18CSE359T - NATURAL LANGUAGE PROCESSING UNIT 2 & 3 - 12M

I HAVE TAKEN ANSWERS FOR 2 UNIT 2 QUESTIONS AND 2 UNIT 3 QUESTIONS

12M:

List and explain with suitable examples about various semantic relationships at the word level

Semantic relationships:

• Semantic relationships are the associations that exist

• Between the meanings of words

• Between the meanings of phrases

• Between the meanings of sentences

Semantic relationships at word level:

• Synonymy

• Antonymy

• Homonymy

• Polysemy 

• Metonymy 

Synonymy: 

• Semantic relationship that exists between two or more words that have the same meaning and belong to the same part of speech, but are spelled differently

• Example: big - large, fast - quickly, begin - start

Antonymy: 

• Semantic relationship that exists between two or more words that have the opposite meaning

• The semantic feature that they do not share is present in one member of the pair and absent in the other

• Types of antonymy:

• Complementary or contradictory antonymy

• Relational antonyms

• Gradable or scalar antonyms

Homonymy:

• Relationships that exist between two (or more) words which belong to the same grammatical category, have the same spelling, may or may not have the same pronunciation, but have different meanings and origins. 

• Eg: 

• To lie (rest) - to lie (not to tell truth)

• Bank (financial institution) - bank (side of river)

Hyponymy:

• Hyponymy is the semantic relationship that exists between two or more words in such a way that the meaning of one word includes the meaning of other words.

 

Polysemy: 

• Semantic relationship that exists between a word and its multiple conceptually and historically related meanings 

• E.g:

• foot = 1. part of the body; 2. lower part of something 

• plain = 1. clear; 2. unadorned; 3. obvious. 

Metonymy:

• Metonymy is the semantic relationship that exists between two words in which one of the words is metaphorically used in place of the other word in particular contexts to convey the same meaning  

• Eg: 

• brass = military officers 

• cops = policemen 

• crown = monarchy 

Discuss on coreference resolution

Coreference resolution:

• Coreference resolution (CR) is the task of finding all linguistic expressions in a given text that refer to the same real-world entity. 

• Coreference resolution is an exceptionally versatile tool and can be applied to a variety of NLP tasks

Types of references: 

• Anaphora and cataphora

• Split antecedents

• Co-referring noun phrases

• Presuppositions / bound variable

Anaphora: 

• Refers to the use of a pronoun or word that refers back to a previously mentioned noun or phrase in a sentence or text

• The word occurring before an anaphora is called antecedent

Cataphora: 

• Refers to the use of a pronoun or word that refers to a subsequent noun or phrase that appears later in a sentence or text

• The word occurring before an cataphora is called postcedent

Split antecedents:

• It’s an anaphoric expression where the pronoun (2) refers to more than one antecedent (1).

Co-referring noun phrases:

• It’s also an anaphoric example of a situation in which the second noun phrase (2) is a reference to an earlier descriptive form of an expression (1).

Presuppositions / bound variable:

• Some argue whether presupposition can be classified as a coreference resolution type. That’s because a pronoun (2) is not exactly referential – in the sense that we can’t replace it with the quantified expression (1). However, after all the pronoun is a variable that is bound by its antecedent

Steps for coreference resolution:

STEP 1:

• The first step in order to apply coreference resolution is to decide whether we would like to work with single words/tokens or spans. 

• Span is most often the case that what we want to swap or what we are swapping for is not a single word but multiple adjacent tokens. Therefore span is a whole expression.

Example: 

STEP 2:

• The next step is to combine the spans into groups.

• Combining items is referred to as clustering or grouping. It is a method of taking arbitrary objects and grouping them together into clusters/groups within which these items share a common theme.

STEP 3:

• The resulting groups are [Sam, his, he, him] as well as [a white star, it]. Notice that “Sam” and “a white star” are marked as entities. This is a crucial step in coreference resolution. 

