Directional Distributional Similarity for Lexical Expansion
Article Structure
Abstract
Distributional word similarity is most commonly perceived as a symmetric relation.
Introduction
Much work on automatic identification of semantically similar terms exploits Distributional Similarity, assuming that such terms appear in similar contexts.
Background
The distributional word similarity scheme follows two steps.
A Statistical Inclusion Measure
Our research goal was to develop a directional similarity measure suitable for learning asymmetric relations, focusing empirically on lexical expansion.
Evaluation and Results
4.1 Evaluation Setting
Conclusions and Future work
This paper advocates the use of directional similarity measures for lexical expansion, and potentially for other tasks, based on distributional inclusion of feature vectors.
Topics
similarity measures
Appears in 10 sentences as: similarity measure (4) similarity measures (6)
In Directional Distributional Similarity for Lexical Expansion
- This paper investigates the nature of directional (asymmetric) similarity measures , which aim to quantify distributional feature inclusion.Page 1, “Abstract”
- Often, distributional similarity measures are used to identify expanding terms (e.g.Page 1, “Introduction”
- More generally, directional relations are abundant in NLP settings, making symmetric similarity measures less suitable for their identification.Page 1, “Introduction”
- Despite the need for directional similarity measures , their investigation counts, to the best of our knowledge, only few works (Weeds and Weir, 2003; Geffet and Dagan, 2005; Bhagat et al., 2007; Szpektor and Dagan, 2008; Michelbacher et al., 2007) and is utterly lacking.Page 1, “Introduction”
- This paper investigates the nature of directional similarity measures .Page 1, “Introduction”
- Then, word vectors are compared by some vector similarity measure .Page 1, “Background”
- Our research goal was to develop a directional similarity measure suitable for learning asymmetric relations, focusing empirically on lexical expansion.Page 2, “A Statistical Inclusion Measure”
- We tested our similarity measure by evaluating its utility for lexical expansion, compared with baselines of the LIN, WeedsPrec and balPrec measuresPage 3, “Evaluation and Results”
- Next, for each similarity measure , the terms found similar to any of the event’s seeds (‘u —> seed’) were taken as expansion terms.Page 3, “Evaluation and Results”
- This paper advocates the use of directional similarity measures for lexical expansion, and potentially for other tasks, based on distributional inclusion of feature vectors.Page 4, “Conclusions and Future work”
See all papers in Proc. ACL 2009 that mention similarity measures.
See all papers in Proc. ACL that mention similarity measures.
Back to top.
feature vector
Appears in 9 sentences as: feature vector (5) Feature vectors (1) feature vectors (3)
In Directional Distributional Similarity for Lexical Expansion
- First, a feature vector is constructed for each word by collecting context words as features.Page 1, “Background”
- where FVgc is the feature vector of a word cc and way (f) is the weight of the feature f in that word’s vector, set to their pointwise mutual information.Page 2, “Background”
- Extending this rationale to the textual entailment setting, Geffet and Dagan (2005) expected that if the meaning of a word it entails that of 2) then all its prominent context features (under a certain notion of “prominence”) would be included in the feature vector of v as well.Page 2, “Background”
- Effectively, this measure penalizes infrequent templates having short feature vectors , as those usually yield low symmetric similarity with the longer vectors of more common templates.Page 2, “Background”
- Amongst these features, those found in 21’s feature vector are termed included features.Page 2, “A Statistical Inclusion Measure”
- In preliminary data analysis of pairs of feature vectors , which correspond to a known set of valid and invalid expansions, we identified the following desired properties for a distributional inclusion measure.Page 2, “A Statistical Inclusion Measure”
- In our case the feature vector of the expanded word is analogous to the set of all relevant documents while tested features correspond to retrieved documents.Page 2, “A Statistical Inclusion Measure”
- Feature vectors were created by parsing the Reuters RCVl corpus and taking the words related to each term through a dependency relation as its features (coupled with the relation name and direction, as in (Lin, 1998)).Page 3, “Evaluation and Results”
- This paper advocates the use of directional similarity measures for lexical expansion, and potentially for other tasks, based on distributional inclusion of feature vectors .Page 4, “Conclusions and Future work”
See all papers in Proc. ACL 2009 that mention feature vector.
See all papers in Proc. ACL that mention feature vector.
Back to top.
distributional similarity
Appears in 6 sentences as: Distributional Similarity (1) distributional similarity (5)
In Directional Distributional Similarity for Lexical Expansion
- Much work on automatic identification of semantically similar terms exploits Distributional Similarity , assuming that such terms appear in similar contexts.Page 1, “Introduction”
- This paper is motivated by one of the prominent applications of distributional similarity , namely identifying lexical expansions.Page 1, “Introduction”
- Often, distributional similarity measures are used to identify expanding terms (e.g.Page 1, “Introduction”
- While distributional similarity is most prominently modeled by symmetric measures, lexical expansion is in general a directional relation.Page 1, “Introduction”
- To date, most distributional similarity research concentrated on symmetric measures, such as the widely cited and competitive (as shown in (Weeds and Weir, 2003)) LIN measure (Lin, 1998):Page 2, “Background”
- In this setting, category names were taken as seeds and expanded by distributional similarity , further measuring cosine similarity with categorized documents similarly to IR query expansion.Page 4, “Evaluation and Results”
See all papers in Proc. ACL 2009 that mention distributional similarity.
See all papers in Proc. ACL that mention distributional similarity.
Back to top.
