acl acl2013 acl2013-335 knowledge-graph by maker-knowledge-mining
Source: pdf
Author: Angelina Ivanova ; Stephan Oepen ; Lilja vrelid
Abstract: In this paper we focus on practical issues of data representation for dependency parsing. We carry out an experimental comparison of (a) three syntactic dependency schemes; (b) three data-driven dependency parsers; and (c) the influence of two different approaches to lexical category disambiguation (aka tagging) prior to parsing. Comparing parsing accuracies in various setups, we study the interactions of these three aspects and analyze which configurations are easier to learn for a dependency parser.
Reference: text
sentIndex sentText sentNum sentScore
1 Survey on parsing three dependency representations for English Angelina Ivanova Stephan Oepen Lilja Øvrelid University of Oslo, Department of Informatics { ange l i| oe | l l ao } @ i . [sent-1, score-0.145]
2 no i i j fi Abstract In this paper we focus on practical issues of data representation for dependency parsing. [sent-3, score-0.086]
3 We carry out an experimental comparison of (a) three syntactic dependency schemes; (b) three data-driven dependency parsers; and (c) the influence of two different approaches to lexical category disambiguation (aka tagging) prior to parsing. [sent-4, score-0.196]
4 Comparing parsing accuracies in various setups, we study the interactions of these three aspects and analyze which configurations are easier to learn for a dependency parser. [sent-5, score-0.212]
5 1 Introduction Dependency parsing is one of the mainstream research areas in natural language processing. [sent-6, score-0.059]
6 There were several shared tasks organized on dependency parsing (CoNLL 2006–2007) and labeled dependencies (CoNLL 2008–2009) and there were a number of attempts to compare various dependencies intrinsically, e. [sent-11, score-0.227]
7 In this paper we focus on practical issues of data representation for dependency parsing. [sent-18, score-0.086]
8 , 2005) and the parser of Bohnet and Nivre (2012); (c) two approaches to wordcategory disambiguation, e. [sent-22, score-0.114]
9 We parse the formats and compare accuracies in all configurations in order to determine how parsers, dependencyrepresentations andgrammatical tagging methods interact with each other in application to automatic syntactic analysis. [sent-27, score-0.123]
10 SB and CD are derived automatically from phrase structures of Penn Treebank to accommodate the needs of fast and accurate dependency parsing, whereas DT is rooted in the formal grammar theory HPSG and is independent from any specific treebank. [sent-28, score-0.126]
11 For DT we gain more expressivity from the underlying linguistic theory, which challenges parsing with statistical tools. [sent-29, score-0.078]
12 The structural analysis of the schemes in Ivanova et al. [sent-30, score-0.034]
13 We recompute similarities on a larger treebank and check whether parsing results reflect them. [sent-32, score-0.086]
14 (2012) investigate which dependency representations of several syntactic structures are easier to parse with supervised versions of the Klein and Manning (2004) parser, ClearParser (Choi and Nicolov, 2009), MST Parser, Malt and the Easy First Non-directional parser (Goldberg and Elhadad, 2010). [sent-35, score-0.246]
15 The results imply that all parsers consistently perform better when (a) coordination has one of the conjuncts as the head rather than the coordinating conjunction; 31 Sofia, BulPgraoricea,ed Ainug us otf 4 t-h9e 2 A0C13L. [sent-36, score-0.22]
16 (2010) present intristic and extristic (event-extraction task) evaluation of six parsers (GDep, Bikel, Stanford, Charniak-Johnson, C&C; and Enju parser) on three dependency formats (Stanford Dependencies, CoNLL-X, and Enju PAS). [sent-41, score-0.24]
17 Intristic evaluation results show that all parsers have the highest accuracies with the CoNLL-X format. [sent-42, score-0.1]
18 2 Parsers In the experiments described in Section 4 we used parsers that adopt different approaches and implement various algorithms. [sent-53, score-0.072]
19 , 2007): transition-based dependency parser with local learning and greedy search. [sent-55, score-0.2]
20 Bohnet and Nivre (2012) parser: transitionbased dependency parser with joint tagger that implements global learning and beam search. [sent-58, score-0.2]
21 3 Dependency schemes In this work we extract DeepBank data in the form of bilexical syntactic dependencies, DELPH-IN Syntactic Derivation Tree (DT) format. [sent-60, score-0.095]
