Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
Biran, Or and McKeown, Kathleen

Article Structure

Abstract

We present a reformulation of the word pair features typically used for the task of disambiguating implicit relations in the Penn Discourse Treebank.

Introduction

Discourse relations such as contrast and causality are part of what makes a text coherent.

Related Work

This line of research began with (Marcu and Echihabi, 2002), who used a small number of unambiguous explicit markers and patterns involving them, such as [Arg1, but Arg2] to collect sets of word pairs from a large corpus using the crossproduct of the words in Argl and Arg2.

Word Pairs

3.1 The Problem: Sparsity

Evaluation of Word Pairs

For our main evaluation, we evaluate the performance of word pair features when used with no additional features.

Other Features

For our secondary evaluation, we include additional features to complement the word pairs.

Evaluation of Additional Features

For our secondary evaluation, we present results for each feature category on its own in Table 3 and for our best system for each of the relation classes in Table 2.

Conclusion

We presented an aggregated approach to word pair features and showed that it outperforms the previous formulation for all relation types but contingency.

Topics

word pairs

Appears in 36 sentences as: word pair (13) word pairs (27)
In Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
  1. We present a reformulation of the word pair features typically used for the task of disambiguating implicit relations in the Penn Discourse Treebank.
    Page 1, “Abstract”
  2. Our word pair features achieve significantly higher performance than the previous formulation when evaluated without additional features.
    Page 1, “Abstract”
  3. Without an explicit marker to rely on, work on this task initially focused on using lexical cues in the form of word pairs mined from large corpora where they appear around an explicit marker (Marcu and Echihabi, 2002).
    Page 1, “Introduction”
  4. The intuition is that these pairs will tend to represent semantic relationships which are related to the discourse marker (for example, word pairs often appearing around but may tend to be antonyms).
    Page 1, “Introduction”
  5. While this approach showed some success and has been used extensively in later work, it has been pointed out by multiple authors that many of the most useful word pairs
    Page 1, “Introduction”
  6. In particular, we present a reformulation of the word pair features that have most often been used for this task in the past, replacing the sparse lexical features with dense aggregated score features.
    Page 1, “Introduction”
  7. We show that our formulation outperforms the original one while requiring less features, and that using a stop list of functional words does not significantly affect performance, suggesting that these features indeed represent semantically related content word pairs .
    Page 1, “Introduction”
  8. In addition, we present a system which combines these word pairs with additional features to achieve near state of the art performance without the use of syntactic parse features and of context outside the arguments of the relation.
    Page 1, “Introduction”
  9. This line of research began with (Marcu and Echihabi, 2002), who used a small number of unambiguous explicit markers and patterns involving them, such as [Arg1, but Arg2] to collect sets of word pairs from a large corpus using the crossproduct of the words in Argl and Arg2.
    Page 1, “Related Work”
  10. Second, it is constructed with the same unsupervised method they use to extract the word pairs -by assuming that the patterns correspond to a particular relation and collecting the arguments from an unannotated corpus.
    Page 2, “Related Work”
  11. They used word pairs as well as additional features to train four binary classifiers, each corresponding to one of the high-level PDTB relation classes.
    Page 2, “Related Work”

See all papers in Proc. ACL 2013 that mention word pairs.

See all papers in Proc. ACL that mention word pairs.

Back to top.

content word

Appears in 3 sentences as: content word (2) content words (1)
In Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
  1. We show that our formulation outperforms the original one while requiring less features, and that using a stop list of functional words does not significantly affect performance, suggesting that these features indeed represent semantically related content word pairs.
    Page 1, “Introduction”
  2. An analysis in (Pitler et al., 2009) also shows that the top word pairs (ranked by information gain) all contain common functional words, and are not at all the semantically-related content words that were imagined.
    Page 2, “Word Pairs”
  3. With this approach, using a stop list does not have a major effect on results for most relation classes, which suggests most of the word pairs affecting performance are content word pairs which may truly be semantically related to the discourse structure.
    Page 5, “Conclusion”

See all papers in Proc. ACL 2013 that mention content word.

