Learning for Semantic Parsing Using
Statistical Syntactic Parsing
Techniques
Ruifang Ge
Ph.D. Final Defense
Supervisor: Raymond J. Mooney
Machine Learning Group
Department of Computer Science
The University of Texas at Austin
1
Semantic Parsing

Semantic Parsing: Transforming natural
language (NL) sentences into completely
formal meaning representations (MRs)

Sample application domains where MRs
are directly executable by another
computer system to perform some task


CLang: Robocup Coach Language
Geoquery: A Database Query Application
2
CLang (RoboCup Coach Language)


Coach
CLang
In RoboCup Coach competition, teams compete to coach
simulated players
The coaching instructions are given in a formal language
called CLang
If our player 2 has the
ball, then position our
player 5 in the
midfield.
Semantic Parsing
Simulated soccer field
((bowner (player our {2}))
(do (player our {5}) (pos (midfield))))
3
GeoQuery: A Database Query Application

Query application for U.S. geography
database [Zelle & Mooney, 1996]
User
What are the
rivers in Texas?
Semantic Parsing
Query answer(x1, (river(x1), loc(x1,x2),
equal(x2,stateid(texas))))
Angelina,
Blanco, …
DataBase
4
Motivation for Semantic
Parsing


Theoretically, it answers the question of how
people interpret language
Practical applications




Question answering
Natural language interface
Knowledge acquisition
Reasoning
5
Motivating Example
If our player 2 has the ball, our player 4 should stay in our half
((bowner (player our {2})) (do our {4} (pos (half our))))
Semantic parsing is a compositional process.
Sentence structures are needed for building meaning representations.
bowner: ball owner
pos: position
6
Syntax-Based Approaches

Meaning composition follows the tree structure of a
syntactic parse

Composing the meaning of a constituent from the
meanings of its sub-constituents in a syntactic parse

Hand-built approaches (Woods, 1970, Warren and Pereira,
1982)

Learned approaches
sentences

Miller et al. (1996): Conceptually simple

Zettlemoyer & Collins (2005)): hand-built Combinatory Categorial
Grammar (CCG) template rules
7
Example
MR: bowner(player(our,2))
S
NP
PRP$
NN
our
player
VP
CD
2
NP
VB
has
DT
NN
the
ball
Use the structure of a syntactic parse
8
Example
MR: bowner(player(our,2))
S
NP
PRP$-our
our
VP
NN-player(_,_) CD-2
player
2
NP
VB-bowner(_)
has
DT-null
NN-null
the
ball
Assign semantic concepts to words
9
Example
MR: bowner(player(our,2))
S
NP-player(our,2)
PRP$-our
our
VP
NN-player(_,_) CD-2
player
2
NP
VB-bowner(_)
has
DT-null
NN-null
the
ball
Compose meaning for the internal nodes
10
Example
MR: bowner(player(our,2))
S
VP-bowner(_)
NP-player(our,2)
PRP$-our
our
NN-player(_,_) CD-2
player
2
VB-bowner(_)
has
NP-null
DT-null
NN-null
the
ball
Compose meaning for the internal nodes
11
Example
MR: bowner(player(our,2))
S-bowner(player(our,2))
NP-player(our,2)
PRP$-our
our
VP-bowner(_)
NN-player(_,_) CD-2
player
2
NP-null
VB-bowner(_)
has
DT-null
NN-null
the
ball
Compose meaning for the internal nodes
12
Semantic Grammars

Non-terminals in a semantic grammar
correspond to semantic concepts in
application domains
Hand-built approaches (Hendrix et al., 1978)

Learned approaches


Tang & Mooney (2001), Kate & Mooney (2006), Wong &
Mooney (2006)
13
Example
MR: bowner(player(our,2))
bowner
player
has
our
our
the
ball
2
player
2
bowner → player has the ball
14
Thesis Contributions

Introduce two novel syntax-based
approaches to semantic parsing


Theoretically well-founded in computational
semantics (Blackburn and Bos, 2005)
Great opportunity: leverage the significant
progress made in statistical syntactic parsing
for semantic parsing (Collins, 1997; Charniak and
Johnson, 2005; Huang, 2008)
15
Thesis Contributions



