Boosting Applied to Word Sense Disambiguation

Abstract

Abstract. In this paper Schapire and Singer's AdaBoost.MH boosting algorithm is applied to the Word Sense Disambiguation (WSD) problem. Initial experiments on a set of 15 selected polysemous words show t h a t the boosting approach surpasses Naive B a yes and Exemplar{based approaches, which represent state{of{the{art accuracy on supervised WSD. In order to make boosting practical for a real learning domain of thousands of words, several ways of accelerating the algorithm by reducing the feature space are studied. The best variant, which w e c a l l LazyBoosting, is tested on a medium/large sense{tagged corpus containing 192,800 examples of the 191 most frequent a n d a m biguous English words. Again, boosting compares favourably to the other benchmark algorithms.

Margin maximization with feed-forward neural networks: a comparative study with SVM and AdaBoost

“…Weighting the sum-of-squares error derivative also implies a kind of weighting the sum-of-squares error, although in a slightly di erent way. 2 The following result justiÿes that the previously suggested weighted error functions (8) and (9) are well founded. In addition, it allows to construct new error functions with the same underlying ideas.…”

“…All experiments were performed with FNN trained with standard BP weighting the sum-of-squares error derivative (9). From now on, we will refer to this method as BPW.…”

“…Although far from obtaining satisfactory results [17], the best WSD systems up to date are based on supervised machine learning algorithms. Among others, we ÿnd Decision Lists [42], Neural Networks [36], Bayesian learning [2], Examplar Based learning [22], AdaBoost [8], and SVM [18] in the recent literature.…”

An Introduction to Boosting and Leveraging

“…Applications of boosting methods to natural language processing has been reported in [1,60,84,194], and approaches to Melanoma Diagnosis are presented in [132]. Some further applications to Pose Invariant Face Recognition [94], Lung Cancer Cell Identification [200] and Volatility Estimation for Financial Time Series [7] have also been developed.…”

Text and Hypertext Categorization

“…TC is used in many application contexts, ranging from automatic document indexing based on a controlled vocabulary (Borko and Bernick 1963;Gray and Harley 1971;Field 1975), to document filtering (Amati and Crestani 1999;Iyer, Lewis et al 2000;Kim, Hahn et al 2000), word sense disambiguation (Gale, Church et al 1992;Escudero, Marquez et al 2000), population of hierarchical catalogues of Web resources (Chakrabarti, Dom et al 1998;Attardi, Gulli et al 1999;Oh, Myaeng et al 2000), and in general any application requiring document organization or selective and adaptive document dispatching.…”