A Combination of Active Learning and Semi-supervised Learning Starting with Positive and Unlabeled Examples for Word Sense Disambiguation: An Empirical Study on Japanese Web Search Query
Article Structure
Abstract
This paper proposes to solve the bottleneck of finding training data for word sense disambiguation (WSD) in the domain of web queries, where a complete set of ambiguous word senses are unknown.
Introduction
In Web mining for sentiment or reputation analysis, it is important for reliable analysis to extract large amount of texts about certain products, shops, or persons with high accuracy.
Topics
word sense
Appears in 7 sentences as: word sense (5) word senses (3)
In A Combination of Active Learning and Semi-supervised Learning Starting with Positive and Unlabeled Examples for Word Sense Disambiguation: An Empirical Study on Japanese Web Search Query
- This paper proposes to solve the bottleneck of finding training data for word sense disambiguation (WSD) in the domain of web queries, where a complete set of ambiguous word senses are unknown.Page 1, “Abstract”
- In this paper, we present a combination of active learning and semi-supervised learning method to treat the case when positive examples, which have an expected word sense in web search result, are only given.Page 1, “Abstract”
- When retrieving texts from Web archive, we often suffer from word sense ambiguity and WSD system is indispensable.Page 1, “Introduction”
- Because target words are often proper nouns, their word senses are rarely listed in handcrafted lexicon.Page 1, “Introduction”
- In selecting pseudo negative dataset, we predict word sense of each unlabeled example using thePage 2, “Introduction”
- 23 s(d) 9 word sense prediction for (1 using I— new 24 c(d, s(d)) 9 the confidence of prediction of d 25 if c(d, s(d)) < cnn-n thenPage 2, “Introduction”
- M-clustering is a variant of b-clustering where the given number of clusters are each number of ambiguous word senses in table 2.Page 3, “Introduction”
See all papers in Proc. ACL 2009 that mention word sense.
See all papers in Proc. ACL that mention word sense.
Back to top.
confidence score
Appears in 6 sentences as: confidence score (6)
In A Combination of Active Learning and Semi-supervised Learning Starting with Positive and Unlabeled Examples for Word Sense Disambiguation: An Empirical Study on Japanese Web Search Query
- The novelty of our approach is to use “pseudo negative examples” with reliable confidence score estimated by a classifier trained with positive and unlabeled examples.Page 1, “Abstract”
- Using naive Bayes classifier, we can estimate the confidence score c(d, s) that the sense of a data instance “(1”, whose features are f1, f2, ..., fn,Page 2, “Introduction”
- 01 foreacth(T—P—N) 02 classify d by WSD system I— (P, T-P) 03 c(d, pos) <— the confidence score that d isPage 3, “Introduction”
- 04 predicted as positive defined in equation (2) 05 c(d, neg) <— the confidence score that d isPage 3, “Introduction”
- 08 (the confidence score that d isPage 3, “Introduction”
- Random sampling with EM, abbreviated as with-EM, is the variant approach where dmin in line 26 of figure 1 is randomly selected without using confidence score .Page 4, “Introduction”
See all papers in Proc. ACL 2009 that mention confidence score.
See all papers in Proc. ACL that mention confidence score.
Back to top.
semi-supervised
Appears in 5 sentences as: semi-supervised (5)
In A Combination of Active Learning and Semi-supervised Learning Starting with Positive and Unlabeled Examples for Word Sense Disambiguation: An Empirical Study on Japanese Web Search Query
- In this paper, we present a combination of active learning and semi-supervised learning method to treat the case when positive examples, which have an expected word sense in web search result, are only given.Page 1, “Abstract”
- McCallum and Nigam (1998) combined active learning and semi-supervised learning techniquePage 1, “Introduction”
- Figure l: A combination of active learning and semi-supervised learning starting with positive and unlabeled examplesPage 2, “Introduction”
- Next we use Nigam’s semi-supervised learning method using EM and a naive Bayes classifier (Nigam et.Page 2, “Introduction”
- This result suggests semi-supervised learning using unlabeled examples is effective.Page 4, “Introduction”
See all papers in Proc. ACL 2009 that mention semi-supervised.
See all papers in Proc. ACL that mention semi-supervised.
Back to top.
