stud
bud
Sign up
Log in
Home
Library
Explore
SAT
More
MLP Quiz Flashcards | StudBud
Explore
/
MLP Quiz
MLP Quiz
15 terms
·
StudBud Administration
·
15 plays
Imported from external platform
Flashcards
15 cards
F
Practice quiz
P
Host live
L
Questions
15
Answers
1
What are approaches to do language generation?
Natural language generation and generation as an inverse of parsing
Hidden Markov Generation and text prediction with large language models
Natural language generation and text prediction with large language models
Generation as an inverse of parsing and Hidden Markov Generation
2
Two words are similar...
...when their cosine similarity is close to 0
...when they're located close to each other in the embedding space
...when they appear in different environments
...when their cosine similarity is close to 1
3
The Porter Stemmer...
...gives a detailed morphological analysis
...removes affixes (mainly suffixes) based on rules
...is a machine learning algorithm
...contains a lexicon
4
What are properties of tf-idf (term frequency - inverse document frequency)?
Sparse vectors that are trained
Dense vectors that are based on occurrence statistics
Sparse vectors that are based on occurrence statistics
Dense vectors that are trained
5
Which pair is true?
n-grams - sequence of n words (or items in general)
Finite State Morphology Analyser - Detailed morphological analysis
f-structure - dependency
Context free grammar - hierarchical structure
6
What is a common ambiguity in syntax?
Coordination
Topicalisation
PP-attachment
Subject verb agreement
7
Which of the following strings are described by this regular expression?
WoooooooowTaco
WWooooooTaco
WoooowTacosss
WoTacos
8
What is a language model?
A very beautiful language
The probabilities in a language model are based on a corpus
A model to compute the probability of a sentence or sequence of words
Model to compute the probability of an upcoming word
9
What does this denote in regular expressions?
The letter "w"
Not the letter "w"
Alphanumeric characters and the underscore
Any digit
10
A probabilistic context free grammar...
...works well for free word order languages
...has probabilities added to its rules
...considers lexical information
...assumes dependence between constituents
11
Common issues in tokenizers are:
Special characters like quotation marks
Multi word expressions
Punctuation
Languages with white spaces
12
A dependency parser provides the following output:
Relations between heads and its dependents
Constituency structure
POS tags
f-structure
13
What does the following First Order Logic expression mean?
The driving test did not fail Daniel.
Daniel shouldn't be allowed to drive.
Daniel did not fail his driving test.
Daniel failed his driving test.
14
The following Finite State Network can parse...
apple
apppppppple
aples
ales
15
Data in machine learning settings are usually split into:
Unsupervised Training, Supervised Validation, Test
Easy Training, Difficult Training, Test
Training, Validation, Test
Training, Development, Test
