Higher-order Constituent Parsing and Parser Combination
Chen, Xiao and Kit, Chunyu

Article Structure

Abstract

This paper presents a higher-order model for constituent parsing aimed at utilizing more local structural context to decide the score of a grammar rule instance in a parse tree.

Introduction

Factorization is crucial to discriminative parsing.

Higher-order Constituent Parsing

Discriminative parsing is aimed to learn a function f : S —> T from a set of sentences 8 to a set of valid parses 7 according to a given CFG, which maps an input sentence 8 E S to a set of candidate parses T The function takes the following discriminative form:

Constituent Recombination

Following Fossum and Knight (2009), our constituent weighting scheme for parser combination uses multiple outputs of independent parsers.

Experiment

Our parsing models are evaluated on both English and Chinese treebanks, i.e., the WSJ section of Penn Treebank 3.0 (LDC99T42) and the Chinese Treebank 5.1 (LDC2005T01U01).

Conclusion

This paper has presented a higher-order model for constituent parsing that factorizes a parse tree into larger parts than before, in hopes of increasing its power of discriminating the true parse from the others without losing tractability.

Topics

parsing model

Appears in 9 sentences as: parsing model (5) parsing models (4)
In Higher-order Constituent Parsing and Parser Combination
  1. Previous discriminative parsing models usually factor a parse tree into a set of parts.
    Page 1, “Introduction”
  2. Then, the previous discriminative constituent parsing models (Johnson, 2001; Henderson, 2004; Taskar et al., 2004; Petrov and Klein, 2008a;
    Page 1, “Introduction”
  3. The first feature ¢0(Q(r), s) is calculated with a PCFG-based generative parsing model (Petrov and Klein, 2007), as defined in (4) below, where 7“ is the grammar rule instance A —> B C that covers the span from the b-th
    Page 2, “Higher-order Constituent Parsing”
  4. With only lexical features in a part, this parsing model backs off to a first-order one similar to those in the previous works.
    Page 2, “Higher-order Constituent Parsing”
  5. Adding structural features, each involving a least a neighboring rule instance, makes it a higher-order parsing model .
    Page 2, “Higher-order Constituent Parsing”
  6. The factorization of the parsing model allows us to develop an exact decoding algorithm for it.
    Page 2, “Higher-order Constituent Parsing”
  7. Our parsing models are evaluated on both English and Chinese treebanks, i.e., the WSJ section of Penn Treebank 3.0 (LDC99T42) and the Chinese Treebank 5.1 (LDC2005T01U01).
    Page 3, “Experiment”
  8. The parameters 6 of each parsing model are estimated from a training set using an averaged perceptron algorithm, following Collins (2002) and Huang (2008).
    Page 4, “Experiment”
  9. The performance of our first— and higher-order parsing models on all sentences of the two test sets is presented in Table 3, where A indicates a tuned balance factor.
    Page 4, “Experiment”

See all papers in Proc. ACL 2012 that mention parsing model.

See all papers in Proc. ACL that mention parsing model.

Back to top.

constituent parsing

Appears in 8 sentences as: constituent parsing (8)
In Higher-order Constituent Parsing and Parser Combination
  1. This paper presents a higher-order model for constituent parsing aimed at utilizing more local structural context to decide the score of a grammar rule instance in a parse tree.
    Page 1, “Abstract”
  2. Similarly, we can define the order of constituent parsing in terms of the number of grammar rules in a part.
    Page 1, “Introduction”
  3. Then, the previous discriminative constituent parsing models (Johnson, 2001; Henderson, 2004; Taskar et al., 2004; Petrov and Klein, 2008a;
    Page 1, “Introduction”
  4. The discriminative re-scoring models (Collins, 2000; Collins and Duffy, 2002; Charniak and Johnson, 2005; Huang, 2008) can be viewed as previous attempts to higher-order constituent parsing , using some parts containing more than one grammar rule as nonlocal features.
    Page 1, “Introduction”
  5. In this paper, we present a higher-order constituent parsing model1 based on these previous works.
    Page 1, “Introduction”
  6. This paper has presented a higher-order model for constituent parsing that factorizes a parse tree into larger parts than before, in hopes of increasing its power of discriminating the true parse from the others without losing tractability.
    Page 4, “Conclusion”
  7. More importantly, it extends the existing works into a more general framework of constituent parsing to utilize more lexical and structural context and incorporate more strength of various parsing techniques.
    Page 4, “Conclusion”
  8. However, higher-order constituent parsing inevitably leads to a high computational complexity.
    Page 4, “Conclusion”

See all papers in Proc. ACL 2012 that mention constituent parsing.

See all papers in Proc. ACL that mention constituent parsing.