SCISSOR: a novel integrated syntacticsemantic parser
SYNSEM: exploits an existing syntactic parser
to produce disambiguated parse trees that
drive the compositional meaning composition
Investigate when the knowledge of syntax
can help
16
Representing Semantic Knowledge in
Meaning Representation Language Grammar
(MRLG)
 Assumes a meaning representation language (MRL)
is defined by an unambiguous context-free grammar.
 Each production rule introduces a single predicate
in the MRL.
 The parse of a MR gives its predicate-argument
structure.
Production
Predicate
CONDITION →(bowner PLAYER)
P_BOWNER
PLAYER →(player TEAM {UNUM})
P_PLAYER
UNUM → 2
P_UNUM
TEAM → our
P_OUR
17
Roadmap




SCISSOR
SYNSEM
Future Work
Conclusions
18
SCISSOR

Semantic Composition that Integrates Syntax
and Semantics to get Optimal Representations

Integrated syntactic-semantic parsing


Allows both syntax and semantics to be used
simultaneously to obtain an accurate combined
syntactic-semantic analysis
A statistical parser is used to generate a
semantically augmented parse tree (SAPT)
19
Syntactic Parse
S
NP
VP
PRP$
NN
CD
VB
our
player
2
has
NP
DT
NN
the
ball
20
SAPT
S-P_BOWNER
NP-P_PLAYER
VP-P_BOWNER
PRP$-P_OUR NN-P_PLAYER CD- P_UNUM VB-P_BOWNER
our
player
2
has
NP-NULL
DT-NULL NN-NULL
the
ball
Non-terminals now have both syntactic and semantic labels
Semantic labels: dominate predicates in the sub-trees
21
SAPT
S-P_BOWNER
NP-P_PLAYER
VP-P_BOWNER
PRP$-P_OUR NN-P_PLAYER CD- P_UNUM VB-P_BOWNER
our
player
2
has
NP-NULL
DT-NULL NN-NULL
the
ball
MR: P_BOWNER(P_PLAYER(P_OUR,P_UNUM))
22
SCISSOR Overview
SAPT Training Examples
learner
Integrated Semantic Parser
TRAINING
23
SCISSOR Overview
NL Sentence
Integrated Semantic Parser
SAPT
TESTING
Compose MR
MR
24
Extending Collins’ (1997) Syntactic
Parsing Model



Find a SAPT with the maximum probability
A lexicalized head-driven syntactic parsing
model
Extending the parsing model to generate
semantic labels simultaneously with syntactic
labels
25
Why Extending Collins’ (1997) Syntactic
Parsing Model

Suitable for incorporating semantic
knowledge



Head dependency: predicate-argument relation
Syntactic subcategorization: a set of arguments
that a predicate appears with
Bikel (2004) implementation: easily
extendable
26
Parser Implementation



Supervised training on annotated SAPTs is
just frequency counting
Testing: a variant of standard CKY chartparsing algorithm
Details in the thesis
27
Smoothing



Each label in SAPT is the combination of a
syntactic label and a semantic label
Increases data sparsity
Break the parameters down
Ph(H | P, w)
= Ph(Hsyn, Hsem | P, w)
= Ph(Hsyn | P, w) × Ph(Hsem | P, w, Hsyn)
28
Experimental Corpora

CLang (Kate, Wong & Mooney, 2005)



300 pieces of coaching advice
22.52 words per sentence
Geoquery (Zelle & Mooney, 1996)



880 queries on a geography database
7.48 word per sentence
MRL: Prolog and FunQL
29
Prolog vs. FunQL (Wong, 2007)
What are the rivers in Texas?
Prolog:
X1: river; x2: texas
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
FunQL:
answer(river(loc_2(stateid(texas))))
Logical forms: widely used as MRLs in
computational semantics, support reasoning
30
Prolog vs. FunQL (Wong, 2007)
What are the rivers in Texas?
Flexible order
Prolog:
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
FunQL:
answer(river(loc_2(stateid(texas))))
Strict order
Better generalization on Prolog
31
Experimental Methodology


standard 10-fold cross validation
Correctness



CLang: exactly matches the correct MR
Geoquery: retrieves the same answers as the
correct MR
Metrics



Precision: % of the returned MRs that are correct
Recall: % of NLs with their MRs correctly returned
F-measure: harmonic mean of precision and recall
32
Compared Systems

COCKTAIL (Tang & Mooney, 2001)


