TAILIEUCHUNG - Báo cáo khoa học: "A Comparison of Loopy Belief Propagation and Dual Decomposition for Integrated CCG Supertagging and Parsing"

Via an oracle experiment, we show that the upper bound on accuracy of a CCG parser is significantly lowered when its search space is pruned using a supertagger, though the supertagger also prunes many bad parses. Inspired by this analysis, we design a single model with both supertagging and parsing features, rather than separating them into distinct models chained together in a pipeline. | A Comparison of Loopy Belief Propagation and Dual Decomposition for Integrated CCG Supertagging and Parsing Michael Auli School of Informatics University of Edinburgh Adam Lopez hltcoE Johns Hopkins University alopez@ Abstract Via an oracle experiment we show that the upper bound on accuracy of a CCG parser is significantly lowered when its search space is pruned using a supertagger though the supertagger also prunes many bad parses. Inspired by this analysis we design a single model with both supertagging and parsing features rather than separating them into distinct models chained together in a pipeline. To overcome the resulting increase in complexity we experiment with both belief propagation and dual decomposition approaches to inference the first empirical comparison of these algorithms that we are aware of on a structured natural language processing problem. On CCGbank we achieve a labelled dependency F-measure of on gold POS tags and on automatic part-of-speeoch tags the best reported results for this task. 1 Introduction Accurate and efficient parsing of Combinatorial Categorial Grammar CCG Steedman 2000 is a longstanding problem in computational linguistics due to the complexities associated its mild context sensitivity. Even for practical CCG that are strongly context-free Fowler and Penn 2010 parsing is much harder than with Penn Treebank-style context-free grammars with vast numbers of nonterminal categories leading to increased grammar constants. Where a typical Penn Treebank grammar may have fewer than 100 nonterminals Hockenmaier and Steedman 2002 we found that a CCG grammar derived from CCGbank contained over 1500. The 470 same grammar assigns an average of 22 lexical categories per word Clark and Curran 2004a resulting in an enormous space of possible derivations. The most successful approach to CCG parsing is based on a pipeline strategy 2 . First we tag or multitag each word of the sentence with a lexical

• Báo cáo khoa học
• A Comparison
• Loopy Belief Propagation and Dual Decomposition
• Integrated CCG Supertagging and Parsing
• báo cáo khoa học
• báo cáo ngôn ngữ
• ngôn ngữ tự nhiên
• Information Sciences
• Relative deviations in pairwise comparison matrices
• Pairwise comparison matrices
• Relative deviation interconnection
• Pairwise comparison matrix
• Nuclear engineering and technology
• Comparison of proliferation resistance among natural uranium
• Comparison of thoriumeuranium
• Comparison of thoriumeplutonium fuels
• CANada Deuterium Uranium in deep geological repository
• Combining multiattribute utility analysis with transport model
• International Journal of Computer Networks and Communications Security
• The optical communications
• Simulating an optical high debit transmission chain
• Optisystem with comparison of optical windows
• Comparison of optical windows
• BMC Cancer
• Indirect comparison
• Matching adjusted comparison
• Prior sunitinib treated patients
• Metastatic renal cell carcinoma
• Mass media
• Tribal groups based
• Social comparison of tribal groups based
• Social comparison
• Wadi project adoption
• BMC Bioinformatics
• Friedman test
• Exact p value
• Rank sum difference
• Multiple comparison
• Nonparametric statistics
• Classifier comparison
• Machine learning
• Review study on design comparison
• Design comparison
• Solar dryer cum solar cooker
• Solar cooking
• Solar drying
• Agricultural products
• Triticum aestivum
• New insecticides
• Termites on wheat
• Field conditions
• Comparison to yield
• Wheat in comparison
• Two new criteria mathematical knowledge
• comparison in the Bruhat order
• mathematical works
• Even circuits
• prescribed clockwise parity
• scientific reports
• scientific research
• medical research
• therapeutic
• pharmaceutical
• Randomized Comparison
• Radial Artery
• Saphenous Vein
• Coronary Bypass Grafts
• trình bày báo cáo
• tài liệu báo cáo khoa học
• nghiên cứu y học
• báo cáo y học
• kiến thức y học
• Chinese Parsers
• Stanford Dependencies
• Measuring Language Divergence
• Intra Lexical Comparison
• Simon Kirby T
• Mark Ellison
• mô hình ngôn ngữ
• xử lý ngôn ngữ tự nhiên
• Models for Sentence Compression
• A Comparison across Domains
• Training Requirements
• Evaluation Measures
• A Comparison and Semi Quantitative Analysis
• Words and Character Bigrams as Features
• Chinese Text Categorization
• A Comparison of Document
• Sentence
• Term Event Spaces
• A Comparison of Alternative Parse
• Tree Paths
• Labeling Semantic Roles
• Paragraph
• word
• and coherence based approaches to sentence ranking
• A comparison of algorithm
• human performance
• A comparison of three adaptive behaviour measures
• Cognitive level of autism
• Severity level of autism
• Assess adaptive behaviours
• VABS Classroom edition
• Comparison
• CFG filtering techniques
• LTAG and HPSG
• On the Evaluation and Comparison of Taggers
• the Effect of Noise
• Testing Corpora
• A Practical Comparison
• Parsing Strategies
• Jonathan Slocum Siemens
• model language
• process natural language
• FACTORING RECURSION
• DEPENDENCIES
• AN ASPECT OF TREE ADJOINING GRAMMARS (TAG) AND A COMPARISON OF SOME FORMAL PROPERTIES OF TAGS
• GPSGS
• PLGS
• AND LPGS