Scribd
Upload
95 views22 pages

Tree Adjoining Grammar

Tree Adjoining Grammar (TAG) is a tree-generating grammar formalism where the basic elements are elementary trees that can be combined through substitution and adjunction operations. TAG sits between context-free and context-sensitive grammars in the Chomsky hierarchy. TAGs have initial trees and auxiliary trees; initial trees represent basic valency relations while auxiliary trees allow for recursion. Derived trees are built from combining elementary trees.

Uploaded by

subramani muthusamy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
95 views22 pages

Tree Adjoining Grammar

Tree Adjoining Grammar (TAG) is a tree-generating grammar formalism where the basic elements are elementary trees that can be combined through substitution and adjunction operations. TAG sits between context-free and context-sensitive grammars in the Chomsky hierarchy. TAGs have initial trees and auxiliary trees; initial trees represent basic valency relations while auxiliary trees allow for recursion. Derived trees are built from combining elementary trees.

Uploaded by

subramani muthusamy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 22

Tree Adjoining Grammar

Dr. VMS
Tree Adjoining Grammar (TAG)
 Tree Adjoining Grammar (TAG) is a formalism originally
proposed by A. K. Joshi, L. S. Levy, and M. Takahashi. in Tree
adjunct grammars. Journal Computer Systems Science, 10(1),
1975.
 Several variations on that formalism are developed, among
which we will be interested in lexicalized (LTAG) and
constraint-based (FTAG) versions.
 A TAG consists of a number of elementary trees, which can
be combined with a substitution, and an adjunction
operation.
A Tree Adjoining Grammar (TAG) is a 5-tu
ple G = {Σ,NT, I
 ,A, S}
 Σ is a finite set of terminal symbols.
 NT is a finite set of non terminal symbols.
 I is a finite set of finite trees called initial trees.
 A is a finite set of finite trees called auxiliary trees.
 S is a distinguished non-terminal symbol.
 The trees in I ∪ A are called elementary trees. The trees of
G are combined using adjunction and substitution.
Description
  Therules in a TAG are trees with a special leaf node known as the foot node,
which is anchored to a word.
 There are two types of basic trees in TAG: initial trees (often represented as '')
and auxiliary trees (''). Initial trees represent basic valency relations, while
auxiliary trees allow for recursion. Auxiliary trees have the root (top) node
and foot node labeled with the same symbol. A derivation starts with an
initial tree, combining via either substitution or adjunction.
 Substitution replaces a frontier node with another tree whose top node has
the same label.
 Adjunction inserts an auxiliary tree into the center of another tree. The
root/foot label of the auxiliary tree must match the label of the node at which
it adjoins.
 Other variants of TAG allow multi-component trees, trees with multiple foot
nodes, and other extensions.
Adjunction

 Stick an appropriate auxiliary tree fragment into the middle of a


tree
 Splits a node into two parts and sticks some material between
them
 Good for adding optional modifiers
 Good for long-distance dependencies
 Why insert into the middle?
 Because the insertion doesn’t affect the specified semantic relations
among nodes in the original fragment
Substitution
 The substitution operation appends one tree at a frontier node of
another tree. That is substitution takes place at non-terminal nodes
on the frontier side, which are marked for substitution by a down
arrow

 Stick an appropriate initial tree fragment at the bottom of a tree (to expand a
childless nonterminal node)

 Fills semantic slots of other words


 Get templates and idioms for free: “NP called NP up,” “NP kicked the bucket”
TAG
 A Tree Adjoining Grammar consists of a set of elementary
trees, divided in initial and auxiliary trees.

 These trees constitute the basic building blocks of the


formalism. Operations of adjunction and substitution are
defined which build derived trees from elementary trees.

 TAG thus constitute a tree-generating system, rather than a


string-generating systems as context-free grammar. The
string language of a TAG is defined as the yields of all the
trees derived by a TAG.
Initial and auxiliary trees

 An initial tree is a tree of which the interior nodes are all


labelled with non-terminal symbols, and the nodes on the
frontier are either labelled with terminal symbols, or with
non-terminal symbols, which are marked with the
substitution marker ( ).
 An auxiliary tree is defined as an initial tree, except that
exactly one of its frontier nodes must be marked as foot
node (`*'). The foot node must be labelled with a non-
terminal symbol which is the same as the label of the root
node.
Initial and auxiliary trees
 In initial tree , the np node is a substitution node, and left is a terminal
symbol associated with a frontier node. In auxiliary tree the foot node
is the leftmost daughter of the root node. with is a terminal symbol at a
frontier node, and the np node is a substitution node.
 Derived trees are built from initial and auxiliary trees by substitution
and adjunction. Substituting a tree in a tree simply replaces a
substitution node in with , under the convention that the non-terminal
symbol of the substitution node is the same as the root node of . For
example, substituting in gives the following tree:

Figure 4.7: Result of adjoining . Figure 4.8: Result of adjoining .
Lexicalised TAG (LTAG)
 There are several extensions of tree adjoining grammars being explored, such as lexicaltree
adjoining grammars (LTAGs). LTAGs consist of sets of initial and auxiliary trees whichcan
correspond to units in human language.
 A TAG is lexicalized if each elementary tree has at least one leaf with a terminal label.
 In lexicalized TAG, each elementary tree has at least one leaf with a terminal label.
• Computationally interesting: if the grammar is finite, the number of analyses for a sentence is
finite.
• Linguistically interesting: each lexical item can be associated with the set of syntactic
constructions it occurs in.
Tree Adjoined Grammars and Context Sensiti
vity
 
 Tree Adjoined Grammars fits in between the Context Free and
Context Sensitive classes in the Chomsky Hierarchy: they describe
Mildly Context-sensitive languages.
• Where as Context Free grammars have the pleasant property of
being parsable in polynomial space and time, they are not adequate
for describing for example the natural languages
Lexicalized TAG

 In LTAG, each elementary tree contains at least one frontier node labelled with a
terminal symbol. Thus each elementary tree is associated with at least one lexical
element. Note that in the example grammar above each of the elementary trees is
lexical; hence the grammar provides an example of a lexicalized TAG.
Derivation in TAG

 We now define by an example the notion of derivation in a TAG. Unlike


CFGs, the tree obtained by derivation (the derived tree) does not give enough
information to determine how it was constructed. The derivation tree is an
object that specifies uniquely how a derived tree was constructed. Both
operations, adjunction and substitution, are considered in a TAG derivation.
Take for example the derived tree a5 in Figure 3; a5 yields the sentence
yesterday a man saw May. It has been built with the elementary trees shown
in Figure 4.
Lexicalized Grammar

A grammar is 'lexicalized' if it consists of: a finite set of structures each associated


with a lexical item; each lexical item will be called the anchor of the corresponding
structure; an operation or operations for composing the structures.
Thank You

You might also like