22 We obtain the exact same sentences in Stanford Basic (SB) format from the automatic conversion of the PTB with the Stanford parser (de Marneffe et al. [sent-61, score-0.175]
23 SB and CD represent the way to convert PTB to bilexical dependencies; in contrast, DT is grounded in linguistic theory and captures decisions taken in the grammar. [sent-63, score-0.037]
24 Figure 1demonstrates the differences between the formats on the coordination structure. [sent-64, score-0.121]
25 (2012), analysis of coordination in SB and CD is easier for a statistical parser to learn; however, as we will see in section 4. [sent-66, score-0.203]
26 4 Part-of-speech tags We experimented with two tag sets: PTB tags and lexical types of the ERG grammar - supertags. [sent-69, score-0.273]
27 PTB tags determine the part of speech (PoS) and some morphological features, such as number for nouns, degree of comparison for adjectives and adverbs, tense and agreement with person and number of subject for verbs, etc. [sent-70, score-0.126]
28 There are 48 tags in the PTB tagset and 1091 supertags in the set of lexical types of the ERG. [sent-79, score-0.707]
29 The state-of-the-art accuracy of PoS-tagging on in-domain test data using gold-standard tokenization is roughly 97% for the PTB tagset and approximately 95% for the ERG supertags (Ytrestøl, 2011). [sent-80, score-0.596]
30 Supertagging for the ERG grammar is an ongoing research effort and an off-the-shelf supertagger for the ERG is not currently available. [sent-81, score-0.044]
31 4 Experiments In this section we give a detailed analysis of parsing into SB, CD and DT dependencies with Malt, MST and the Bohnet and Nivre (2012) parser. [sent-82, score-0.1]
32 1 Setup For Malt and MST we perform the experiments on gold PoS tags, whereas the Bohnet and Nivre (2012) parser predicts PoS tags during testing. [sent-84, score-0.4]
33 Prior to each experiment with Malt, we used MaltOptimizer to obtain settings and a feature model; for MST we exploited default configuration; for the Bohnet and Nivre (2012) parser we set the beam parameter to 80 and otherwise employed the default setup. [sent-85, score-0.114]
34 With regards to evaluation metrics we use labelled attachment score (LAS), unlabeled attachment score (UAS) and label accuracy (LACC) excluding punctuation. [sent-86, score-0.072]
35 Statistical significance was assessed using Dan Bikel’s parsing evaluation comparator1 at the 0. [sent-91, score-0.059]
36 We inspect three different aspects in the interpretation of these results: parser, dependency format and tagset. [sent-93, score-0.128]
37 From the parser perspective Malt and MST are not very different in the traditional setup with gold 1http : / / next en s . [sent-95, score-0.221]
38 nl / deppar s e-wiki / S o ftwareP age # s coring PTB tags (Table 1, Gold PTB tags). [sent-97, score-0.154]
39 The Bohnet and Nivre (2012) parser outperforms Malt on CD and DT and MST on SB, CD and DT with PTB tags even though it does not receive gold PTB tags during test phase but predicts them (Table 2, Predicted PTB tags). [sent-98, score-0.507]
40 This is explained by the fact that the Bohnet and Nivre (2012) parser implements a novel approach to parsing: beam-search algorithm with global structure learning. [sent-99, score-0.114]
41 MST “loses” more than Malt when parsing SB with gold supertags (Table 1, Gold supertags). [sent-100, score-0.707]
42 This parser exploits context features “POS tag of each intervening word between head and dependent” (McDonald et al. [sent-101, score-0.139]
43 Due to the far larger size of the supertag set compared to the PTB tagset, such features are sparse and have low fre- quencies. [sent-103, score-0.042]
44 This leads to the lower scores of parsing accuracy for MST. [sent-104, score-0.059]
45 For the Bohnet and Nivre (2012) parser the complexity of supertag prediction has significant negative influence on the attachment and labeling accuracies (Table 2, Predicted supertags). [sent-105, score-0.22]
46 The addition of gold PTB tags as a feature lifts the performance of the Bohnet and Nivre (2012) parser to the level of performance of Malt and MST on CD with gold supertags and Malt on SB with gold supertags (compare Table 2, Predicted supertags + gold PTB, and Table 1, Gold supertags). [sent-106, score-2.291]
47 Both Malt and MST benefit slightly from the combination of gold PTB tags and gold supertags (Table 1, Gold PTB tags + gold supertags). [sent-107, score-1.114]
