Week 1

Question 1

What was one of the impresive results of early AI efforts?

Answer 1

The Lisp-based ‘Macsyma', a powerful symbolic mathematical system.

Question 2

What limitations did computers and programs of the 1980s have in terms of AI?

Answer 2

They were far from capturing the fluid, adaptive intelligence of humans in tasks like object perception, language understanding, and contextual actions.

Question 3

what is the difference between thinking and acting humanly regarding machines

Answer 3

Thinking: creating cognitive processes the same as a human.
Acting: we don’t really care about cognitive processes as long as it behaves as a human and delivers the output.

Question 4

Define the terms and relationships between them:

• algorithm
• Function
• arguments

Answer 4

algorithm: a set of rules that defines a sequence of operations=computations using arguments
arguments=input values
function= list of instructions that can take arguments, apply an algorithm, and return output.

Question 5

what is a difference in looking at problems between psychology and AI?

Answer 5

psychology=from top to bottom. behavior or phenomena down to neural explanations: inverse. further down, we don’t have the knowledge (yet) that explains behavior

AI: from bottom to top= forward: We know how behavior is produced because we programmed it.

Question 6

timeline of important AI related events

Question 7

what is defined by weak AI according to the turing test

Answer 7

sentient=consciouss

Question 8

How does de Kleijn et al. define intelligence
in a Turing test?

Answer 8

Question 9

what is Searle’s notion about strong and weak AI?

Answer 9

He was skeptical in a machine’s capabillity to have human like consciousness (strong AI). –>chinese room. His views emphasize the difference between mere computational processing and genuine understanding or consciousness.

Question 10

describe the chinese room argument of Searle

Answer 10

Rule-based manipulation of symbols does not constitute intelligence: the
inhabitant of the Chinese room does not understand Chinese.

Question 11

What is strong AI?

Answer 11

Question 12

what is a problem in the definition of strong AI according to Dijkstra?

Answer 12

Question 13

what was Minsky’s contribution to the development of AI?

Answer 13

Question 14

What is named as the birth of AI?

Answer 14

Question 15

what is the definition of a symbol system?

Answer 15

a system that:
* uses symbols: In humans=numbers, words, abstract representations. Computers=binary code, letters, or abstract networs
* manipulates symbols occording to known rules

Question 16

• describe how symbolic AI works:
• name a form of symbolic AI
• what is another name for symbolic systems?

Answer 16

Question 17

What is the ELIZA system?

Answer 17

Question 18

How does ELIZA work?

Answer 18

Question 19

some early planners where susceptible to the sussman anomaly. describe this anomaly by using the “stacking blocks assignment”

Answer 19

regardless with which goal we start we won’t end up with the correct solution. Simple lineaire planningapproaces won’t work, because we need to undo cerctain actions to reach the solution

Question 20

What was the MYCIN system?

Answer 20

system that diagnosed blood infections.

Question 21

what happend in the 70s regarding AI?

Answer 21

Question 22

In the 80’s Connectionism was revived. What was McClelland’s attribution to this?

Answer 22

He contributed to the Model of human memory. founding of PDP work

Question 23

What does this early PDP work allows us to do regarding retrieving information about for example members?

Answer 23

• Find properties of members
• identify a member by properties
• identify general characteristics

Question 24

what does generalization mean in the context of AI?

Answer 24

a model’s ability to adapt properly to new, previously unseen data, drawn from the same distribution as the one used to create the model.

Question 25

What does it mean to say that connectionist AI is: biologically inspired, lesion tolerant and capable of generalization?

Answer 25

Question 26

What made deep learning in AI possible during the 2000’s?

Answer 26

the inventions that made storing huge amounts of data more affordable, which we could use to train AI programs.

Question 27

what is the key difference between symbolic and connectionist AI?

Answer 27

• Symbolic AI: Uses explicit, predefined rules and symbols to represent knowledge. It’s akin to using human-readable code or logical statements to encode information and decision processes.
• Connectionist AI: Represents knowledge implicitly through the architecture and weights of neural networks. Learning occurs by adjusting these weights based on input data, rather than by manipulating symbolic rules.

Question 28

what is the key difference between weak and strong AI?

Answer 28

Question 29

make sure at the end you can answer these questions