Identifying Generic Noun Phrases
Reiter, Nils and Frank, Anette

Article Structure

Abstract

This paper presents a supervised approach for identifying generic noun phrases in context.

Introduction

Generic expressions come in two basic forms: generic noun phrases and generic sentences.

Topics

feature sets

Appears in 15 sentences as: Feature set (1) feature Set (1) feature set (6) feature sets (8)
In Identifying Generic Noun Phrases
  1. In section 4 we motivate the choice of feature sets for the automatic identification of generic NPs in context.
    Page 2, “Introduction”
  2. 4.2 Feature set and feature classes
    Page 4, “Introduction”
  3. The feature set includes NP-local and global features.
    Page 4, “Introduction”
  4. Feature classes We performed evaluation runs for different combinations of feature sets : NP- vs. S-level features (with further distinction between syntactic and semantic NP-/S-level features), as well as overall syntactic vs. semantic features.
    Page 7, “Introduction”
  5. Table 4 shows the resulting feature sets .
    Page 8, “Introduction”
  6. In ablation testing, a single feature in turn is temporarily omitted from the feature set .
    Page 8, “Introduction”
  7. This process is repeated until we are left with an empty feature set .
    Page 8, “Introduction”
  8. From the ranked list of features f1 to fn we evaluate increasingly extended feature sets f1.. fi for i = 2..n. We select the feature set that yields the best balanced performance, at 45.7% precision and 53.6% f-measure.
    Page 8, “Introduction”
  9. The respective feature sets are given as Set 1 to Set 3 in Table 4.
    Page 8, “Introduction”
  10. Most of the features and feature sets yield precision values above the results of Suh.
    Page 8, “Introduction”
  11. The results for each feature set are given in Table 6.
    Page 8, “Introduction”

See all papers in Proc. ACL 2010 that mention feature sets.

See all papers in Proc. ACL that mention feature sets.

Back to top.

noun phrases

Appears in 9 sentences as: noun phrases (9)
In Identifying Generic Noun Phrases
  1. This paper presents a supervised approach for identifying generic noun phrases in context.
    Page 1, “Abstract”
  2. Generic expressions come in two basic forms: generic noun phrases and generic sentences.
    Page 1, “Introduction”
  3. According to the second view, generic noun phrases denote kinds.
    Page 2, “Introduction”
  4. We are not aware of any detailed assessment of the proportion of generic noun phrases in educational text genres or ency-clopaedic resources like Wikipedia.
    Page 3, “Introduction”
  5. Suh (2006) applied a rule-based approach to automatically identify generic noun phrases .
    Page 3, “Introduction”
  6. Suh used patterns based on part of speech tags that identify bare plural noun phrases , reporting a precision of 28.9% for generic entities, measured against an annotated corpus, the ACE 2005 (Ferro et al., 2005).
    Page 3, “Introduction”
  7. Two annotators annotated 48 noun phrases from the British National Corpus for their genericity (and specificity) properties, obtaining a kappa value of 0.744.
    Page 3, “Introduction”
  8. Thus, while at first sight the guidelines do not fully correspond to the characterisation of generics we find in the formal semantics literature, we argue that both characterisations have similar extensions, i.e., include largely overlapping sets of noun phrases .
    Page 6, “Introduction”
  9. In fact, all of the examples for generic noun phrases presented in this paper would also be classified as generic according to the ACE-2 guidelines.
    Page 6, “Introduction”

See all papers in Proc. ACL 2010 that mention noun phrases.

See all papers in Proc. ACL that mention noun phrases.

Back to top.

f-measure

Appears in 7 sentences as: f-measure (7)
In Identifying Generic Noun Phrases
  1. Neither recall nor f-measure are reported.
    Page 3, “Introduction”
  2. The feature whose omission causes the biggest drop in f-measure is set aside as a strong feature.
    Page 8, “Introduction”
  3. From the ranked list of features f1 to fn we evaluate increasingly extended feature sets f1.. fi for i = 2..n. We select the feature set that yields the best balanced performance, at 45.7% precision and 53.6% f-measure .
    Page 8, “Introduction”
  4. Even though our results achieve a lower precision than the Person baseline, in terms of f-measure , we achieve a result of over 50%, which is almost three times the baseline.
    Page 8, “Introduction”
  5. With respect to overall f-measure , the best single features are strong on the unbalanced data.
    Page 9, “Introduction”
  6. The best performing feature in terms of f-measure on both balanced and unbalanced data is Set 5 with Set 4 as a close followup.
    Page 9, “Introduction”
  7. In terms of f-measure on the generic class, all feature sets performed above the baseline(s).
    Page 9, “Introduction”

See all papers in Proc. ACL 2010 that mention f-measure.