48 For the Bohnet and Nivre (2012) parser we also observe small rise of accuracy when gold supertags are provided as a feature for prediction of PTB tags (compare Predicted PTB tags and Predicted PTB tags + gold supertags sections of Table 2). [sent-108, score-1.788]
49 The parsers have different running times: it takes minutes to run an experiment with Malt, about 2 hours with MST and up to a day with the Bohnet and Nivre (2012) parser. [sent-109, score-0.072]
50 From the point of view of the dependency format, SB has the highest LACC and CD is first-rate on UAS for all three parsers in most of the configurations (Tables 1 and 2). [sent-110, score-0.175]
51 This means that SB is easier to label and CD is easier to parse structurally. [sent-111, score-0.044]
52 DT appears to be a more difficult target format because it is both hard to label and attach in most configurations. [sent-112, score-0.042]
53 It is not an unexpected result, since SB and CD are both derived from PTB phrase-structure trees and are oriented to ease dependency parsing task. [sent-113, score-0.145]
54 DT is not custom-designed 33 Gold PTB tags Predicted PTB tags SDCBT8 M95a. [sent-114, score-0.252]
55 Gold PTB tags: Malt and MST are trained and tested on gold PTB tags. [sent-145, score-0.107]
56 Gold supertags: Malt and MST are trained and tested on gold supertags. [sent-146, score-0.107]
57 Gold PTB tags + gold supertags: Malt and MST are trained on gold PTB tags and gold supertags. [sent-147, score-0.573]
58 1 denotes a feature model in which gold PTB tags function as PoS and gold supertags act as additional features (in CPOSTAG field); 2 stands for the feature model which exploits gold supertags as PoS and uses gold PTB tags as extra features (in CPOSTAG field). [sent-148, score-1.762]
59 C816 7C2) Table 2: Parsing results of the Bohnet and Nivre (2012) parser on Stanford Basic (SB), CoNLL Syntactic Dependencies (CD) and DELPH-IN Syntactic Derivation Tree (DT) formats. [sent-152, score-0.114]
60 Predicted PTB: parser predicts PTB tags during the test phase. [sent-155, score-0.274]
61 Predicted supertags: parser predicts supertags during the test phase. [sent-156, score-0.689]
62 Predicted PTB + gold supertags: parser receives gold supertags as feature and predicts PTB tags during the test phase. [sent-157, score-1.029]
63 Predicted supertags + gold PTB: parser receives PTB tags as feature and predicts supertags during test phase. [sent-158, score-1.463]
64 to dependency parsing and is independent from parsing questions in this sense. [sent-159, score-0.204]
65 Contrary to our expectations, the accuracy scores of parsers do not suggest that CD and DT are particularly similar to each other in terms of parsing. [sent-166, score-0.072]
66 Inspecting the aspect of tagset we conclude that traditional PTB tags are compatible with SB and CD but do not fit the DT scheme well, while ERG supertags are specific to the ERG framework and do not seem to be appropriate for SB and CD. [sent-167, score-0.707]
67 Neither of these findings seem surprising, as PTB tags were developed as part of the treebank from which CD and SB are derived; whereas ERG supertags are closely related to the HPSG syntactic structures captured in DT. [sent-168, score-0.737]
68 PTB tags were designed to simplify PoS-tagging whereas supertags were developed to capture information that is required to analyze syntax of HPSG. [sent-169, score-0.686]
69 For each PTB tag we collected corresponding supertags from the gold-standard training set. [sent-170, score-0.541]
70 For open word classes such as nouns, adjectives, adverbs and verbs the relation between PTB tags and supertags is many-to-many. [sent-171, score-0.687]
71 Thus, supertags do not provide extra level of detalization for PTB tags, but PTB tags and supertags are complementary. [sent-173, score-1.208]
72 For this reason their combination results in slight increase of accuracy for all three parsers on all dependency formats (Table 1, Gold PTB tags + gold supertags, and Table 2, Predicted PTB + gold supertags and Predicted supertags + gold PTB). [sent-176, score-1.741]
73 The Bohnet and Nivre (2012) parser predicts supertags with an average accuracy of 89. [sent-177, score-0.689]
74 3 Error analysis Using the CoNLL-07 evaluation script2 on our test set, for each parser we obtained the error rate distribution over CPOSTAG on SB, CD and DT. [sent-182, score-0.114]
75 We automatically collected all sentences that contain 1) attachment errors, 2) label errors, 3) attachment and label errors for VBP, VBZ and VBG made by Malt parser on DT format with PTB PoS. [sent-185, score-0.251]
