Docsity
Docsity
Prepare for your exams
Prepare for your exams

Study with the several resources on Docsity

Earn points to download
Earn points to download

Earn points by helping other students or get them with a premium plan

Guidelines and tips
Guidelines and tips
Sell on Docsity
Sell on Docsity
Docsity AI
Docsity AI
Log inSign up

Prepare for your exams

Study with the several resources on Docsity

Find documents

Prepare for your exams with the study notes shared by other students like you on Docsity

Search for your university

Find the specific documents for your university's exams

Docsity AINEW

Summarize your documents, ask them questions, convert them into quizzes and concept maps

Explore questions

Clear up your doubts by reading the answers to questions asked by your fellow students

Earn points to download

Earn points by helping other students or get them with a premium plan

Share documents
download point icon20 Points

For each uploaded document

Answer questions
download point icon5 Points

For each given answer (max 1 per day)

All the ways to get free points
Get points immediately

Choose a premium plan with all the points you need

Study Opportunities

Choose your next study program

Get in touch with the best universities in the world. Search through thousands of universities and official partners

Community

Ask the community

Ask the community for help and clear up your study doubts

Free resources

Our save-the-student-ebooks!

Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors

Natural Language Semantics, Lecture notes of Logic

Washington State University (WSU or Wazzu)Logic

– This assumption is not always true, especially for words/phrases which have idiomatic/non-literal meaning. • e.g. “It's raining cats and dogs”. 3. Semantic ...

Typology: Lecture notes

2022/2023

Uploaded on 03/01/2023

lalitlallit
lalitlallit 🇺🇸

4.1

(10)

226 documents

1 / 8

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
1
1
CSC 594 Topics in AI –
Applied Natural Language Processing
Fall 2009/2010
5. Semantics
2
Natural Language Semantics
• Semantics of a sentence is the meani ng of the sentence.
• And the meaning comes from:
– meanings of the words/phrases in the sentence; plus
– (semantic) relations between the words/phrases
• Underlying assumption – Compositionality
–The meaning of a sentence is a composition of the meanings of
its parts.
– This assumption is not always true, especially for words/phrases
which have idiomatic/non-literal meaning.
• e.g. “It’s raining cats and dogs”
3
Semantic Analysis
Semantic Analysis
Derive the meaning of a sentence.
Derive the meaning of a sentence.
Often applied to the result of syntactic analysis.
Often applied to the result of syntactic analysis.
“
“John
John ate
ate the cake
the cake.
.”
”
NP V NP
NP V NP
((action INGEST) ; syntactic verb
((action INGEST) ; syntactic verb
(actor
(actor JOHN
JOHN-
-01) ; syntactic
01) ; syntactic subj
subj
(object FOOD)) ; syntactic
(object FOOD)) ; syntactic obj
obj
To do semantic analysis, we need a (semantic)
To do semantic analysis, we need a (semantic)
dictionary (e.g.
dictionary (e.g. WordNet
WordNet,
,
http://
http://www.cogsci.princeton.edu/~wn
www.cogsci.princeton.edu/~wn/
/).
).
pf3
pf4
pf5
pf8

Partial preview of the text

Download Natural Language Semantics and more Lecture notes Logic in PDF only on Docsity!

CSC 594 Topics in AI –

Applied Natural Language Processing

Fall 2009/

5. Semantics

Natural Language Semantics

  • Semantics of a sentence is the meaning of the sentence.
  • And the meaning comes from:
    • meanings of the words/phrases in the sentence; plus
    • (semantic) relations between the words/phrases
  • Underlying assumption – Compositionality
    • The meaning of a sentence is a composition of the meanings of

its parts.

  • This assumption is not always true, especially for words/phrases

which have idiomatic/non-literal meaning.

  • e.g. “It’s raining cats and dogs”

Semantic Analysis Semantic Analysis

„„^ Derive the meaning of a sentence.Derive the meaning of a sentence.

„„ Often applied to the result of syntactic analysis.Often applied to the result of syntactic analysis.

““JohnJohn ateate the cakethe cake..””

NPNP VV NPNP

((action((action INGEST)INGEST) ; syntactic verb; syntactic verb

(actor(actor JOHN-JOHN-01)01) ; syntactic; syntactic subjsubj

(object(object FOOD))FOOD)) ; syntactic; syntactic objobj

„„ To do semantic analysis, we need a (semantic)To do semantic analysis, we need a (semantic)

dictionary (e.g.dictionary (e.g. WordNetWordNet,,

http://http://www.cogsci.princeton.edu/~wnwww.cogsci.princeton.edu/~wn//).).