WASP (Wong & Mooney, 2006)


Semantic grammar, string kernels
Z&C (Zettleymoyer & Collins, 2007)


Semantic grammar, machine translation
KRISP (Kate & Mooney, 2006)


Deterministic, inductive logic programming
Syntax-based, combinatory categorial grammar (CCG)
LU (Lu et al., 2008)

Semantic grammar, generative parsing model
33
Compared Systems

COCKTAIL (Tang & Mooney, 2001)


WASP (Wong & Mooney, 2006)


Semantic grammar, string kernels
Z&C (Zettleymoyer & Collins, 2007)


Semantic grammar, machine translation
KRISP (Kate & Mooney, 2006)


Deterministic, inductive logic programming
Hand-built
lexicon
for Geoquery
Manual CCG
Template rules
Syntax-based, combinatory categorial grammar (CCG)
LU (Lu et al., 2008)

Semantic grammar, generative parsing model
34
Compared Systems

COCKTAIL (Tang & Mooney, 2001)


WASP (Wong & Mooney, 2006)


handling
logical forms
Semantic grammar, string kernels
Z&C (Zettleymoyer & Collins, 2007)


Semantic grammar, machine translation
λ-WASP,
KRISP (Kate & Mooney, 2006)


Deterministic, inductive logic programming
Syntax-based, combinatory categorial grammar (CCG)
LU (Lu et al., 2008)

Semantic grammar, generative parsing model
35
Results on CLang
Precision
Recall
F-measure
COCKTAIL
-
-
-
SCISSOR
89.5
73.7
80.8
WASP
88.9
61.9
73.0
KRISP
85.2
61.9
71.7
-
-
-
82.4
57.7
67.8
Z&C
LU
Memory
overflow
Not reported
(LU: F-measure after reranking is 74.4%)
36
Results on CLang
Precision
Recall
F-measure
SCISSOR
89.5
73.7
80.8
WASP
88.9
61.9
73.0
KRISP
85.2
61.9
71.7
LU
82.4
57.7
67.8
(LU: F-measure after reranking is 74.4%)
37
Results on Geoquery
Precision
Recall
F-measure
SCISSOR
92.1
72.3
81.0
WASP
87.2
74.8
80.5
KRISP
93.3
71.7
81.1
LU
86.2
81.8
84.0
COCKTAIL
89.9
79.4
84.3
λ-WASP
92.0
86.6
89.2
Z&C
95.5
83.2
88.9
FunQL
Prolog
(LU: F-measure after reranking is 85.2%)
38
Results on Geoquery (FunQL)
Precision
Recall
F-measure
SCISSOR
92.1
72.3
81.0
WASP
87.2
74.8
80.5
KRISP
93.3
71.7
81.1
LU
86.2
81.8
84.0
competitive
(LU: F-measure after reranking is 85.2%)
39
Why Knowledge of Syntax
does not Help


Geoquery: 7.48 word per sentence
Short sentence


Sentence structure can be feasibly learned from
NLs paired with MRs
Gain from knowledge of syntax vs.
flexibility loss
40
Limitation of Using Prior
Knowledge of Syntax
Traditional syntactic analysis
N1
N2
What
is the smallest
state
answer(smallest(state(all)))
41
Limitation of Using Prior
Knowledge of Syntax
Traditional syntactic analysis
Semantic grammar
N1
N1
N2
What
is the smallest
state
answer(smallest(state(all)))
What
N2
state is the smallest
answer(smallest(state(all)))
Isomorphic syntactic structure with MR
Better generalization
42
Why Prior Knowledge of
Syntax does not Help


Geoquery: 7.48 word per sentence
Short sentence



Sentence structure can be feasibly learned from
NLs paired with MRs
Gain from knowledge of syntax vs.
flexibility loss
LU vs. WASP and KRISP

Decomposed model for semantic grammar
43
Detailed Clang Results on
Sentence Length
0-10
(7%)
11-20
(33%)
21-30
(46%)
31-40
(13%)
44
SCISSOR Summary



Integrated syntactic-semantic parsing
approach
Learns accurate semantic interpretations
by utilizing the SAPT annotations
knowledge of syntax improves performance
on long sentences
45
Roadmap




