Scribd
Upload
18 views20 pages

Dependancy Parsing

Uploaded by

saisuraj1510
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
18 views20 pages

Dependancy Parsing

Uploaded by

saisuraj1510
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

Dependency

Parsing
Introduction

Syntactic structure of a sentence is described in terms of

✓ Words (or lemmas) in a sentence

✓ Directed binary grammatical relations that hold among the words.


Dependency Structure

✓Idea of dependency structure goes long way


✓ Panini grammar, 5th century BCB
✓ CFG - 1947 (R.S.Wells) 1953 (chemistry)
✓ DG – 1962 (Hays) Built dependency parses.
Typed dependency structure

✓ Relations among the words are illustrated with directed, labelled arcs from heads to
dependents.
✓ root node that explicitly marks the root of the tree the head of the entire structure.
✓ Labels are drawn from a fixed inventory of grammatical relations.
✓ head-dependent relationship : directly linking heads to the words that are immediately
dependent on them, bypassing the need for constituent structures.
Dependency Parsing
Typed dependency structure
Advantage of ✓ Deals morphologically rich and a relatively
free word order languages
Dependency ✓ A dependency-based approach need one link

Parsing type representing this particular relation.


Dependency vs Phrase structure
Head & Dependents

✓ Head of a phrase is the word that determines the syntactic category of that phrase.
✓ Head word of a constituent is the central organizing word of a larger constituent
✓ primary noun in a noun phrase, or verb in a verb phrase
✓ head of the noun phrase boiling hot water is the noun water.
✓ Other elements of the phrase modify the head, therefore they are head's dependents
✓ The remaining words in the constituent are either direct, or indirect, dependents of their
head.
✓ Binary relations that comprise these dependency
structures.
✓ The arguments to these relations consist of a head
and a dependent.
Binary Relations
✓ nsubjpass(submitted, Bills)
✓ auxpass(submitted, were)
Dependency ✓ agent(submitted, Brownback)
relation ✓ nn(Brownback, Senator)
✓ appos(Brownback, Republican)
✓ prep_of(Republican, Kansas)
✓ prep_on(Bills, ports)
✓ conj_and(ports, immigration)
✓ prep_on(Bills, immigration)
Stanford Dependency parser
Binary Relations
✓ nsubjpass(submitted, Bills)
✓ auxpass(submitted, were)
✓ agent(submitted, Brownback)
✓ nn(Brownback, Senator)
✓ appos(Brownback, Republican)
✓ prep_of(Republican, Kansas)
✓ prep_on(Bills, ports)
✓ conj_and(ports, immigration)
✓ prep_on(Bills, immigration)
Bills on ports and immigration were submitted by Senator Brownback, Republican of Kansas
✓ Dependent relation between predicates and their
arguments is useful for coreference resolution,
question answering and information
extraction.
Applications ✓ Constituent-based approaches to parsing provide
similar information
✓ It has to be distilled from the trees via techniques
head-finding rules
Relations
Example
Examples
1. Mark the head child of each node in a phrase
structure, using the appropriate head rules.
2. In the dependency structure, make the head of
each non-head child depend on the head of the
head-child.

Simple When a phrase-structure parse contains additional


information in the form of grammatical relations
Algorithm and function tags these tags can be used to label
the edges.
Limitation
✓ Limited by the information present in the
original constituent trees.
✓ Failure to integrate morphological information
with the phrase structure trees
Methods of dependency parsing
✓Dynamic programming (1996) O(𝑛3 )
✓Graph algorithms
Minimum spanning tree for a sentence
Neural graph based Parser (2017)
✓Constraint satisfaction 1990
✓ Transition based dependency parsing
✓Processes an input sentence and predicts a sequence of parsing actions in a
left-to-right manner.
Transition based dependency parsing
✓ Exact match (EM) — how many sentences are
parsed correctly.
✓ Pessimistic, with most sentences being
marked wrong.
✓ Metrics to be sensitive enough to tell if actual

Evaluation improvements are being made.


✓ labelled and unlabelled attachment
accuracy.
✓ Labelled attachment :
✓ proper assignment of a word to its head
along with the correct dependency relation.
✓ Unlabelled attachment :
✓ Correctness of the assigned head, ignoring
the dependency relation.
✓ Given a system output and a corresponding
reference parse, accuracy is
✓ Percentage of words in an input that
are assigned the correct head with the
Evaluation correct relation.
✓ Labelled attachment score (LAS)
✓ Unlabelled attachment score (UAS).
SPEECH and LANGUAGE
PROCESSING, An Introduction to Natural
Language Processing, Computational
References Linguistics, and Speech Recognition by
Daniel Jurafsky and James H. Martin
(Second Edition)

You might also like