Various Semantic Analyses

  • There are many aspects and levels in the meaning of a

sentence:

  • Word meanings/senses, concepts
  • Semantic relations
  • Quantifiers (e.g. “Every boy loves a dog”)
  • Truth value in a model
  • Inference
    • e.g. entailment – “He was snoring” entails “He was sleeping”
  • Various representations of meaning
  • First-order Predicate Logic (FOPL)
  • Slot-filler structures or frames
  • Semantic networks
  • Web Ontology Language (OWL) Í new
  • etc.

Word Senses

  • Many words have more than one sense ( ambiguous words)
    • an applicable sense varies depending on the context.
  • A word may also have several parts-of-speech.

Ontology

  • Word senses (not words) can be grouped into a set of

broad classes – semantic concepts.

  • The set of semantic concepts is called ontology.
  • Typical major classes:
    • substance, quantity, quality, relation, place, time, position, state,

action, affection, event, situation

  • For each class, there are many sub-classes, which are

often organized in a hierarchy.

e.g. WordNet

  • Relations between actions/events (as used in WordNet)

Source: Jurafsky & Martin “Speech and Language Processing”

Thematic Roles (1)

  • Consider the following examples:
    • John broke the window with the hammer.
    • The hammer broke the window.
    • The window broke.
  • Although in each sentence the surface position of the

three words are different, they play the same semantic

roles (deep roles; thematic roles)

  • John – AGENT
  • the window – THEME (or OBJECT)
  • the hammer -- INSTRUMENT

Thematic Roles (2)

  • There are several commonly used thematic roles:

Source: Jurafsky & Martin “Speech and Language Processing”

Some prototypical examples of various thematic roles

Source: Jurafsky & Martin “Speech and Language Processing”

But it’s difficult to come up with a standard set of roles or to define them.

Linking Syntax and Semantics

  • Same thematic role could be realized in various syntactic

positions.

  • John broke the window with the hammer.
  • The hammer broke the window.
  • The window broke.
  • To link syntax and semantics to derive a semantic

representation of a sentence, one common approach is

by Lexical Semantics – encode the semantic

constraints which a word imposes upon other words in

specific relations ( arguments ).

  • [BREAK [AGENT animate] [THEME inanimate] [INSTR utensil] ]
  • [BREAK [AGENT utensil] [THEME inanimate] ]
  • [BREAK [AGENT inanimate] ]

Selectional Restrictions

  • Specifications of the legal combinations of senses that

can co-occur are called Selectional Restrictions –

constraints which a predicate imposes on the semantic

types of its arguments.

  • e.g. [READ [AGENT person] [THEME text] ]
  • Semantic constraints (on nouns) specify the most

general types – all sub-types/hyponyms in the ontology

are legal.

thing

animate inanimate

person animal

child adult

text

book magazine

  • Unification is the operation for

* combining information (concatenation)

* checking compatibility

  • Subsumption is an ordering on feature structures
foo
w
bar
cat
F
h
f
h
h
cat
F
cat
D
foo
g
h
h
cat
F
cat
D
foo
g

w

bar
foo
f

g

foo
cat
F f

g

foo
cat
F f h
bar
p g
foo

cat

F f h
bar
p

(define-word (cat) = V (word) = “ate” (head tense) = past (head sem cat) = EAT (head sem actor) = (head subj sem) (head sem object) = (head dobj sem))

head

object cat EAT

cat

V

subj sem actorsem

dobj

sem

tense past

word

“ate”

cat

EAT

actor object instru- ment cat

cat

cat FOOD

ANIMATE
UTENSIL

(define-semcat (cat) = EAT (actor cat) = ANIMATE (object cat) = FOOD (instrument cat) = UTENSIL)

head

object cat

EAT

cat

V

subj sem actorsem

dobj

sem

tense past

word

“ate”

FOOD

cat

cat

ANIMATE

UTENSIL

instru- ment

cat

Syntax/Semantics Integrated Processing

h 1 h

2 cat E

1 cat D

cat C

Grammar

word " d"

cat D

r1:

r2:

cat

cat

C
E

cat

D

hh

“d”

cat

cat

C
E

cat

D

word

hh

τ i^ ⇒ τ j

Derivation using Unification Grammar

R

1 headagr 2 headagr
headsubj 1 head
head 2 head
2 cat VP
1 cat NP
cat S

head 1 head

1 catV

catVG

R 1 ,

headagr 3S

word"Mary"

catNP

R 5 = "Mary" ,

Grammar

headagr 3S

word"John"

catNP

R 3 = "John" ,

headtype trans

headdobj 2 head

head 1 head

2 catNP

1 catVG

catVP

R 2 ,

headtype trans

headagr3S

word"likes"

catV

R 4 = "likes" ,