A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
Kawahara, Daisuke and Peterson, Daniel W. and Palmer, Martha

Article Structure

Abstract

We present an unsupervised method for inducing verb classes from verb uses in gigaword corpora.

Introduction

A verb plays a primary role in conveying the meaning of a sentence.

Related Work

As stated in Section 1, most of the previous studies on verb clustering assume that verbs are monosemous.

Our Approach

3.1 Overview

Experiments and Evaluations

We first describe our experimental settings and define evaluation metrics to evaluate induced soft clusterings of verb classes.

Conclusion

We presented a stepwise unsupervised method for inducing verb classes from instances in gigaword corpora.

Topics

clusterings

Appears in 9 sentences as: clusterings (9)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. Moreover, to the best of our knowledge, none of the following approaches attempt to quantitatively evaluate soft clusterings of verb classes induced by polysemy-aware unsupervised approaches (Korhonen et al., 2003; Lapata and Brew, 2004; Li and Brew, 2007; Schulte im Walde et al., 2008).
    Page 1, “Introduction”
  2. We first describe our experimental settings and define evaluation metrics to evaluate induced soft clusterings of verb classes.
    Page 5, “Experiments and Evaluations”
  3. This kind of normalization for soft clusterings was performed for other evaluation metrics as in Springorum et al.
    Page 5, “Experiments and Evaluations”
  4. (2003) evaluated hard clusterings based on a gold standard with multiple classes per verb.
    Page 5, “Experiments and Evaluations”
  5. They reported only precision measures including modified purity, and avoided extending the evaluation metrics for soft clusterings .
    Page 5, “Experiments and Evaluations”
  6. (2003) actually hardened the clusterings and left
    Page 6, “Experiments and Evaluations”
  7. the evaluations of soft clusterings for their future work.
    Page 6, “Experiments and Evaluations”
  8. clusterings with the NN and IB methods are obtained by using the VALEX subcategorization lexicon.
    Page 8, “Experiments and Evaluations”
  9. To harden the clusterings of the IB method and the LDA-frames, the class with the highest probability is selected for each verb.
    Page 8, “Experiments and Evaluations”

See all papers in Proc. ACL 2014 that mention clusterings.

See all papers in Proc. ACL that mention clusterings.

Back to top.

gold-standard

Appears in 8 sentences as: gold-standard (8)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. The effectiveness of our approach is verified through quantitative evaluations based on polysemy-aware gold-standard data.
    Page 1, “Abstract”
  2. They evaluated their result with a gold-standard test set, Where a single class is assigned to a verb.
    Page 2, “Related Work”
  3. They considered multiple classes only in the gold-standard data used for their evaluations.
    Page 2, “Related Work”
  4. We also evaluate our induced verb classes on this gold-standard data, which was created on the basis of Levin’s classes (Levin, 1993).
    Page 2, “Related Work”
  5. Where there is no frequency information available for class distribution, such as the gold-standard data described in Section 4.3, we use a uniform distribution across the verb’s classes.
    Page 5, “Experiments and Evaluations”
  6. Table 1: An excerpt of the gold-standard verb classes for several verbs from Korhonen et al.
    Page 6, “Experiments and Evaluations”
  7. We evaluate the single-class output for each verb based on the predominant gold-standard classes, which are defined for each verb in the test set of Korhonen et al.
    Page 7, “Experiments and Evaluations”
  8. Then, we compare the induced verb classes of the SemLink instances with their gold-standard VerbNet classes.
    Page 8, “Experiments and Evaluations”

See all papers in Proc. ACL 2014 that mention gold-standard.

See all papers in Proc. ACL that mention gold-standard.

Back to top.

gold standard

Appears in 5 sentences as: gold standard (5)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. (2003) evaluated hard clusterings based on a gold standard with multiple classes per verb.
    Page 5, “Experiments and Evaluations”
  2. Interestingly, this result indicates that slot distributions are more effective than lexical information in slot-word pairs for inducing verb classes similar to the gold standard .
    Page 7, “Experiments and Evaluations”
  3. This result is consistent with expectations, given a gold standard based on Levin’s verb classes, which are organized according to the syntactic behavior of verbs.
    Page 7, “Experiments and Evaluations”
  4. (2003), using the gold standard with multiple classes, which we also use for our multi-class evaluations.
    Page 7, “Experiments and Evaluations”
  5. When we evaluate against the multiple classes in the gold standard , we do normalize the inverse purity.
    Page 7, “Experiments and Evaluations”

See all papers in Proc. ACL 2014 that mention gold standard.