76 For each of these three lexical categories we manually analyzed a random sample of sentences with errors and their corresponding gold-standard versions. [sent-186, score-0.023]
77 In many cases such errors are related to the root of the sentence when the verb is either treated as complement or adjunct instead of having a root status or vice versa. [sent-187, score-0.023]
78 Sentences with coordination are particularly difficult for the correct attachment and labeling of the VBP (see Figure 2 for an example). [sent-189, score-0.103]
79 The error rate of Malt, MST and the Bohnet and Nivre (2012) parser for the coordination is not so high for SB and CD ( 1% and 2% correspondingly with MaltParser, PTB tags) whereas for DT the error rate on the CPOSTAGS is especially high (26% with MaltParser, PTB tags). [sent-191, score-0.2]
80 It means that there are many errors on incoming dependency arcs for coordinating conjunctions when parsing DT. [sent-192, score-0.219]
81 On outgoing arcs parsers also make more mistakes on DT than on SB and CD. [sent-193, score-0.087]
82 , 2012), it is harder to parse coordination headed by coordinating con- junction. [sent-196, score-0.147]
83 Although the approach used in DT is harder for parser to learn, it has some advantages: using SB and CD annotations, we cannot distinguish the two cases illustrated with the sentences (a) and (b): 2http : / / next ens . [sent-197, score-0.139]
84 nl / deppars e-wiki / So ftwareP age # s coring 35 ro t HD-CSMB-PHDVP-VPMRK-NH The figures VBP show that spending VBD rose 0. [sent-199, score-0.028]
85 In the sentence a) “the national fast-food” refers only to the conjunct “chains”, while in the sentence b) “the planned” refers to both conjuncts and “MGM Grand” refers only to the first conjunct. [sent-207, score-0.043]
86 The Bohnet and Nivre (2012) parser succeeds in finding the correct conjucts (shown in bold font) on DT and makes mistakes on SB and CD in some difficult cases like the following ones: a) <. [sent-208, score-0.114]
87 > investors hoard gold and help underpin its price <. [sent-211, score-0.107]
88 5 Conclusions and future work In this survey we gave a comparative experimental overview of (i) parsing three dependency schemes, viz. [sent-216, score-0.145]
89 , Stanford Basic (SB), CoNLL Syntactic Dependencies (CD) and DELPH-IN Syntactic Derivation Tree (DT), (ii) with three leading dependency parsers, viz. [sent-217, score-0.086]
90 , Malt, MST and the Bohnet and Nivre (2012) parser (iii) exploiting two different tagsets, viz. [sent-218, score-0.114]
91 From the parser perspective, the Bohnet and Nivre (2012) parser performs better than Malt and MST not only on conventional formats but also on the new representation, although this parser solves a harder task than Malt and MST. [sent-220, score-0.421]
92 From the dependency format perspective, DT appeares to be a more difficult target dependency representation than SB and CD. [sent-221, score-0.214]
93 This suggests that the expressivity that we gain from the grammar theory (e. [sent-222, score-0.04]
94 for coordination) is harder to learn with state-of-the-art dependency parsers. [sent-224, score-0.111]
95 PTB tags and supertags are complementary, and for all three parsers we observe slight benefits from being supplied with both types of tags. [sent-226, score-0.739]
96 As future work we would like to run more experiments with predicted supertags. [sent-227, score-0.049]
97 In the absence of a specialized supertagger, we can follow the pipeline of (Ytrestøl, 2011) who reached the stateof-the-art supertagging accuracy of 95%. [sent-228, score-0.023]
98 applied to semantic role labeling and question-answering in order to find out if the usage of the DT format grounded in the computational grammar theory is beneficial for such tasks. [sent-231, score-0.063]
99 K-best, locally pruned, transition-based dependency parsing using robust risk minimization. [sent-245, score-0.145]
100 Corpus-based induction of syntactic structure: models of dependency and constituency. [sent-280, score-0.11]
wordName wordTfidf (topN-words)
