tailieunhanh - Báo cáo khoa học: "Part of Speech Tagging Using a Network of Linear Separators"

We present an architecture and an on-line learning algorithm and apply it to the problem of part-ofspeech tagging. The architecture presented, SNOW, is a network of linear separators in the feature space, utilizing the Winnow update algorithm. Multiplicative weight-update algorithms such as Winnow have been shown to have exceptionally good behavior when applied to very high dimensional problems, and especially when the target concepts depend on only a small subset of the features in the feature space. . | Part of Speech Tagging Using a Network of Linear Separators Dan Roth and Dmitry Zelenko Department of Computer Science University of Illinois at Urbana-Champaign 1304 w Springfield Ave. Urbana IL 61801 danr zelenko Abstract We present an architecture and an on-line learning algorithm and apply it to the problem of part-of-speech tagging. The architecture presented SNOW is a network of linear separators in the feature space utilizing the Winnow update algorithm. Multiplicative weight-update algorithms such as Winnow have been shown to have exceptionally good behavior when applied to very high dimensional problems and especially when the target concepts depend on only a small subset of the features in the feature space. In this paper we describe an architecture that utilizes this mistake-driven algorithm for multi-class prediction - selecting the part of speech of a word. The experimental analysis presented here provides more evidence to that these algorithms are suitable for natural language problems. The algorithm used is an on-line algorithm every example is used by the algorithm only once and is then discarded. This has significance in terms of efficiency as well as quick adaptation to new contexts. We present an extensive experimental study of our algorithm under various conditions in particular it is shown that the algorithm performs comparably to the best known algorithms for POS. 1 Introduction Learning problems in the natural language domain often map the text to a space whose dimensions are the measured features of the text . its words. Two characteristic properties of this domain are that its dimensionality is very high and that both the learned concepts and the instances reside very sparsely in the feature space. In this paper we present a learning algorithm and an architecture with properties suitable for this domain. The SNOW algorithm presented here builds on recently introduced theories of multiplicative weight-updating learning .

• Báo cáo khoa học
• Part of Speech Tagging
• a Network of Linear Separators
• Long Papers
• báo cáo khoa học
• mô hình ngôn ngữ
• xử lý ngôn ngữ tự nhiên
• Journal of Science and Technology
• A formula to calculate pruning threshold
• The part of speech tagging problem
• The Wall Street Journal
• Hidden Markov model
• Fast and Robust
• Dynamic Model Selection
• báo cáo ngôn ngữ
• ngôn ngữ tự nhiên
• A Cost Sensitive Part of Speech Tagging
• Differentiating Serious Errors
• Minor Errors
• Capturing Paradigmatic
• Syntagmatic Lexical Relations
• Towards Accurate Chinese Part of Speech Tagging
• A Stacked Sub Word Model
• Joint Chinese Word Segmentation
• Efficient Optimization
• an MDL Inspired Objective Function
• Unsupervised Part of Speech Tagging
• Part of Speech Tagging for Twitter
• Annotation
• Features
• and Experiments
• Semisupervised condensed nearest neighbor
• Simultaneous Tokenization
• Part of Speech Tagging for Arabic
• Morphological Analyzer
• Minimized Models
• Bilingual Graph Based Projections
• Dipanjan Das
• A Cascaded Linear Model
• A Fully Bayesian
• Sharon Goldwater
• Unsupervised Part of Speech
• Tagging Employing Efficient Graph Clustering
• Chris Biemann
• Automatic Part of Speech Tagging for Bengali
• An Approach for Morphologically Rich Languages
• Poor Resource Scenario
• Arabic Tokenization
• Morphological Disambiguation in One Fell Swoop
• Considering Surface Form
• an Agglutinative Language
• Self Organizing Markov Models
• Their Application to Part of Speech Tagging
• Hae Chang Rim Jin Dong Kim
• Machine Aided Error Correction Environment
• Korean Morphological Analysis
• A VARIABLE MEMORY MARKOV MODEL
• Hinrich Schiitze
• Revision Learning
• its Application to Part of Speech Tagging
• Tetsuji Nakagawa
• Feature Rich Part of speech Tagging
• Morphologically Complex Languages
• Application to Bulgarian
• scientific reports
• model language
• process natural language
• Weakly Supervised Part of Speech Tagging
• Morphologically Rich
• Resource Scarce Languages
• Categorial Fluidity
• Chinese and its Implications