Back to top.

parse tree

Appears in 6 sentences as: parse tree (6)
In Higher-order Constituent Parsing and Parser Combination
  1. This paper presents a higher-order model for constituent parsing aimed at utilizing more local structural context to decide the score of a grammar rule instance in a parse tree .
    Page 1, “Abstract”
  2. Previous discriminative parsing models usually factor a parse tree into a set of parts.
    Page 1, “Introduction”
  3. It allows multiple adjacent grammar rules in each part of a parse tree , so as to utilize more local structural context to decide the plausibility of a grammar rule instance.
    Page 1, “Introduction”
  4. Figure l: A part of a parse tree centered at NP —> NP VP
    Page 2, “Higher-order Constituent Parsing”
  5. A part in a parse tree is illustrated in Figure 1.
    Page 2, “Higher-order Constituent Parsing”
  6. This paper has presented a higher-order model for constituent parsing that factorizes a parse tree into larger parts than before, in hopes of increasing its power of discriminating the true parse from the others without losing tractability.
    Page 4, “Conclusion”

See all papers in Proc. ACL 2012 that mention parse tree.

See all papers in Proc. ACL that mention parse tree.

Back to top.

Treebank

Appears in 4 sentences as: Treebank (3) treebank (1) treebanks (2)
In Higher-order Constituent Parsing and Parser Combination
  1. EXperiments on English and Chinese treebanks confirm its advantage over its first-order version.
    Page 1, “Abstract”
  2. Evaluated on the PTB WSJ and Chinese Treebank , it achieves its best Fl scores of 91.86% and 85.58%, respectively.
    Page 1, “Introduction”
  3. Our parsing models are evaluated on both English and Chinese treebanks, i.e., the WSJ section of Penn Treebank 3.0 (LDC99T42) and the Chinese Treebank 5.1 (LDC2005T01U01).
    Page 3, “Experiment”
  4. For parser combination, we follow the setting of Fossum and Knight (2009), using Section 24 instead of Section 22 of WSJ treebank as development set.
    Page 3, “Experiment”

See all papers in Proc. ACL 2012 that mention Treebank.

See all papers in Proc. ACL that mention Treebank.

Back to top.

development set

Appears in 3 sentences as: development set (3)
In Higher-order Constituent Parsing and Parser Combination
  1. The parameters )V; and p are tuned by the Powell’s method (Powell, 1964) on a development set , using the F1 score of PARSEVAL (Black et al., 1991) as objective.
    Page 3, “Constituent Recombination”
  2. For parser combination, we follow the setting of Fossum and Knight (2009), using Section 24 instead of Section 22 of WSJ treebank as development set .
    Page 3, “Experiment”
  3. It is tuned on a development set using the gold sec-
    Page 3, “Experiment”

See all papers in Proc. ACL 2012 that mention development set.

See all papers in Proc. ACL that mention development set.

Back to top.

F1 scores

Appears in 3 sentences as: F1 score (1) F1 scores (2)
In Higher-order Constituent Parsing and Parser Combination
  1. It achieves its best F1 scores of 91.86% and 85.58% on the two languages, respectively, and further pushes them to 92.80% and 85.60% via combination with other high—performance parsers.
    Page 1, “Abstract”
  2. Combined with other high-performance parsers under the framework of constituent recombination (Sagae and Lavie, 2006; Fossum and Knight, 2009), this model further enhances the F1 scores to 92.80% and 85.60%, the highest ones achieved so far on these two data sets.
    Page 1, “Introduction”
  3. The parameters )V; and p are tuned by the Powell’s method (Powell, 1964) on a development set, using the F1 score of PARSEVAL (Black et al., 1991) as objective.
    Page 3, “Constituent Recombination”

See all papers in Proc. ACL 2012 that mention F1 scores.

See all papers in Proc. ACL that mention F1 scores.

Back to top.

structural features

Appears in 3 sentences as: structural features (3)
In Higher-order Constituent Parsing and Parser Combination
  1. Some back-off structural features are used for smoothing, which cannot be presented due to limited space.
    Page 2, “Higher-order Constituent Parsing”
  2. Adding structural features , each involving a least a neighboring rule instance, makes it a higher-order parsing model.
    Page 2, “Higher-order Constituent Parsing”
  3. Because all structures above the current rule instance is not determined yet, the computation of its nonlocal structural features , e. g., parent and sibling features, has to be delayed until it joins an upper level structure.
    Page 2, “Higher-order Constituent Parsing”

See all papers in Proc. ACL 2012 that mention structural features.

See all papers in Proc. ACL that mention structural features.

Back to top.

PDF available from www.aclweb.org.
Send us an email