Transition Network Grammars for Natural Language Analysis W. A. Woods

1
Transition Network Grammars for Natural Language
Analysis - W. A. Woods

• In-Su Yoon
• Pusan National University
• School of Electrical and Computer Engineering

Pusan National University
2
Whats a grammar?

• A grammar contains the knowledge about legal
  syntactic structure, represented as rewrite
  rules.
• The grammar defines the language.

Pusan National University
3
Heres a very simple example

• S ? NP VP
• VP ? VERB NP
• NP ? ART NOUN
• NP ? POSS NOUN

• My dog ate a frog.
• NOUN ? frog dog
• ART ? a
• VERB ? ate
• POSS ? my

Pusan National University
4
Parse tree
The tree notation is difficult to compute with
directly, so we can convert the representation
into more useful (S (NP (POSS my)
(NOUN dog)) (VP (VERB ate) (NP
(ART a) (NOUN frog))))
Pusan National University
5
From grammars to transition nets

• As grammars get more rules, they become difficult
  to understand and computationally more demanding.
• We can make our jobs easier by converting the
  grammar to a more convenient representation known
  as a FSM or transition network (TN).

Pusan National University
6
Transition Network
NP
VP
S0
S1
S2
VERB
NP
VP0
VP1
VP2
ART
ART
NOUN
NP0
NP1
NP2
NOUN
NP0
NP1
NP2
POSS
NOUN
NP3
NP4
NP5
POSS
When the lexicon gets really big, drawing them
takes forever!
Pusan National University
7
Why do we want recursive rules in our grammar?

• Natural languages allow us to express an infinite
  range of ideas using a finite set of rules and
  symbols.
• The boy drowned.
• The boy with the raft drowned.
• The boy with the raft near the island drowned.
• The boy with the raft near the island in the
  ocean drowned.
• The boy ....

Pusan National University
8
From transition nets to recursive transition nets

• The ability to push a destination on a stack,
  jump to a subnetwork, and return to the pushed
  destination.

Pusan National University
9
A sample recursive transition network(1)
The boy broke the window with a hammer.
Pusan National University
10
A sample recursive transition network(2)
Pusan National University
11
RTNs are not enough

• RTNs are interesting from a theoretical
  standpoint but are not of much use themselves.
• From a computational perspective, a black box
  that accepts English input and just say yes or
  no doesnt buy us much.
• What we need is a black box that records the
  structure of the input as well as providing an
  evaluation of syntactic correctness.

Pusan National University
12
Augmented Transition Network (ATN) (1)

• Adding procedures to the arcs of the RTN.
• These procedures are then performed when the
  corresponding arcs are traversed.
• The resulting network is called an augmented
  transition network (ATN).

Pusan National University
13
Augmented Transition Network (ATN) (2)

• One of the things we can do with these added
  procedures or augmentations is to store
  information in registers when arcs are traversed.
• To record the structure of the input, we add an
  action to each arc which stores the word that was
  processed while traversing that arc in an
  appropriate register.

Pusan National University
14
Register assignment (1)

• If we now process the noun phrase
• the vicious dog with the ATN.
• Well have made the following register
  assignments
• ART the
• ADJS vicious
• NOUN dog

Pusan National University
15
Register assignment (2)

• When we take the pop arc from this network, we
  can then accumulate the register contents into a
  larger structure that we call NP
• (NP (ART the)
• (ADJS vicious)
• (NOUN dog))

Pusan National University
16
the vicious dog ate the wimpy frog

• the resulting structure would be
• (S (SUBJ (NP (ART the)
• (ADJS vicious)
• (NOUN dog)))
• (VERB ate)
• (OBJ (NP (ART the)
• (ADJS wimpy)
• (NOUN frog))))

Pusan National University
17
Conclusion

• The great majority of natural language processing
  (NLP) systems in actual use in the world today
  are based on the ATN formalism.
• They are often used as front-ends or interfaces
  to database systems for question answering.
• One often cited example of a system that followed
  this approach successfully is a program called
  LUNAR, written by William Woods.
• They are often designed to perform both the
  syntactic and semantic analysis at the same time.

Pusan National University