[('supertags', 0.541), ('ptb', 0.425), ('sb', 0.295), ('cd', 0.267), ('malt', 0.266), ('dt', 0.216), ('mst', 0.174), ('bohnet', 0.155), ('nivre', 0.127), ('tags', 0.126), ('parser', 0.114), ('gold', 0.107), ('erg', 0.101), ('dependency', 0.086), ('deepbank', 0.083), ('parsers', 0.072), ('coordination', 0.067), ('parsing', 0.059), ('vbp', 0.057), ('cpostag', 0.055), ('ytrest', 0.055), ('formats', 0.054), ('hpsg', 0.053), ('ivanova', 0.049), ('vrelid', 0.049), ('predicted', 0.049), ('format', 0.042), ('supertag', 0.042), ('dependencies', 0.041), ('tagset', 0.04), ('bilexical', 0.037), ('lilja', 0.037), ('coordinating', 0.036), ('attachment', 0.036), ('schemes', 0.034), ('predicts', 0.034), ('conll', 0.033), ('schwartz', 0.032), ('maltparser', 0.028), ('coring', 0.028), ('ftwarep', 0.028), ('intristic', 0.028), ('lacc', 0.028), ('mgm', 0.028), ('accuracies', 0.028), ('treebank', 0.027), ('vbz', 0.026), ('treebanks', 0.026), ('stanford', 0.026), ('harder', 0.025), ('joakim', 0.025), ('head', 0.025), ('derivation', 0.025), ('bunt', 0.025), ('grand', 0.025), ('yakushiji', 0.025), ('syntactic', 0.024), ('pos', 0.024), ('mcdonald', 0.023), ('stroudsburg', 0.023), ('errors', 0.023), ('conjunct', 0.023), ('angelina', 0.023), ('supertagging', 0.023), ('vbg', 0.023), ('enju', 0.023), ('supertagger', 0.023), ('miyao', 0.022), ('easier', 0.022), ('grammar', 0.021), ('miwa', 0.021), ('penn', 0.02), ('adverbs', 0.02), ('flickinger', 0.02), ('bikel', 0.02), ('conjuncts', 0.02), ('oepen', 0.02), ('ichi', 0.02), ('expressivity', 0.019), ('structurally', 0.019), ('headed', 0.019), ('prepositional', 0.019), ('whereas', 0.019), ('vbd', 0.019), ('setups', 0.019), ('conversion', 0.019), ('pa', 0.018), ('tagsets', 0.018), ('configurations', 0.017), ('planned', 0.017), ('stacking', 0.017), ('bernd', 0.017), ('noun', 0.016), ('csli', 0.016), ('uas', 0.016), ('goldberg', 0.015), ('classic', 0.015), ('tokenization', 0.015), ('arcs', 0.015), ('johansson', 0.015)]
simIndex simValue paperId paperTitle
same-paper 1 0.99999976 335 acl-2013-Survey on parsing three dependency representations for English
Author: Angelina Ivanova ; Stephan Oepen ; Lilja vrelid
Abstract: In this paper we focus on practical issues of data representation for dependency parsing. We carry out an experimental comparison of (a) three syntactic dependency schemes; (b) three data-driven dependency parsers; and (c) the influence of two different approaches to lexical category disambiguation (aka tagging) prior to parsing. Comparing parsing accuracies in various setups, we study the interactions of these three aspects and analyze which configurations are easier to learn for a dependency parser.
2 0.14012964 358 acl-2013-Transition-based Dependency Parsing with Selectional Branching
Author: Jinho D. Choi ; Andrew McCallum
Abstract: We present a novel approach, called selectional branching, which uses confidence estimates to decide when to employ a beam, providing the accuracy of beam search at speeds close to a greedy transition-based dependency parsing approach. Selectional branching is guaranteed to perform a fewer number of transitions than beam search yet performs as accurately. We also present a new transition-based dependency parsing algorithm that gives a complexity of O(n) for projective parsing and an expected linear time speed for non-projective parsing. With the standard setup, our parser shows an unlabeled attachment score of 92.96% and a parsing speed of 9 milliseconds per sentence, which is faster and more accurate than the current state-of-the-art transitionbased parser that uses beam search.
3 0.12858488 132 acl-2013-Easy-First POS Tagging and Dependency Parsing with Beam Search
Author: Ji Ma ; Jingbo Zhu ; Tong Xiao ; Nan Yang
Abstract: In this paper, we combine easy-first dependency parsing and POS tagging algorithms with beam search and structured perceptron. We propose a simple variant of “early-update” to ensure valid update in the training process. The proposed solution can also be applied to combine beam search and structured perceptron with other systems that exhibit spurious ambiguity. On CTB, we achieve 94.01% tagging accuracy and 86.33% unlabeled attachment score with a relatively small beam width. On PTB, we also achieve state-of-the-art performance. 1
4 0.12544009 372 acl-2013-Using CCG categories to improve Hindi dependency parsing
Author: Bharat Ram Ambati ; Tejaswini Deoskar ; Mark Steedman
Abstract: We show that informative lexical categories from a strongly lexicalised formalism such as Combinatory Categorial Grammar (CCG) can improve dependency parsing of Hindi, a free word order language. We first describe a novel way to obtain a CCG lexicon and treebank from an existing dependency treebank, using a CCG parser. We use the output of a supertagger trained on the CCGbank as a feature for a state-of-the-art Hindi dependency parser (Malt). Our results show that using CCG categories improves the accuracy of Malt on long distance dependencies, for which it is known to have weak rates of recovery.
