... have mainly employed bottom- up parsingalgorithms for processing. Conventional prediction techniques to improve the efficiency of the • parsing process, appear to fall short when parsing CUG. ... Prediction inChartParsingAlgorithms for Categorial Unification Grammar Gosse Bouma Computational Linguistics Department University of Groningen, P.O. box 716 NL-9700 AS Groningen, The ... only in this way the effect of top-down filtering will pay-off against the increased overhead of having to check the left-corner table. 6 Some Results The performance of the parsing algorithms...
... Survey of Rule-Invocation Strategies This section surveys the fundamental rule-invocation strategies in context-flee chart parsing. 3 In a chart- parsing framework, different rule-invocation strate- ... empirically. The experiments clearly indicate that it is possible to significantly increase efficiency inchartparsing by fine-tuning the rule-invocation strategy. Fine-tuning however also requires ... being generated, though. 228 Kay (1982) is the principal source, providing a very general exposition of the control strategies and data structures involved inchart parsing. In con- sidering...
... which can attach into other (mini)charts where combinations are 1In effect, hypotheticals belong on additional subor- derings, which can connect into the main ordering of the chart at various ... and so the indexing system allows the encoding of dominance relations. 4 A New ChartParsing Method for Lambek Grammars 4.1 Lambek to SLMG Conversion The first task of the parsing approach ... present a new chartparsing method for Lambek grammars, inspired by a method for D- Tree grammar parsing. The formulae of a Lam- bek sequent are firstly converted into rules of an indexed grammar...
... crossing (s,t)}[ : the number of constituents in TG correct according to Consistent Brackets. Following are the definitions of the six metrics used in this paper for evaluating binary branching ... grammar with nonterminals not comparable to those in the training, a different experiment is needed to evaluate the Labelled Recall Algorithm, one in which the nonterminals in the induced grammar ... nonterminals in the test set. 5.2.1 Grammar Induction by Counting For this experiment, a very simple grammar was induced by counting, using a portion of the Penn Tree Bank, version 0.5. In particular,...
... entropy model forpart-of- sp e e c h tagging. In Proc. EMNLP.B. Taskar, D. Klein, M. Collins, D. Koller, and C. Man-ning. 2004. Max-margin parsing. In Proc. EMNLP.H. Yamada and Y. Matsumoto. ... large-margin dependency2We are not concerned with violating the tree constraint. parsing. As usual for supervised learning, we as-sume a training set T = {(xt, yt)}Tt=1, consist-ing of ... are very promising andfurther show the robustness of discriminative on-line learning with approximate parsing algorithms. 6 DiscussionWe described approximate dependency parsing al-gorithms...
... collaborative-filtering techniques. In content-based filtering, wesolve the problem of data sparsity in online con-tent by smoothing using latent topic information. In collaborative filtering, we model ... mutual information between users afterdoing an average-link clustering on their pairwisemutual information. In a clean clustering, intra-cluster mutual information should be high, whileinter-cluster ... using the bag-of-word model sincethat would require the exact words to appear in thetraining set. In order to take advantage of the topic level in- formation while not losing the “fine-grained”...
... Klein and Christopher D. Manning. 2003a. Ac-curate unlexicalized parsing. In Proceedings of the41st Annual Meeting of the Association for Com-putational Linguistics, pages 423–430.Dan Klein ... D. Manning. 2003b. Fastexact inference with a factored model for naturallanguage parsing. In Advances in Neural Informa-tion Processing Systems, volume 15. MIT Press.Ravi Kumar, Jasmine Novak, ... are not used in ourclassification experiments. We then pruned thelist of interests by excluding any interest thatoccurred within a +/-10 window (based on itsposition in the list) in multiple...
... training speed of discrimina-tive parsing. Our discriminative parsing method has no generative component, yetsurpasses a generative baseline on con-stituent parsing, and does so with mini-mal ... & Norvig, 1995).3 Training Method3.1 General SettingOur training set I consists of candidate inferencesfrom the parse trees in the training data. Fromeach training inference i ∈ I we generate ... development set as training progressed. Eventhough the baseline training made progress morequickly than the kitchen sink, the kitchen sink’s F1surpassed the baseline’s F1early in training, and at6.3K...
... and to outline how this could be embedded in an interactive parsing system. 1.2 Incremental Parsing The word "incremental ~ has been used in two dif- fering senses in the (parsing) literature. ... gives a brief outline of the combined interactive and incremental parsing system, and section 5 summarizes the con- clusions. 2 Incremental ChartParsing 2.1 ChartParsing The incremental parser ... algorithm for incremental chart parsing, outlines how this could be embed- ded in an interactive parsing system, and discusses why this might be useful. Incremental parsing here means that input...
... beneveling. After likelihood determination and or- dering of the two remaining analyses in (8), the cor- rect analysis be + nevel + ing is in topmost position: (14) 1 be -t- nevel 4- ing N ... Head Rule be combined into a word of category B 4. In addition to this general rule for com- pounding, the grammar contains a small set of rules defining productive compounding. An analysis ... tion) is shown in (3)z: (3) a be + neef + eling b be+neef+e+ling c be + nevel + ing d been + e + veel + ing e be+n +e+veel+ing f be +neef + eel + ing All of the parts in the segmentations...
... using Collin’sparser (Vadas and Curran, 2007c), but has yet to beanalyzed for its effects on dependency parsing. Parsing is very early in the NLP pipeline. There-fore, improvements inparsing ... components insteadof improperly being a leaf node.5 ConclusionThis paper has demonstrated the benefit of addi-tional noun phrase bracketing in training data for use in dependency parsing and machine ... Computational LinguisticsEffects of Noun Phrase Bracketing in Dependency Parsing and MachineTranslationNathan GreenCharles University in PragueInstitute of Formal and Applied LinguisticsFaculty...
... consider-able speedups.1 IntroductionThe increasing demand for accuracy and ro-bustness for today’s unification-based gram-mar parsers brings on an increasing demandfor computing power. In addition, as ... Engineering, 6(1):1–18.[Caroll1994] John Caroll. 1994. Relating complex-ity to practical performance inparsing withwide-coverage unification grammars. In Proc.of the 32ndAnnual Meeting ... non-empty. In this case, newWork willbe set and step 2 is executed again. This cyclecontinues until all work is completed.The remaining steps serve several purposes:load balancing, preventing double...