SCISSOR
SYNSEM
Future Work
Conclusions
46
SYNSEM Motivation



SCISSOR requires extra SAPT annotation for
training
Must learn both syntax and semantics from
same limited training corpus
High performance syntactic parsers are
available that are trained on existing large
corpora (Collins, 1997; Charniak & Johnson,
2005)
47
SCISSOR Requires SAPT Annotation
S-P_BOWNER
NP-P_PLAYER
VP-P_BOWNER
PRP$-P_OUR NN-P_PLAYER CD- P_UNUM VB-P_BOWNER
our
player
2
has
NP-NULL
DT-NULL NN-NULL
the
ball
Time consuming.
Automate it!
48
Part I: Syntactic Parse
S
NP
VP
PRP$
NN
CD
VB
our
player
2
has
NP
DT
NN
the
ball
Use a statistical syntactic parser
49
Part II: Word Meanings
P_OUR
P_PLAYER
our
player
our
P_UNUM
2
player
P_BOWNER NULL
has
2
has
the
the
NULL
ball
ball
P_BOWNER P_PLAYER P_OUR P_UNUM
Use a word alignment model (Wong and
Mooney (2006) )
50
Learning a Semantic Lexicon



IBM Model 5 word alignment (GIZA++)
top 5 word/predicate alignments for each training
example
Assume each word alignment and syntactic parse
defines a possible SAPT for composing the correct
MR
51
S
VP
NP
P_OUR
NP
λa1λa2P_PLAYER
our
player
λa1P_BOWNER
P_UNUM
2
has
NP
NULL
NULL
the
ball
Introducing λvariables in semantic labels for
missing arguments
(a1: the first argument)
52
Part III: Internal Semantic Labels
S
P_BOWNER
P_PLAYER
VP
NP
P_OUR
P_OUR
NP
λa1λa2P_PLAYER
our
player
λa1P_BOWNER
P_UNUM
2
has
P_UNUM
NP
NULL
NULL
the
ball
How to choose the dominant predicates?
53
Learning Semantic Composition Rules
?
λa1λa2P_PLAYER
player
λa1λa2PLAYER
P_BOWNER
P_UNUM
2
+ P_UNUM 
P_PLAYER
P_OUR
P_UNUM
λa1 P_PLAYER , a2=c2
(c2: child 2)
54
Learning Semantic Composition Rules
S
P_BOWNER
P_PLAYER
?
VP
P_OUR
P_OUR
λa1P_PLAYER
λa1λa2P_PLAYER
our
player
λa1P_BOWNER
P_UNUM
2
has
P_UNUM
NP
NULL
NULL
the
ball
λa1λa2PLAYER + P_UNUM  {λa1P_PLAYER, a2=c2}
55
Learning Semantic Composition Rules
S
P_BOWNER
P_PLAYER
P_PLAYER
VP
P_OUR
P_OUR
λa1P_PLAYER
λa1λa2P_PLAYER
our
player
λa1P_BOWNER
P_UNUM
2
has
?
NULL
NULL
the
ball
P_OUR +λa1P_PLAYER  {P_PLAYER, a1=c1}
P_UNUM
56
Learning Semantic Composition Rules
?
P_BOWNER
P_PLAYER
λa1P_BOWNER
P_PLAYER
P_OUR
P_OUR
λa1P_PLAYER
λa1λa2P_PLAYER
our
player
P_UNUM
λa1P_BOWNER NULL
P_UNUM
2
has
NULL
NULL
the
ball
57
Learning Semantic Composition Rules
P_BOWNER
P_BOWNER
P_PLAYER
λa1P_BOWNER
P_PLAYER
P_OUR
P_OUR
λa1P_PLAYER
λa1λa2P_PLAYER
our
player
P_UNUM
λa1P_BOWNER NULL
P_UNUM
2
has
NULL
NULL
the
ball
P_PLAYER + λa1P_BOWNER  {P_BOWNER, a1=c1}
58
Ensuring Meaning Composition
N1
N2
What
is the smallest
state
answer(smallest(state(all)))
Non-isomorphism
59
Ensuring Meaning Composition