5 0.11852951 44 acl-2013-An Empirical Examination of Challenges in Chinese Parsing
Author: Jonathan K. Kummerfeld ; Daniel Tse ; James R. Curran ; Dan Klein
Abstract: Aspects of Chinese syntax result in a distinctive mix of parsing challenges. However, the contribution of individual sources of error to overall difficulty is not well understood. We conduct a comprehensive automatic analysis of error types made by Chinese parsers, covering a broad range of error types for large sets of sentences, enabling the first empirical ranking of Chinese error types by their performance impact. We also investigate which error types are resolved by using gold part-of-speech tags, showing that improving Chinese tagging only addresses certain error types, leaving substantial outstanding challenges.
6 0.11411456 368 acl-2013-Universal Dependency Annotation for Multilingual Parsing
7 0.10684102 85 acl-2013-Combining Intra- and Multi-sentential Rhetorical Parsing for Document-level Discourse Analysis
8 0.098210976 26 acl-2013-A Transition-Based Dependency Parser Using a Dynamic Parsing Strategy
9 0.097103983 155 acl-2013-Fast and Accurate Shift-Reduce Constituent Parsing
10 0.094240539 112 acl-2013-Dependency Parser Adaptation with Subtrees from Auto-Parsed Target Domain Data
11 0.089207783 343 acl-2013-The Effect of Higher-Order Dependency Features in Discriminative Phrase-Structure Parsing
12 0.085446626 94 acl-2013-Coordination Structures in Dependency Treebanks
13 0.083705559 208 acl-2013-Joint Inference for Heterogeneous Dependency Parsing
14 0.079372145 80 acl-2013-Chinese Parsing Exploiting Characters
15 0.075783469 331 acl-2013-Stop-probability estimates computed on a large corpus improve Unsupervised Dependency Parsing
16 0.07128603 98 acl-2013-Cross-lingual Transfer of Semantic Role Labeling Models
17 0.066187479 19 acl-2013-A Shift-Reduce Parsing Algorithm for Phrase-based String-to-Dependency Translation
18 0.064355262 362 acl-2013-Turning on the Turbo: Fast Third-Order Non-Projective Turbo Parsers
19 0.063627087 13 acl-2013-A New Syntactic Metric for Evaluation of Machine Translation
20 0.063121922 28 acl-2013-A Unified Morpho-Syntactic Scheme of Stanford Dependencies
topicId topicWeight
[(0, 0.122), (1, -0.081), (2, -0.173), (3, 0.007), (4, -0.114), (5, -0.021), (6, 0.049), (7, -0.013), (8, 0.069), (9, -0.09), (10, 0.011), (11, 0.017), (12, -0.022), (13, 0.052), (14, 0.044), (15, 0.018), (16, -0.07), (17, -0.064), (18, -0.051), (19, -0.028), (20, -0.003), (21, -0.015), (22, -0.035), (23, -0.034), (24, 0.029), (25, 0.034), (26, -0.034), (27, -0.037), (28, 0.004), (29, -0.011), (30, 0.002), (31, -0.0), (32, -0.033), (33, 0.008), (34, -0.014), (35, -0.035), (36, -0.046), (37, -0.032), (38, 0.014), (39, -0.04), (40, 0.017), (41, -0.022), (42, -0.039), (43, -0.057), (44, 0.011), (45, -0.054), (46, -0.041), (47, 0.012), (48, 0.037), (49, 0.024)]
simIndex simValue paperId paperTitle
same-paper 1 0.95975322 335 acl-2013-Survey on parsing three dependency representations for English
Author: Angelina Ivanova ; Stephan Oepen ; Lilja vrelid
Abstract: In this paper we focus on practical issues of data representation for dependency parsing. We carry out an experimental comparison of (a) three syntactic dependency schemes; (b) three data-driven dependency parsers; and (c) the influence of two different approaches to lexical category disambiguation (aka tagging) prior to parsing. Comparing parsing accuracies in various setups, we study the interactions of these three aspects and analyze which configurations are easier to learn for a dependency parser.
2 0.8090983 208 acl-2013-Joint Inference for Heterogeneous Dependency Parsing
Author: Guangyou Zhou ; Jun Zhao
Abstract: This paper is concerned with the problem of heterogeneous dependency parsing. In this paper, we present a novel joint inference scheme, which is able to leverage the consensus information between heterogeneous treebanks in the parsing phase. Different from stacked learning methods (Nivre and McDonald, 2008; Martins et al., 2008), which process the dependency parsing in a pipelined way (e.g., a second level uses the first level outputs), in our method, multiple dependency parsing models are coordinated to exchange consensus information. We conduct experiments on Chinese Dependency Treebank (CDT) and Penn Chinese Treebank (CTB), experimental results show that joint infer- ence can bring significant improvements to all state-of-the-art dependency parsers.
