Old exam Flashcards
Map the following techniques for creating large ontologies to the ontologies where they have been primarily used
- Collaborative creation of the knowledge base -
- Manual creation of the knowledge base -
- Automated extraction from Wikipedia -
- Automated learning from Web documents -
a) NELL
b) Cyc
c) DBPedia and Yago
d) Freebase and WikiData
Collaborative creation of the knowledge base - Freebase and WikiData
Manual creation of the knowledge base - Cyc
Automated extraction from Wikipedia - DBPedia and Yago
Automated learning from Web documents - NELL
Recall the blocks world from the lecture. We want to add a new operator for the robot so that it is able to kick the top-most block of a stack
down to the table. The remainder of the stack will not change.
Action( kick-down(X),
PRECOND: block(X),
clear(X),
on(X,Y),
block(Y),
handempty
ADD: ???
DELETE: ???
)
Which of the following literals would you put in the ADD-list, the DELETE-list, or neither of the two?
handempty:
holding(X):
clear(Y):
on(X,Y):
on(Y,table):
on(X,table):
handempty: neither
holding(X): neither
clear(Y): ADD
on(X,Y): DELETE
on(Y,table): neither
on(X,table): ADD
In Searle’s Chinese Room Argument: Who does certainly not speak Chinese?
The person who executes the rules inside the room.
In the lecture, we have seen Eliza and ChatGPT at work. Which of the following statements is not true?
- ChatGPT works with patterns automatically learned from human text, whereas Eliza’s patterns are handcrafted
- ChatGPT has a deep understanding of the world, while Eliza has not
- Eliza often inserts parts of the input into the reply, whereas ChatGPT can generate entirely new replies
- ChatGPT can dynamically adapt the style of its reply, whereas Eliza relies on a fixed number of pre-defined patterns
ChatGPT has a deep understanding of the world, while Eliza has not
Map the type of ambiguity in natural language processing to their definition:
1. the same sentence can be read in different ways
2. the correct interpretation of a sentence depends on the context
3. the meaning of a sentence may depend on previous sentences in the conversation
4. the same word may have different meanings
a lexical ambiguity
b syntactic ambiguity
c ambiguity of discourse
d semantic ambiguity
- b) syntactic ambiquity
- d) semantic ambiquity
- c) ambiquity of discourse
- a) lexical ambiquity
What is the Feigenbaum Bottleneck?
The difficulty of encoding human knowledge in logic
What three are the main parts of Watson, the machine which played ‘Jeopardy!”?
d. information retrieval for retrieving candidate answer
e. natural language understanding to parse inputs
f. knowledge and reasoning for matching candidate answers with evidence
Which of the following statements about ontologies is not true?
Wählen Sie eine Antwort:
a. they may be viewed as semantic networks
b. they are frequently implemented with a language called OWL
c. they provide a formalized description of the knowledge in a domain
d. they are based on inductive reasoning for automatically expanding their knowledge bas
they are based on inductive reasoning for automatically expanding their knowledge base
Map the following components of Searle’s Chinese Room argument to the components of a modern computer.
1. Rule Book
2. English-speaking man inside the room
3. Baskets with symbols
4. Letter slot of the door
a) Memory
b) Program
c) CPU
d) Input/Output Units
- Program
2.CPU
3.Memory
4.Input/Output Units
How is the semantic embedding in Word2Vec determined?
train a neural network to predict a word from its surrounding words
If an action C is in conflict with a causal link A ➔ p ➔ B, which three of the following options are certainly true?
a. the conflict cannot be resolved and backtracking must be performed
b. p is an effect of C
c. not(p) is an effect of C
d. C could be ordered before A unless a previous ordering constraint prevents this
e. C could be ordered after B unless a previous ordering constraint prevents this
f. C could be ordered between A and B unless a previous ordering constraint prevents this
c. not(p) is an effect of C
d. C could be ordered before A unless a previous ordering constraint prevents this
e. C could be ordered after B unless a previous ordering constraint prevents this
Put the correct variables into the head of the following rule for great_aunt(X,Y), which should describe that Xis the
great aunt of Y, i.e., Xis the sister of one of Y’s grandparents.
great_ aunt( __ , __ ) :- parent(C,A), sister(B,C), parent(A,D).
great_ aunt( B , D )