Non-isomorphism between NL parse and MR
parse





Various linguistic phenomena
Machine translation between NL and MRL
Use automated syntactic parses
Introduce macro-predicates that combine
multiple predicates.
Ensure that MR can be composed using a
syntactic parse and word alignment
60
SYNSEM Overview
training/test sentence, S
Syntactic parser
syntactic parse tree,T
Before training & testing
Unambiguous CFG of MRL
Semantic knowledge acquisition
Training set, {(S,T,MR)}
Semantic lexicon & composition rules
Parameter estimation
Probabilistic parsing model
Training
Input sentence parse T
Testing
Output MR
Semantic parsing
61
SYNSEM Overview
training/test sentence, S
Syntactic parser
syntactic parse tree,T
Before training & testing
Unambiguous CFG of MRL
Semantic knowledge acquisition
Training set, {(S,T,MR)}
Semantic lexicon & composition rules
Parameter estimation
Probabilistic parsing model
Training
Input sentence, S
Testing
Output MR
Semantic parsing
62
Parameter Estimation
•
•
•
•
Apply the learned semantic knowledge to all training
examples to generate possible SAPTs
Use a standard maximum-entropy model similar to that
of Zettlemoyer & Collins (2005), and Wong & Mooney
(2006)
Training finds a parameter that (approximately)
maximizes the sum of the conditional log-likelihood of
the training set including syntactic parses
Incomplete data since SAPTs are hidden variables
63
Features

Lexical features:



Unigram features: # that a word is assigned a
predicate
Bigram features: # that a word is assigned a
predicate given its previous/subsequent word.
Rule features: # a composition rule applied in
a derivation
64
Handling Logical Forms
What are the rivers in Texas?
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
λv1P_ANSWER(x1)
λv1P_RIVER(x1) λv1λv2P_LOC(x1,x2) λv1P_EQUAL(x2)
Handle shared logical variables
Use Lambda Calculus (v: variable)
65
Prolog Example
What are the rivers in Texas?
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
λv1P_ANSWER(x1)
(λv1P_RIVER(x1) λv1 λv2P_LOC(x1,x2) λv1P_EQUAL(x2))
Handle shared logical variables
Use Lambda Calculus (v: variable)
66
Prolog Example
What are the rivers in Texas?
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
λv1P_ANSWER(x1)
(λv1P_RIVER(x1) λv1λv2P_LOC(x1,x2) λv1P_EQUAL(x2))
Handle shared logical variables
Use Lambda Calculus (v: variable)
67
Prolog Example
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
SBARQ
Start from a
syntactic parse
SQ
NP
PP
WHNP
VBP
NP
IN
NP
What
are
the rivers
in
Texas
68
Prolog Example
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
SBARQ
Add predicates to
words
SQ
NP
PP
λv1λa1P_ANSWER
What
NULL λv1P_RIVER
are
the rivers
λv1λv2P_LOC λv1P_EQUAL
in
Texas
69
Prolog Example
answer(x1, (river(x1), loc(x1,x2), equal(x2,stateid(texas))))
SBARQ
Learn a rule with
variable unification
SQ
NP
λv1P_LOC
λv1λa1P_ANSWER
What
NULL λv1P_RIVER
are
the rivers
λv1λv2P_LOC λv1P_EQUAL
in
λv1λv2P_LOC(x1,x2) + λv1P_EQUAL(x2)  λv1P_LOC
Texas
70
Experimental Results


CLang
Geoquery (Prolog)
71
Syntactic Parsers (Bikel,2004)

WSJ only



WSJ + in-domain sentences



CLang(SYN0): F-measure=82.15%
Geoquery(SYN0) : F-measure=76.44%
CLang(SYN20): 20 sentences, F-measure=88.21%
Geoquery(SYN40): 40 sentences, F-measure=91.46%
Gold-standard syntactic parses (GOLDSYN)
72
Questions

Q1. Can SYNSEM produce accurate semantic
interpretations?

Q2. Can more accurate Treebank syntactic parsers
produce more accurate semantic parsers?

Q3. Does it also improve on long sentences?

Q4. Does it improve on limited training data due
to the prior knowledge from large treebanks?

