You are on page 1of 14

Chapter – 6

Disambiguation

Department of Computer Science


School of Computing
Dire Dawa Institute of Technology
Dire Dawa University

Te ssf u Ge t e ye (Ph D)

2020/2021-Sem e st er- II
Introduction
Morphological
Ambiguo us Text
Level Syntactic
Ambiguity Resolution
Level Semantic
Level Discourse
Level
Introduction: Ambiguous Text

 A text is said to be ambiguous if multiple or alternative linguistic structures can be built


for it.

 For example, given the following lexical entry in a lexicon.

duck
V: to move (head or body) quickly downwards or
away N: waterfowl
N: a heavy cotton fabric used for clothing and
tents

 We can see that at least the following linguistic meanings can be derived from
the sentence “I made her duck”.

(1) I cooked waterfowl for her


(2) I cooked waterfowl belonging to her
(3) I created the fabric she owns
(4) I caused her to quickly lower her head or body

Department of Computer Science, SC, DDIT, DDU Disambiguation 2/14


Introduction
Morphological
Ambiguo us Text
Level Syntactic
Ambiguity Resolution
Level Semantic
Level Discourse
Level
Introduction: Ambiguity Resolution

 Ambiguity may occur at:


 Phonological level - multiple orthographic representations
 Morphological level - multiple word classes
 Syntactic level - different ways to parse the tree
 Semantic level - different meanings of the same parse tree
 Discourse level - different references of the same anaphora
 Pragmatic level - cannot be clearly interpreted

 Disambiguation (Ambiguity Resolution) refers to the resolution of ambiguities that


occur at different levels of linguistic analysis.

 Most tasks in language processing can be viewed as resolving ambiguity at one of


these levels.

 Ambiguity resolution at each level of the linguistic analysis is required by almost all NLP
applications.

Department of Computer Science, SC, DDIT, DDU Disambiguation 3/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Morphological Level: Sources of Ambiguity

 Morphological ambiguity occurs due to multiple word classes.

 For example, consider the following cases.

(i)
duck: verb, noun
her: dative pronoun,
possessive pronoun
“I made her duck” can
mean:

(1) I cooked
waterfowl for her
(2) I cooked
waterfowl
belonging to her
(3) I created the
fabric she owns
(4) I caused her to
Department of Computer Science, SC, DDIT, DDU Disambiguation 4/14
Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Morphological Level: Sources of Ambiguity

(ii)

አበበች: noun, verb


“ጽጌረዳዋ አበበች” can mean:
(1) The rose plant flowered
(blossomed)
(2) The beautiful Abebech

 Ambiguities (1) and (2) arise at morphological level.

Department of Computer Science, SC, DDIT, DDU Disambiguation 5/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Morphological Level: Ambiguity Resolution

 Ambiguity at morphological level can be resolved by


 Part-of-Speech Tagger (POST)
 Parsing (Syntactic Analysis)

 For example

(i) “I made her duck” can be tagged as:


I (N) made (V) her (dative pro.) duck (N)  (1)

I (N) made (V) her (possessive pro.) duck (N)  (2)

I (N) made (V) her (N) duck (V)  (4)

(1) I cooked waterfowl for her


(2) I cooked waterfowl belonging to her
(3) I created the fabric she owns
(4) I caused her to quickly lower her head or
body

Department of Computer Science, SC, DDIT, DDU Disambiguation 6/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Morphological Level: Ambiguity Resolution

(i) “ጽጌረዳዋ አበበች ” can be tagged as:

ጽጌረዳዋ (Definite N) አበበች ( V )  (1)

ጽጌረዳዋ (Modifier, N) አበበች (N)  (2)

(1) The rose plant flowered


(blossomed)
(2) The beautiful Abebech

Department of Computer Science, SC, DDIT, DDU Disambiguation 7/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Syntactic Level: Sources of Ambiguity

 Syntactic ambiguity occurs when there are multiple ways to parse a sentence.

 For example, consider the following parse trees for “I made her duck”.

(i) S (ii) S

NP VP NP VP

I V NP VP I V NP

made her V made Det N

duck her duck


 Ambiguities (1) and (4) arise as a result of syntactic ambiguity.

Department of Computer Science, SC, DDIT, DDU Disambiguation 8/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Syntactic Level: Ambiguity Resolution

 Syntactic ambiguity can be resolved by


 Probabilistic Parsing
 Semantic Analysis

 For example, the following parsing result resolves “I made her duck” to mean (4).
S
(1) I cooked waterfowl for her
(2) I cooked waterfowl belonging to her
NP VP (3) I created the fabric she owns
(4) I caused her to quickly lower her head or
I V NP VP body

made her V

duck

Department of Computer Science, SC, DDIT, DDU Disambiguation 9/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Semantic Level: Sources of Ambiguity

 Semantic ambiguity occurs when there are different meanings for the same parse tree.

 For example, consider the following parse tree for “I made her duck”.
S
(1) I cooked waterfowl for her
(2) I cooked waterfowl belonging to
NP VP (3) her
I created the fabric she owns
(4) I caused her to quickly lower her head or
I body
V NP
cooked
made
N created
made Det

her duck

 Ambiguities (2) and (3) arise as a result of semantic ambiguity.

Department of Computer Science, SC, DDIT, DDU Disambiguation 10/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Semantic Level: Ambiguity Resolution

 Semantic ambiguity can be resolved by


 Word Sense Disambiguation (WSD)
 Discourse Analysis

 For example, consider the following sentences.

(1) I went to the bank to deposit money.

(2) I went to the bank to look at the river.

(3) I went to the bank. Unfortunately, it was closed.

(4) I went to the bank. Unfortunately, it was flooded due to


heavy rainfall.

 POST and WSD can be supported by a lexicon.

Department of Computer Science, SC, DDIT, DDU Disambiguation 11/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Discourse Level: Sources of Ambiguity

 Discourse ambiguity occurs due to different references of the same anaphora.

 Consider the following examples.

(1) The city councilors refused the demonstrators a


permit because they feared violence.

(2) The city councilors refused the demonstrators a


permit because they advocated revolution.

 The ambiguities in (1) and (2) arise because “they” may refer to either


“The city councilors”; or


“the demonstrators”.

Department of Computer Science, SC, DDIT, DDU Disambiguation 12/14


Introduction
Morphological
Sources of Ambiguity
Level Syntactic
Ambiguity
Level Semantic
Resolution
Level Discourse
Level
Discourse Level: Ambiguity Resolution

 Discourse ambiguity can be resolved by


 Pragmatic Analysis

 For example, consider the following background knowledge.

City councilors are members of the governing body who would


like to see their city free of any violence. They know that
violence could undermine their mission, vision and goals. They
also want to stay in power and do not like any revolution.

City residents usually feel that there is no governing body who


listens to their quest for change in city administration. They do
not have any other means except staging a demonstration and make
their demand fulfilled.

 With this background knowledge “they” refers to “The city councilors” in (1)
and “the demonstrators” in (2)

 A text is finally concluded as ambiguous is it can not be resolved by Pragmatic


Analysis.

Department of Computer Science, SC, DDIT, DDU Disambiguation 13/14


TOC: Course Syllabus

Previous: Discourse and Pragmatic Processing

Current: Disambiguation

Next: Approaches to NLP

You might also like