• We need to not only identify similar spans but also determine which one of them is, often referred to as, the real-world entity.

Explain any 3 types of syntactic disambiguation

Syntactic disambiguation:

• Selection of correct parse tree from various parse trees by the natural language processing system is known as syntactic disambiguation

• Types: 

• Attachment problem

• Gap finding and filling

• Analytical ambiguities

• The interaction between categorical and structural ambiguity

• Structural ambiguity as a closure problem

Gap finding and filling:

• Gap-finding ambiguities occur when a moved constituent has to be returned to its pre-transformational starting point, and there is more than one place that it might go. For example:

• Eg: Those are the boys that the police debated about fighting

• Taking the first gap gives the meaning that the police debated with the boys on the topic of fighting. The second gives the police debated (among themselves) about fighting the boys. 

The interaction between categorical and structural ambiguity:

• If a word is categorically ambiguous, a sentence containing it can be structurally ambiguous, and the possibilities will correspond to those for the word. 

• For example:

• The Japanese push bottles up the Chinese.

• The words push and bottle could be verb and noun respectively, or noun and verb; the writer intended the latter, though there is a strong preference for the former.

Structural ambiguity as a closure problem:

• Another way to look at many structural ambiguities is to view them as CLOSURE PROBLEMS. 

• In parsing, a constituent of the parse tree is said to be OPEN if it has not been declared complete, and so other constituents may still be attached to it. 

• When a constituent is complete, it is CLOSED, and that subtree may no longer be changed. 

• Eg: 

Explain in detail the technique behind automatically assigning the coherence in a given set of sentences. give a suitable example

Understanding Coherence in Discourse:

• Coherence is a primary characteristic of discourse, ensuring that the sentences or utterances "hang together" and make sense collectively. 

• It involves establishing meaningful connections, known as coherence relations, between the components of the text.

Techniques to Determine Coherence

1. Discourse Structure and Segmentation

• Discourse Segments: Each sentence in a discourse is treated as a segment.

• Combining Segments: Smaller segments (sentences) can be combined into larger segments if a coherence relation can be established between them.

2. Coherence Relations

• Subordinating Relations: These involve one segment being more central (nucleus) and the other providing additional, supporting information (satellite). For example, in the sentence "John won the contest but he does not sing well," "John won the contest" is the nucleus, and "he does not sing well" is the satellite.

• Coordinating Relations: These involve segments that are equally important and provide parallel or contrasting information. 

3. Reference Relations

• Coreference or Anaphora Resolution: This technique determines which entity a referring expression (like a pronoun) refers to, ensuring that the text maintains coherence by correctly linking references.

 4. Discourse Relations

• Informational Relations: These convey semantic relationships such as cause, condition, and temporal connections between segments.

• Intentional Relations: These specify the intended effects of the discourse, like explaining why a statement was made.

Example of Coherence in Discourse

Consider the following set of sentences:

1. John went to the bank to deposit his paycheck.

2. He then took a train to Bill’s car dealership.

3. He needed to buy a car.

4. The company he works for now isn’t near any public transportation.

5. He also wanted to talk to Bill about their softball league.

Practical Example Using RST

Using RST, the coherence of our example can be visualized as follows:

- Nucleus: John went to the bank to deposit his paycheck.

- Nucleus: He then took a train to Bill’s car dealership.

- Nucleus: He needed to buy a car.

- Satellite: The company he works for now isn’t near any public transportation.

- Satellite: He also wanted to talk to Bill about their softball league.

Methods to Automatically Assign Coherence:

Automatic coherence evaluation involves several computational techniques, often employing machine learning models trained on large corpora of text. Some methods include:

1. Lexical Cohesion

• Lexical Chains

2. Surface Features

• Part-of-Speech Tagging.

• N-gram Models

3. Discourse Parsing

• Rhetorical Structure Theory (RST)

4. Machine Learning Models

• Supervised Learning