Báo cáo khoa học: "Enhancing Unlexicalized Parsing Performance using a Wide Coverage Lexicon, Fuzzy Tag-set Mapping, and EM-HMM-based Lexical Probabilities" ppt

... 4), and in a more re- alistic one in which parsing and segmentation are handled jointly by the parser (Goldberg and Tsar- faty, 2008) (Sec. 5). External lexical informa- tion enhances unlexicalized ... training data. count(·) is a counting function over the training data, rare stands for any rare event, and w rare is a specific rare event. KCA(·) is the KC Analyzer function...

Ngày tải lên: 17/03/2014, 22:20

... standard, and computed precision and recall figures over the dependencies. Recall that a dependency is defined as a 4-tuple: a head of a functor, a functor category, an argument slot, and a head ... the data. If a word ap- pears at least K times in the data, the supertagger only considers categories that appear in the word’s category set, rather than all lexical categories...

Ngày tải lên: 31/03/2014, 06:20

... between a compact grammar and useful markov histories. 3 External vs. Internal Annotation The two major previous annotation strategies, par- ent annotation and head lexicalization, can be seen as ... motivates various class- or similarity- based approaches to combating sparseness, and this remains a promising avenue of work, but success in this area has proven somewhat elusive, and...

Ngày tải lên: 08/03/2014, 04:22

... and citations therein); (ii) compu- tational ease—with just around a thousand cate- gories, the word–category matrix has a manage- able size; (iii) widespread availability—thesauri are available ... thesaurus. Since human an- notation is both expensive and time intensive, we present an alternative approach of artificially gen- erating thesaurus-sense-tagged data following the ideas of...

Ngày tải lên: 08/03/2014, 21:20

... similarly as the standard shift-reduce parsing algorithm. In the training phase, each target-style parse tree in the training data is transformed into a binary tree (Charniak et al., 1998) and then ... choose actions for state transition. Moreover, beam search strategies can be used to expand the search space of a shift-reduce-based heterogeneous parser (Sagae and Lavie, 200 6a) . T...

Ngày tải lên: 17/03/2014, 00:20

... data comes from a different source than the train- ing data, and will allow the algorithm to adapt to the new context. For example, a language acquisition system with a tagger trained on a ... and eval- uate its performance under various conditions. In the second set SNOW is compared with a naive Bayes algorithm and with Brill's TBL, all trained and tested o...

Ngày tải lên: 23/03/2014, 19:20

... assumptions for the static tag probabilities, are encouraging. VARIABLE MEMORY MARKOV MODELS Markov models are a natural candidate for lan- guage modeling and temporal pattern recognition, ... fixed-length histories, variable memory Markov models dynamically adapt their history length based on the training data, and hence may use fewer parameters. In a test of a VMM based t...

Ngày tải lên: 23/03/2014, 20:21

... inhibitors of ADP and thromboxane A 2 , both saratin and 6F1, a blocking a 2 b 1 mAb, abro- gated platelet adhesion to fibrillar and soluble collagen. Additionally, sara- tin eliminated a 2 b 1 -dependent ... kDa leech antiplatelet protein isola- ted from Haementeria officinalis) and calin and sara- tin (approximately 65 kDa and 12 kDa proteins, respectively, both isolated f...

Ngày tải lên: 30/03/2014, 09:20

... U.K. Abstract This paper describes a wide- coverage Japanese grammar based on HPSG. The aim of this work is to see the coverage and accuracy attain- able using an underspecified grammar. Under- ... when a relative clause modifies a phrase. Head-marker schema Applied when a marker like a postposition marks a phrase. Head-adjacent schema Applied when a suffix att...

Ngày tải lên: 31/03/2014, 04:20

... thresholds, and results are shown in Table 1. Since a gold stan- dard treebank for our data set was available, it was possible to evaluate the accuracy of the parser. Eval- uation of deep parsing ... likely to have lexicalised grammars) as a POS tagger can massively increase the parser coverage on unseen text. While annotat- ing with named entity data or a lexical type supertag-...

Ngày tải lên: 17/03/2014, 02:20