See all papers in Proc. ACL that mention f-measure.

Back to top.

lexical semantic

Appears in 5 sentences as: Lexical semantic (1) lexical semantic (4)
In Identifying Generic Noun Phrases
  1. Lexical semantic factors, such as the semantic type of the clause predicate (5.c,e), or “well-established” kinds (5.g) may favour a generic reading, but such lexical factors are difficult to capture in a rule-based setting.
    Page 4, “Introduction”
  2. In the following, we will structure this feature space along two dimensions, distinguishing NP- and sentence-level factors as well as syntactic and semantic (including lexical semantic ) factors.
    Page 4, “Introduction”
  3. Semantic features include semantic features abstracted from syntax, such as tense and aspect or type of modification, but also lexical semantic features such as word sense classes, sense granularity or verbal predicates.
    Page 4, “Introduction”
  4. This holds especially for the lexical semantic features.
    Page 8, “Introduction”
  5. In particular, we will address lexical semantic features, as they tend to be effected by sparsity.
    Page 9, “Introduction”

See all papers in Proc. ACL 2010 that mention lexical semantic.

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

Back to top.

machine learning

Appears in 5 sentences as: machine learning (5)
In Identifying Generic Noun Phrases
  1. We will argue that the automatic identification of generic expressions should be cast as a machine learning problem instead of a rule-based approach, as there is (i) no transparent marking of genericity in English (as in most other European languages) and (ii) the phenomenon is highly context dependent.
    Page 2, “Introduction”
  2. In this paper, we build on insights from formal semantics to establish a corpus-based machine learning approach for the automatic classification of generic expressions.
    Page 2, “Introduction”
  3. In our view, these observations call for a corpus-based machine learning approach that is able to capture a variety of factors indicating genericity in combination and in context.
    Page 4, “Introduction”
  4. We presented a data-driven machine learning approach for identifying generic NPs in context that in turn can be used to improve tasks such as knowledge acquisition and organisation.
    Page 9, “Introduction”
  5. Therefore, a machine learning approach seemed promising, both for the identification of relevant features as for capturing contex-
    Page 9, “Introduction”

See all papers in Proc. ACL 2010 that mention machine learning.

See all papers in Proc. ACL that mention machine learning.

Back to top.

rule-based

Appears in 3 sentences as: rule-based (3)
In Identifying Generic Noun Phrases
  1. We will argue that the automatic identification of generic expressions should be cast as a machine learning problem instead of a rule-based approach, as there is (i) no transparent marking of genericity in English (as in most other European languages) and (ii) the phenomenon is highly context dependent.
    Page 2, “Introduction”
  2. Suh (2006) applied a rule-based approach to automatically identify generic noun phrases.
    Page 3, “Introduction”
  3. Lexical semantic factors, such as the semantic type of the clause predicate (5.c,e), or “well-established” kinds (5.g) may favour a generic reading, but such lexical factors are difficult to capture in a rule-based setting.
    Page 4, “Introduction”

See all papers in Proc. ACL 2010 that mention rule-based.

See all papers in Proc. ACL that mention rule-based.

Back to top.

sentence-level

Appears in 3 sentences as: Sentence-level (1) sentence-level (2)
In Identifying Generic Noun Phrases
  1. In the following, we will structure this feature space along two dimensions, distinguishing NP- and sentence-level factors as well as syntactic and semantic (including lexical semantic) factors.
    Page 4, “Introduction”
  2. Sentence-level features are extracted from the clause (in which the NP appears), as well as sentential and non-sentential adjuncts of the clause.
    Page 4, “Introduction”
  3. Using syntactic features on the NP-or sentence-level only, however, leads to a drop in precision as well as recall.
    Page 8, “Introduction”

See all papers in Proc. ACL 2010 that mention sentence-level.

See all papers in Proc. ACL that mention sentence-level.

Back to top.

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