3 0.80306822 94 acl-2013-Coordination Structures in Dependency Treebanks
Author: Martin Popel ; David Marecek ; Jan StÄłpanek ; Daniel Zeman ; ZdÄłnÄłk Zabokrtsky
Abstract: Paratactic syntactic structures are notoriously difficult to represent in dependency formalisms. This has painful consequences such as high frequency of parsing errors related to coordination. In other words, coordination is a pending problem in dependency analysis of natural languages. This paper tries to shed some light on this area by bringing a systematizing view of various formal means developed for encoding coordination structures. We introduce a novel taxonomy of such approaches and apply it to treebanks across a typologically diverse range of 26 languages. In addition, empirical observations on convertibility between selected styles of representations are shown too.
4 0.76688719 368 acl-2013-Universal Dependency Annotation for Multilingual Parsing
Author: Ryan McDonald ; Joakim Nivre ; Yvonne Quirmbach-Brundage ; Yoav Goldberg ; Dipanjan Das ; Kuzman Ganchev ; Keith Hall ; Slav Petrov ; Hao Zhang ; Oscar Tackstrom ; Claudia Bedini ; Nuria Bertomeu Castello ; Jungmee Lee
Abstract: We present a new collection of treebanks with homogeneous syntactic dependency annotation for six languages: German, English, Swedish, Spanish, French and Korean. To show the usefulness of such a resource, we present a case study of crosslingual transfer parsing with more reliable evaluation than has been possible before. This ‘universal’ treebank is made freely available in order to facilitate research on multilingual dependency parsing.1
5 0.74829322 358 acl-2013-Transition-based Dependency Parsing with Selectional Branching
Author: Jinho D. Choi ; Andrew McCallum
Abstract: We present a novel approach, called selectional branching, which uses confidence estimates to decide when to employ a beam, providing the accuracy of beam search at speeds close to a greedy transition-based dependency parsing approach. Selectional branching is guaranteed to perform a fewer number of transitions than beam search yet performs as accurately. We also present a new transition-based dependency parsing algorithm that gives a complexity of O(n) for projective parsing and an expected linear time speed for non-projective parsing. With the standard setup, our parser shows an unlabeled attachment score of 92.96% and a parsing speed of 9 milliseconds per sentence, which is faster and more accurate than the current state-of-the-art transitionbased parser that uses beam search.
6 0.74287349 155 acl-2013-Fast and Accurate Shift-Reduce Constituent Parsing
7 0.73949337 343 acl-2013-The Effect of Higher-Order Dependency Features in Discriminative Phrase-Structure Parsing
8 0.72689664 26 acl-2013-A Transition-Based Dependency Parser Using a Dynamic Parsing Strategy
9 0.68018597 204 acl-2013-Iterative Transformation of Annotation Guidelines for Constituency Parsing
10 0.67847264 372 acl-2013-Using CCG categories to improve Hindi dependency parsing
11 0.6783002 28 acl-2013-A Unified Morpho-Syntactic Scheme of Stanford Dependencies
12 0.67672807 331 acl-2013-Stop-probability estimates computed on a large corpus improve Unsupervised Dependency Parsing
13 0.67518783 132 acl-2013-Easy-First POS Tagging and Dependency Parsing with Beam Search
14 0.65052551 362 acl-2013-Turning on the Turbo: Fast Third-Order Non-Projective Turbo Parsers
15 0.61974341 288 acl-2013-Punctuation Prediction with Transition-based Parsing
16 0.59806943 112 acl-2013-Dependency Parser Adaptation with Subtrees from Auto-Parsed Target Domain Data
17 0.59007913 133 acl-2013-Efficient Implementation of Beam-Search Incremental Parsers
18 0.57603031 44 acl-2013-An Empirical Examination of Challenges in Chinese Parsing
19 0.54621458 357 acl-2013-Transfer Learning for Constituency-Based Grammars
20 0.54259425 19 acl-2013-A Shift-Reduce Parsing Algorithm for Phrase-based String-to-Dependency Translation
topicId topicWeight
[(0, 0.046), (6, 0.018), (11, 0.089), (14, 0.029), (24, 0.019), (26, 0.069), (28, 0.022), (32, 0.244), (35, 0.069), (42, 0.086), (48, 0.035), (70, 0.036), (88, 0.031), (90, 0.016), (95, 0.059)]
simIndex simValue paperId paperTitle
same-paper 1 0.8154304 335 acl-2013-Survey on parsing three dependency representations for English
Author: Angelina Ivanova ; Stephan Oepen ; Lilja vrelid
Abstract: In this paper we focus on practical issues of data representation for dependency parsing. We carry out an experimental comparison of (a) three syntactic dependency schemes; (b) three data-driven dependency parsers; and (c) the influence of two different approaches to lexical category disambiguation (aka tagging) prior to parsing. Comparing parsing accuracies in various setups, we study the interactions of these three aspects and analyze which configurations are easier to learn for a dependency parser.