See all papers in Proc. ACL that mention content word.

Back to top.

semantically related

Appears in 3 sentences as: semantic relationships (1) semantically related (2)
In Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
  1. The intuition is that these pairs will tend to represent semantic relationships which are related to the discourse marker (for example, word pairs often appearing around but may tend to be antonyms).
    Page 1, “Introduction”
  2. We show that our formulation outperforms the original one while requiring less features, and that using a stop list of functional words does not significantly affect performance, suggesting that these features indeed represent semantically related content word pairs.
    Page 1, “Introduction”
  3. With this approach, using a stop list does not have a major effect on results for most relation classes, which suggests most of the word pairs affecting performance are content word pairs which may truly be semantically related to the discourse structure.
    Page 5, “Conclusion”

See all papers in Proc. ACL 2013 that mention semantically related.

See all papers in Proc. ACL that mention semantically related.

Back to top.

state of the art

Appears in 3 sentences as: state of the art (3)
In Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
  1. In addition, we present results for a full system using additional features which achieves close to state of the art performance without resorting to gold syntactic parses or to context outside the relation.
    Page 1, “Abstract”
  2. In addition, we present a system which combines these word pairs with additional features to achieve near state of the art performance without the use of syntactic parse features and of context outside the arguments of the relation.
    Page 1, “Introduction”
  3. (2009), who used mostly similar features, for comparison and temporal and is competitive with the most recent state of the art systems for contingency and expansion without using any syntactic or context features.
    Page 5, “Conclusion”

See all papers in Proc. ACL 2013 that mention state of the art.

See all papers in Proc. ACL that mention state of the art.

Back to top.

syntactic parses

Appears in 3 sentences as: syntactic parse (1) syntactic parses (2)
In Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
  1. In addition, we present results for a full system using additional features which achieves close to state of the art performance without resorting to gold syntactic parses or to context outside the relation.
    Page 1, “Abstract”
  2. In addition, we present a system which combines these word pairs with additional features to achieve near state of the art performance without the use of syntactic parse features and of context outside the arguments of the relation.
    Page 1, “Introduction”
  3. 1Reliable syntactic parses are not always available in domains other than newswire, and context (preceding relations, especially explicit relations) is not always available in some applications such as generation and question answering.
    Page 1, “Related Work”

See all papers in Proc. ACL 2013 that mention syntactic parses.

See all papers in Proc. ACL that mention syntactic parses.

Back to top.

Treebank

Appears in 3 sentences as: Treebank (3)
In Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
  1. We present a reformulation of the word pair features typically used for the task of disambiguating implicit relations in the Penn Discourse Treebank .
    Page 1, “Abstract”
  2. More recently, implicit relation prediction has been evaluated on annotated implicit relations from the Penn Discourse Treebank (Prasad et al., 2008).
    Page 2, “Related Work”
  3. Previous work has relied on features based on the gold parse trees of the Penn Treebank (which overlaps with PDTB) and on contextual information from relations preceding the one being disambiguated.
    Page 4, “Other Features”

See all papers in Proc. ACL 2013 that mention Treebank.

See all papers in Proc. ACL that mention Treebank.

Back to top.

WordNet

Appears in 3 sentences as: WordNet (4)
In Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
  1. 1 WordNet 20.07 34.07 52.96 11.58 2 Verb Class 14.24 24.84 49.6 10.04 3 MPN 23.84 38.58 49.97 13.16 4 Modality 17.49 28.92 13.84 10.72 5 Polarity 16.46 26.36 65.15 11.58 6 Affect 18.62 31.59 59.8 13.37 7 8 9
    Page 4, “Evaluation of Word Pairs”
  2. WordNet Features: We define four features based on WordNet (Fellbaum, 1998) - Synonyms, Antonyms, Hypernyms and Hyponyms.
    Page 4, “Other Features”
  3. In addition, we introduced the new and useful WordNet , Aflect, Length and Negation feature categories.
    Page 5, “Conclusion”

See all papers in Proc. ACL 2013 that mention WordNet.

See all papers in Proc. ACL that mention WordNet.

Back to top.

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