Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

CS5012 > Week 4 > Flashcards

Week 4 Flashcards

1
Q

What is a tagset?

A

Set of POSs that we use to classify words in a sentence

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What are some applications for POS tagging?

A 
<ul>
<li>Speech generation - learning how to pronounce words</li>
<li>Information extraction</li>
<li>Parsing</li>
<li>Machine translation</li>
</ul>
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the closed class and what are some tags that are in the closed class?

A

A class of tags that have small and fixed membership. Examples include prepositions, auxiliary verbs and pronouns

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are some prepositions?

A

of, in, by

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What are some auxiliary verbs?

A

may, can, will, had, been

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is the open class and what are some tags that are in the open class?

A

Many new words are created in this class all the time, examples include nouns, verbs and adjectives

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is a standard used to tell if two words should be in the same class?

A

Can we exchange word A for word B?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What are nouns?

A

Words for animals, places, things,…

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What are proper nouns?

A

Nouns that are names, capitalized and not normally preceded by articles

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What are common nouns?

A

Articles that are not proper nouns

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What are the two types of common nouns?

A

<ul>
<li>Count nouns</li>
<li>Mass nouns</li>
</ul>

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What are count nouns, and give an example?

A

Can occur in singular/plural e.g. goat(s), relationship(s) and can be counted

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What are mass nouns, and give an example?

A

Don’t get counted e.g. snow, communism

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Describe verbs

A

Mostly describe actions and processes

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Give some examples of auxiliary verbs

A

be, have, do

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What are modal verbs?

A

Auxiliary verbs that indicate mood (can - ability), may (possibility), must (necessity)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What are adjectives?

A

Terms that describe properties or qualities

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What are adverbs?

A

Words that modify verbs, adjectives or other adverbs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

Give an example of a manner adverb

A

Walking <b>slowly</b>

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

Give an example of a directional adverb

A

Driving <b>downhill</b>

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

Give an example of a temporal adverb

A

She is leaving <b>tomorrow</b>

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

Give an example of a degree adverb

A

<b>Extremely</b> good

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

What are prepositions and give an example?

A

Words that indicate spatial, temporal or other relation to noun phrase e.g. to, on, by

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

What are a verb and a particle together called?

A

Phrasal verb

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

What is a particle?

A

Similar to a preposition or an adverb but tied to a verb e.g. turn the paper <b>over</b>

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What is <b>this</b> chapter an example of?

A

Demonstrative determiners

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

What is a coordinating conjunction?

A

When two elements of equal status are joined

28
Q

What can be used to make a coordinating conjunction?

A

and, or, but

29
Q

What is a subordinating conjunction?

A

When one element is embedded into another

30
Q

What can be used to make a subordinating conjunction?

A

that

31
Q

What are pronouns?

A

Way to refer to some entity without naming it explicitly

32
Q

What are the three type of pronouns?

A 
<ul>
<li>Personal pronouns</li>
<li>Possessive pronouns</li>
<li>Interrogative pronouns</li>
</ul>
33
Q

What is a benefit of having a coarser tag set?

A

Easier to do tagging by hand

34
Q

What is the benefit of having a finer tagset?

A

More information can be obtained from annotated text

35
Q

How are morphology and tag sets related?

A

The richer the morphology, the more information can be put into tag sets

36
Q

What is the main difficulty related to tagging?

A

Ambiguity

37
Q

What are the three main approaches for automatic tagging?

A 
<ul>
<li>Rule-based</li>
<li>Transformation based</li>
<li>Stochastic</li>
</ul>
38
Q

Describe rule based tagging

A

<ul>
<li>Use dictionary to find all possible tags for each word token</li>
<li>Repeatedly apply hand made rules to narrow down to one tag per word</li>
</ul>

39
Q

Describe a typical rule in rule based tagging

A

Negative constraints i.e. if this condition doesn’t hold, eliminate the tag

40
Q

What is transformation based tagging?

A

Where the tag of a word is changed based on the previous and following tags

41
Q

Describe a transformation rule in transformation based tagging

A

If the previous and following words have certain tags, then change current tag X into tag Y

42
Q

How is training done for transformation based tagging?

A

<ul>
<li>Assign most likely tag to each word token</li>
<li>Consider all possible rules</li>
<li>The one that improves the accuracy the most is taken to be rule number 1</li>
<li>Apply rule 1 everywhere</li>
<li>Repeat above steps until no more substantial improvement is possible</li>
</ul>

43
Q

What are some benefits of rule based tagging?

A

<ul>
<li>Requires no training data</li>
</ul>

44
Q

What are some disadvantages of rule based tagging?

A

<ul>
<li>Costly hand crafted rules</li>
<li>Requires data to validate rules</li>
</ul>

45
Q

What is the benefit of transformation based tagging?

A

<ul>
<li>Able to apply machine learning</li>
</ul>

46
Q

What is the disadvantage of transformation based tagging?

A

<ul>
<li>Requires tagged data</li>
<li>No clear statistical interpretation</li>
</ul>

47
Q

What are the advantages of using HMM for tagging?

A

<ul>
<li>Can do machine supervised learning</li>
<li>Can do unsupervised learning with untagged data</li>
</ul>

48
Q

What is the disadvantage of using HMM for tagging?

A

<ul>
<li>Requires tagged data</li>
</ul>

49
Q

What are the four main open classes in English?

A 
<ol>
<li>Nouns</li>
<li>Adjectives</li>
<li>Adverbs</li>
<li>Verbs</li>
</ol>
50
Q

Describe directional adverbs

A

Adverbs that specify the direction or location of some action

51
Q

Describe degree adverbs

A

Specify the extent of some action, process or property

52
Q

Describe manner adverbs

A

Describe the manner of some action

53
Q

Describe temporal adverbs

A

Describe the time that some action or event took place

54
Q

Describe auxiliary verbs

A

Words that mark certain semantic features of main verb including the tense, whether it’s completed and the mood of an action

55
Q

How are N grams and Markov Chains related?

A

N grams are just (N-1)th order Markov chains

56
Q

What is the transition probability?

A

The probability that a word has a given tag given the tag of the preceeding word

57
Q

What is the emission probability?

A

The probability of a word given a certain tag

58
Q

What are the two probabilities that HMMs take into account?

A

<ul>
<li>Transition probabilities</li>
<li>Emission probabilities</li>
</ul>

59
Q

Describe how HMMs are double stochastic models

A

<ul>
<li>Hidden markov process on the tags</li>
<li>Directly observable process on the words</li>
</ul>

60
Q

What are the two ways a HMM can be decoded?

A

<ul>
<li>Local decoding</li>
<li>Global decoding</li>
</ul>

61
Q

What is decoding of HMMs?

A

Optimising the posterior of the tags

62
Q

Loosely describe local decoding

A

Using marignal probabilities to consider each tag in isolation to maximise the posterior of the tags

63
Q

Loosely describe global decoding

A

More suitable for applications where the focus is the whole sequence and considers the surronding tags when trying to maximise the posterior

64
Q

From a given tag, how many ways can you reach the previous tags?

A

K, where K is the number of tags

65
Q

Why do we use log probabilities in our HMM algorithms?

A

Repeatedly taking product of probabilities will eventually underflow, especially if you are working with longer sentences

66
Q

What is a confusion matrix?

A

2 dimensional table showing the frequency that one tag is mistaken for another

CS5012 (12 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions