 25

2

Question 26

State the PSSH

Answer 26

A physical symbol system has the necessary and sufficient means for general intelligent action. Necessary: Any intelligent thing is a PSS Sufficient: Any PSS is intelligent - Newell and Simon

Question 27

What is the PSSH response to the challenge of cognitive scientific uncertainty?

Answer 27

Every intelligent organism is a PSS and every intelligent action results from rule-based symbol manipulation

Question 28

How would PSSH describe tasks like decision making?

Answer 28

A series of transitions between symbolic states

Question 29

What are the questions typically asked by Symbolic AI researches when designing “intelligent systems”

Answer 29

Which symbols should be used to represent the task environment? Which rules should be deployed to arrive at a particular goal? How should those symbols and rules be implemented?

Question 30

Explain the Symbolic AI’s Sense -> Think -> Act loop

Answer 30

Sense: Transform perceptual stimuli of the physical world into symbol strings. (Objects, Indicators) Think: Solve the relevant “interesting” problems by applying transition rules. (Calculate/predict/identify) — Rules may perform heuristic search through a space of possible solutions Act: Transform symbol strings into physical motion (Steering, braking, accelerating) REPEAT

Question 31

What is the Frame Problem?

Answer 31

It is difficult to distinguish the relevant from the irrelevant problems in any given context — Dennett, 1984