... proposes a new web parallel data mining scheme. Given a pair of parallel web pages as seeds, the Document Object Model 1 (DOM) is used to represent theweb pages as a pair of DOM trees. Then a ... Pietra, and R. L. Mercer. 1993. The Mathematics of Statistical Ma-chine Translation: Parameter Estimation. Computa-tional Linguistics, V19(2). Callison-Burch, C. and C. Bannard. 2005. Paraphras-ing ... web pages manually labeled as parallel or non-parallel. The Iterative Scaling algorithm (Pietra, Pietra and Lafferty 1995) is used forthe training. 7 Experimental Results The DOM tree alignment...
... NounPhrase (baseNP) is an important subtask for many natural language processing applications,such as partial parsing, information retrieval andmachine translation. A baseNP is a simple nounphrase ... Treebank II,and the definition of baseNP is the same asRamshaw’s, Table 1 summarizes the averageperformance on both baseNP tagging and POStagging, each section of the whole PennTreebank was ... three other approaches tobaseNP identification have been evaluated usingPenn Treebank-Ramshaw & Marcus’stransformation-based chunker, Argamon et al.’sMBSL, and Cardie’s Treebank_lex in Table...
... SR for each pair. The reader can consult Budanitskyand Hirst (2006) to confirm that all the other mea-sures of semantic relatedness we compare to, donot follow the same pattern as the human ratings,as ... 2) a new GVSM re-trieval model, which incorporates the aforemen-tioned semantic relatedness measure; 3) exploita-tion of all thesemantic information a thesauruscan offer, including semantic ... ratings,as closely as our measure of relatedness does (lowy values for small x values and high y values for high x). The same pattern applies in the M&C and353-C data sets.4.2 Evaluation of the...
... panel-data analytic techniques. As more years of data on individual vehicles become available, it may be advan-tageous to adopt a panel-data approach. xii Improving Recapitalization Planning: Toward ... Management Modelforthe HMMWVTable 4.1Fleet Management Model Assumptions in Sensitivity Analyses and Base CaseReplace Earlier Base Case Replace LaterNon-EDA costs Vary in same way as EDA ... for Army Analysis (CAA) (East, 2002), drew on the CBO figure of 1 to 3 percent to build a mathematical model optimizing Army RECAP rates. Specifically, CAA used an estimated age escalation factor...
... transition-based strategy, the integrated algorithm we are looking for has tobe transition-based at the top level. The advan-tages of the graph-based approach – a more glob-ally informed basis forthe decision ... punctuation marks for these corporaand follow in that the evaluation schema of the CoNLL Shared Task 2009. Table 3 presents the results as obtained for these data set. The transition-based parser ... the incremental calculation of the scores of the com-pletion model, and the parallel feature extractionas well as the parallelized transition-based pars-ing strategy play an important role in carrying...
... notuse the lambda calculus formalism to define our taskbut rather treat it as an instance of frame -semantic parsing, or a specific type of semantic role label-ing (Gildea and Jurafsky, 2002). The ... monoclonal antibody, ab, an-tisera, mab12 tnfalpha, tnf-alpha, il-6, tnfTable 1: Examples of the induced semantic classes.realizations have a clear semantic connection. Clus-ter 6, for example, ... 2005. A statistical semantic parser that integrates syntax and semantics.In Proceedings of the Ninth Conference on Computa-tional Natural Language Learning (CONLL-05), AnnArbor, Michigan.Daniel...
... consid-ered as a syntactic constraint. Therefore wecan use thousands of syntactic constraints toguide phrase translation.ã The SDB model maintains and protects the strength of the phrase-based approach ... in a better way than the CMVC does. It is able toreward non-syntactic translations by assign-ing an adequate probability to them if thesetranslations are appropriate to particular syn-tactic ... to the word align-ments, we define bracketable and unbracketableinstances. For each of these instances, we auto-matically extract relevant syntactic features from the source parse tree as bracketing...
... is the fact that thereis often a good deal of overlap in words between the reparandum and the alteration, as speakers maytrace back several words when restarting after an er-ror. For instance, ... modified for use in a specialrepair grammar, which not only reduces the amountof available training data, but violates our intuitionthat most reparanda are fluent up until the actual editoccurs. The ... Communication Re-search Centre, University of Edinburgh.John Hale, Izhak Shafran, Lisa Yung, Bonnie Dorr, MaryHarper, Anna Krasnyanskaya, Matthew Lease, YangLiu, Brian Roark, Matthew Snover, and...
... groups and domains can be modeled separately without accessing and adapting the language model of the MT system for each SMS application. Another advantage is that the normalization module can ... normalization as a translation problem from the SMS language to the English language1 and we propose to adapt a phrase-based statistical MT modelforthe task. Evaluation by 5-fold cross validation ... Normalization We view the SMS language as a variant of Eng-lish language with some derivations in vocabu-lary and grammar. Therefore, we can treat SMS normalization as a MT problem where the...