Q5. Can it handle syntactic errors?
73
Results on CLang
Precision
Recall
F-measure
GOLDSYN
84.7
74.0
79.0
SYN20
85.4
70.0
76.9
SYN0
87.0
67.0
75.7
SCISSOR
89.5
73.7
80.8
WASP
88.9
61.9
73.0
KRISP
85.2
61.9
71.7
LU
82.4
57.7
67.8
SYNSEM
SAPTs
(LU: F-measure after reranking is 74.4%)
GOLDSYN > SYN20 > SYN0
74
Questions

Q1. Can SynSem produce accurate semantic
interpretations? [yes]

Q2. Can more accurate Treebank syntactic parsers
produce more accurate semantic parsers? [yes]

Q3. Does it also improve on long sentences?
75
Detailed Clang Results on Sentence Length
Prior
+
Knowledge
Flexibility
0-10
(7%)
+
11-20
(33%)
Syntactic
error
21-30
(46%)
=
?
31-40
(13%)
76
Questions

Q1. Can SynSem produce accurate semantic
interpretations? [yes]

Q2. Can more accurate Treebank syntactic parsers
produce more accurate semantic parsers? [yes]

Q3. Does it also improve on long sentences? [yes]

Q4. Does it improve on limited training data due
to the prior knowledge from large treebanks?
77
Results on Clang
(training size = 40)
Precision
Recall
F-measure
GOLDSYN
61.1
35.7
45.1
SYN20
57.8
31.0
40.4
SYN0
53.5
22.7
31.9
SCISSOR
85.0
23.0
36.2
WASP
88.0
14.4
24.7
KRISP
68.35
20.0
31.0
SYNSEM
SAPTs
The quality of syntactic parser is critically important!
78
Questions

Q1. Can SynSem produce accurate semantic
interpretations? [yes]

Q2. Can more accurate Treebank syntactic parsers
produce more accurate semantic parsers? [yes]

Q3. Does it also improve on long sentences? [yes]

Q4. Does it improve on limited training data due
to the prior knowledge from large treebanks? [yes]

Q5. Can it handle syntactic errors?
79
Handling Syntactic Errors


Training ensures meaning composition from
syntactic parses with errors
For test NLs that generate correct MRs, measure
the F-measures of their syntactic parses


SYN0: 85.5%
SYN20: 91.2%
If DR2C7 is true
then players 2 , 3 , 7 and 8 should pass to player 4
80
Questions

Q1. Can SynSem produce accurate semantic
interpretations? [yes]

Q2. Can more accurate Treebank syntactic parsers
produce more accurate semantic parsers? [yes]

Q3. Does it also improve on long sentences? [yes]

Q4. Does it improve on limited training data due
to the prior knowledge of large treebanks? [yes]

Q5. Is it robust to syntactic errors? [yes]
81
Results on Geoquery (Prolog)
Precision
Recall
F-measure
GOLDSYN
91.9
88.2
90.0
SYN40
90.2
86.9
88.5
SYN0
81.8
79.0
80.4
COCKTAIL
89.9
79.4
84.3
λ-WASP
92.0
86.6
89.2
Z&C
95.5
83.2
88.9
SYNSEM
SYN0 does not perform well
All other recent systems perform competitively
82
SYNSEM Summary




Exploits an existing syntactic parser to drive
the meaning composition process
Prior knowledge of syntax improves
performance on long sentences
Prior knowledge of syntax improves
performance on limited training data
Handle syntactic errors
83
Discriminative Reranking for
semantic Parsing



Adapt global features used for reranking
syntactic parsing for semantic parsing
Improvement on CLang
No improvement on Geoquery where
sentences are short, and are less likely for
global features to show improvement on
84
Roadmap




SCISSOR
SYNSEM
Future Work
Conclusions
85
Future Work

Improve SCISSOR




Discriminative SCISSOR (Finkel, et al., 2008)
Handling logical forms
SCISSOR without extra annotation (Klein and
Manning, 2002, 2004)
Improve SYNSEM

Utilizing syntactic parsers with improved accuracy
and in other syntactic formalism
86
Future Work

Utilizing wide-coverage semantic
representations (Curran et al., 2007)


Better generalizations for syntactic variations
Utilizing semantic role labeling (Gildea and
Palmer, 2002)
 Provides a layer of correlated semantic
information
87
Roadmap




SCISSOR
SYNSEM
Future Work
Conclusions
88
Conclusions





SCISSOR: a novel integrated syntactic-semantic
parser.
SYNSEM: exploits an existing syntactic parser to
produce disambiguated parse trees that drive the
compositional meaning composition.
Both produce accurate semantic interpretations.
Using the knowledge of syntax improves
performance on long sentences.
SYNSEM also improves performance on limited
training data.
89
Thank you!

Questions?
90