See all papers in Proc. ACL that mention gold standard.

Back to top.

dependency parsing

Appears in 4 sentences as: dependency parses (1) dependency parsing (3)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. 1. apply dependency parsing to a raw corpus and extract predicate-argument structures for each verb from the automatic parses,
    Page 3, “Our Approach”
  2. We apply dependency parsing to a large raw corpus.
    Page 4, “Our Approach”
  3. Then, we extract predicate-argument structures from the dependency parses .
    Page 4, “Our Approach”
  4. For instance, Kawahara and Kurohashi (2006) improved accuracy of dependency parsing based on Japanese semantic frames automatically induced from a raw corpus.
    Page 9, “Conclusion”

See all papers in Proc. ACL 2014 that mention dependency parsing.

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

Back to top.

evaluation metrics

Appears in 4 sentences as: Evaluation Metrics (1) evaluation metrics (3)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. We first describe our experimental settings and define evaluation metrics to evaluate induced soft clusterings of verb classes.
    Page 5, “Experiments and Evaluations”
  2. 4.2 Evaluation Metrics
    Page 5, “Experiments and Evaluations”
  3. This kind of normalization for soft clusterings was performed for other evaluation metrics as in Springorum et al.
    Page 5, “Experiments and Evaluations”
  4. They reported only precision measures including modified purity, and avoided extending the evaluation metrics for soft clusterings.
    Page 5, “Experiments and Evaluations”

See all papers in Proc. ACL 2014 that mention evaluation metrics.

See all papers in Proc. ACL that mention evaluation metrics.

Back to top.

sense disambiguation

Appears in 4 sentences as: sense disambiguation (4)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. Such verb classes have been used in many NLP applications that need to consider semantics in particular, such as word sense disambiguation (Dang, 2004), semantic parsing (Swier and Stevenson, 2005; Shi and Mihalcea, 2005) and discourse parsing (Subba and Di Eugenio, 2009).
    Page 1, “Introduction”
  2. They conducted several evaluations including predominant class induction and token-level verb sense disambiguation , but did not evaluate multiple classes output by their models.
    Page 3, “Related Work”
  3. For each predicate-argument structure of a verb, we couple the verb and an argument to make a unit for sense disambiguation .
    Page 4, “Our Approach”
  4. As applications of the resulting semantic frames and verb classes, we plan to integrate them into syntactic parsing, semantic role labeling and verb sense disambiguation .
    Page 9, “Conclusion”

See all papers in Proc. ACL 2014 that mention sense disambiguation.

See all papers in Proc. ACL that mention sense disambiguation.

Back to top.

im

Appears in 3 sentences as: im (3)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. Most of these approaches assume that all target verbs are monosemous (Stevenson and Joanis, 2003; Schulte im Walde, 2006; Joanis et al., 2008; Li and Brew, 2008; Sun et al., 2008; Sun and Korhonen, 2009; Vlachos et al., 2009; Parisien and Stevenson, 2010; Parisien and Stevenson, 2011; Falk et al., 2012; Lippincott et al., 2012; Reichart and Korhonen, 2013; Sun et al., 2013).
    Page 1, “Introduction”
  2. Moreover, to the best of our knowledge, none of the following approaches attempt to quantitatively evaluate soft clusterings of verb classes induced by polysemy-aware unsupervised approaches (Korhonen et al., 2003; Lapata and Brew, 2004; Li and Brew, 2007; Schulte im Walde et al., 2008).
    Page 1, “Introduction”
  3. Schulte im Walde et al.
    Page 3, “Related Work”

See all papers in Proc. ACL 2014 that mention im.

See all papers in Proc. ACL that mention im.

Back to top.

semantic roles

Appears in 3 sentences as: semantic role (1) semantic roles (3)
In A Step-wise Usage-based Method for Inducing Polysemy-aware Verb Classes
  1. (2012) extended the model of Titov and Klementiev (2012), which is an unsupervised model for inducing semantic roles, to jointly induce semantic roles and frames across verbs using the Chinese Restaurant Process (Aldous, 1985).
    Page 2, “Related Work”
  2. available on the web site.6 This frame data was induced from the BNC and consists of 1,200 frames and 400 semantic roles .
    Page 7, “Experiments and Evaluations”
  3. As applications of the resulting semantic frames and verb classes, we plan to integrate them into syntactic parsing, semantic role labeling and verb sense disambiguation.
    Page 9, “Conclusion”

See all papers in Proc. ACL 2014 that mention semantic roles.

See all papers in Proc. ACL that mention semantic roles.

Back to top.