2 0.59072316 358 acl-2013-Transition-based Dependency Parsing with Selectional Branching
Author: Jinho D. Choi ; Andrew McCallum
Abstract: We present a novel approach, called selectional branching, which uses confidence estimates to decide when to employ a beam, providing the accuracy of beam search at speeds close to a greedy transition-based dependency parsing approach. Selectional branching is guaranteed to perform a fewer number of transitions than beam search yet performs as accurately. We also present a new transition-based dependency parsing algorithm that gives a complexity of O(n) for projective parsing and an expected linear time speed for non-projective parsing. With the standard setup, our parser shows an unlabeled attachment score of 92.96% and a parsing speed of 9 milliseconds per sentence, which is faster and more accurate than the current state-of-the-art transitionbased parser that uses beam search.
3 0.58850873 132 acl-2013-Easy-First POS Tagging and Dependency Parsing with Beam Search
Author: Ji Ma ; Jingbo Zhu ; Tong Xiao ; Nan Yang
Abstract: In this paper, we combine easy-first dependency parsing and POS tagging algorithms with beam search and structured perceptron. We propose a simple variant of “early-update” to ensure valid update in the training process. The proposed solution can also be applied to combine beam search and structured perceptron with other systems that exhibit spurious ambiguity. On CTB, we achieve 94.01% tagging accuracy and 86.33% unlabeled attachment score with a relatively small beam width. On PTB, we also achieve state-of-the-art performance. 1
4 0.58183801 343 acl-2013-The Effect of Higher-Order Dependency Features in Discriminative Phrase-Structure Parsing
Author: Greg Coppola ; Mark Steedman
Abstract: Higher-order dependency features are known to improve dependency parser accuracy. We investigate the incorporation of such features into a cube decoding phrase-structure parser. We find considerable gains in accuracy on the range of standard metrics. What is especially interesting is that we find strong, statistically significant gains on dependency recovery on out-of-domain tests (Brown vs. WSJ). This suggests that higher-order dependency features are not simply overfitting the training material.
5 0.57104278 155 acl-2013-Fast and Accurate Shift-Reduce Constituent Parsing
Author: Muhua Zhu ; Yue Zhang ; Wenliang Chen ; Min Zhang ; Jingbo Zhu
Abstract: Shift-reduce dependency parsers give comparable accuracies to their chartbased counterparts, yet the best shiftreduce constituent parsers still lag behind the state-of-the-art. One important reason is the existence of unary nodes in phrase structure trees, which leads to different numbers of shift-reduce actions between different outputs for the same input. This turns out to have a large empirical impact on the framework of global training and beam search. We propose a simple yet effective extension to the shift-reduce process, which eliminates size differences between action sequences in beam-search. Our parser gives comparable accuracies to the state-of-the-art chart parsers. With linear run-time complexity, our parser is over an order of magnitude faster than the fastest chart parser.
6 0.57042199 80 acl-2013-Chinese Parsing Exploiting Characters
7 0.56929296 225 acl-2013-Learning to Order Natural Language Texts
8 0.56875318 137 acl-2013-Enlisting the Ghost: Modeling Empty Categories for Machine Translation
9 0.56659669 208 acl-2013-Joint Inference for Heterogeneous Dependency Parsing
10 0.56551111 276 acl-2013-Part-of-Speech Induction in Dependency Trees for Statistical Machine Translation
11 0.56470686 57 acl-2013-Arguments and Modifiers from the Learner's Perspective
12 0.56468725 70 acl-2013-Bilingually-Guided Monolingual Dependency Grammar Induction
13 0.56462902 38 acl-2013-Additive Neural Networks for Statistical Machine Translation
14 0.56302732 280 acl-2013-Plurality, Negation, and Quantification:Towards Comprehensive Quantifier Scope Disambiguation
15 0.56242508 133 acl-2013-Efficient Implementation of Beam-Search Incremental Parsers
16 0.56071532 98 acl-2013-Cross-lingual Transfer of Semantic Role Labeling Models
17 0.56011075 56 acl-2013-Argument Inference from Relevant Event Mentions in Chinese Argument Extraction
18 0.55925447 264 acl-2013-Online Relative Margin Maximization for Statistical Machine Translation
19 0.55921686 174 acl-2013-Graph Propagation for Paraphrasing Out-of-Vocabulary Words in Statistical Machine Translation
20 0.55854267 18 acl-2013-A Sentence Compression Based Framework to Query-Focused Multi-Document Summarization