| { | |
| "paper_id": "U16-1017", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T03:10:43.389919Z" | |
| }, | |
| "title": "An empirical study for Vietnamese dependency parsing", | |
| "authors": [ | |
| { | |
| "first": "Dat", | |
| "middle": [ | |
| "Quoc" | |
| ], | |
| "last": "Nguyen", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Macquarie University", | |
| "location": { | |
| "country": "Australia" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "Mark", | |
| "middle": [], | |
| "last": "Dras", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Macquarie University", | |
| "location": { | |
| "country": "Australia" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Mark", | |
| "middle": [], | |
| "last": "Johnson", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Macquarie University", | |
| "location": { | |
| "country": "Australia" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "This paper presents an empirical comparison of different dependency parsers for Vietnamese, which has some unusual characteristics such as copula drop and verb serialization. Experimental results show that the neural network-based parsers perform significantly better than the traditional parsers. We report the highest parsing scores published to date for Vietnamese with the labeled attachment score (LAS) at 73.53% and the unlabeled attachment score (UAS) at 80.66%.", | |
| "pdf_parse": { | |
| "paper_id": "U16-1017", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "This paper presents an empirical comparison of different dependency parsers for Vietnamese, which has some unusual characteristics such as copula drop and verb serialization. Experimental results show that the neural network-based parsers perform significantly better than the traditional parsers. We report the highest parsing scores published to date for Vietnamese with the labeled attachment score (LAS) at 73.53% and the unlabeled attachment score (UAS) at 80.66%.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "Dependency parsing has become a key research topic in natural language processing in the last decade, boosted by the success of the CoNLL 2006 and 2007 shared tasks on multilingual dependency parsing (Buchholz and Marsi, 2006; Nivre et al., 2007a) . McDonald and Nivre (2011) identify two types of approaches for dependency parsing: graph-based approaches (McDonald et al., 2005) and transition-based approaches (Nivre et al., 2007b) . Most traditional graph-or transitionbased dependency parsers (McDonald et al., 2005; Nivre et al., 2007b; Bohnet, 2010; Zhang and Nivre, 2011; Martins et al., 2013; Choi and Mc-Callum, 2013) manually define a set of core and combined features associated with one-hot representations.", | |
| "cite_spans": [ | |
| { | |
| "start": 200, | |
| "end": 226, | |
| "text": "(Buchholz and Marsi, 2006;", | |
| "ref_id": "BIBREF2" | |
| }, | |
| { | |
| "start": 227, | |
| "end": 247, | |
| "text": "Nivre et al., 2007a)", | |
| "ref_id": "BIBREF22" | |
| }, | |
| { | |
| "start": 250, | |
| "end": 275, | |
| "text": "McDonald and Nivre (2011)", | |
| "ref_id": "BIBREF14" | |
| }, | |
| { | |
| "start": 356, | |
| "end": 379, | |
| "text": "(McDonald et al., 2005)", | |
| "ref_id": "BIBREF15" | |
| }, | |
| { | |
| "start": 412, | |
| "end": 433, | |
| "text": "(Nivre et al., 2007b)", | |
| "ref_id": "BIBREF23" | |
| }, | |
| { | |
| "start": 497, | |
| "end": 520, | |
| "text": "(McDonald et al., 2005;", | |
| "ref_id": "BIBREF15" | |
| }, | |
| { | |
| "start": 521, | |
| "end": 541, | |
| "text": "Nivre et al., 2007b;", | |
| "ref_id": "BIBREF23" | |
| }, | |
| { | |
| "start": 542, | |
| "end": 555, | |
| "text": "Bohnet, 2010;", | |
| "ref_id": "BIBREF1" | |
| }, | |
| { | |
| "start": 556, | |
| "end": 578, | |
| "text": "Zhang and Nivre, 2011;", | |
| "ref_id": "BIBREF33" | |
| }, | |
| { | |
| "start": 579, | |
| "end": 600, | |
| "text": "Martins et al., 2013;", | |
| "ref_id": "BIBREF13" | |
| }, | |
| { | |
| "start": 601, | |
| "end": 626, | |
| "text": "Choi and Mc-Callum, 2013)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Recent work shows that neural network-based parsers obtain the state-of-the-art parsing results across many languages. Chen and Manning (2014) , Weiss et al. (2015) , Pei et al. (2015) , and Andor et al. (2016) represent the core features with dense vector embeddings and then feed them as inputs to neural network-based classifiers, while Dyer et al. (2015) , Kiperwasser and Goldberg (2016a) , and Kiperwasser and Goldberg (2016b) propose novel neural network architectures to solve the feature-engineering problem.", | |
| "cite_spans": [ | |
| { | |
| "start": 119, | |
| "end": 142, | |
| "text": "Chen and Manning (2014)", | |
| "ref_id": "BIBREF3" | |
| }, | |
| { | |
| "start": 145, | |
| "end": 164, | |
| "text": "Weiss et al. (2015)", | |
| "ref_id": "BIBREF31" | |
| }, | |
| { | |
| "start": 167, | |
| "end": 184, | |
| "text": "Pei et al. (2015)", | |
| "ref_id": "BIBREF24" | |
| }, | |
| { | |
| "start": 191, | |
| "end": 210, | |
| "text": "Andor et al. (2016)", | |
| "ref_id": "BIBREF0" | |
| }, | |
| { | |
| "start": 340, | |
| "end": 358, | |
| "text": "Dyer et al. (2015)", | |
| "ref_id": "BIBREF7" | |
| }, | |
| { | |
| "start": 361, | |
| "end": 393, | |
| "text": "Kiperwasser and Goldberg (2016a)", | |
| "ref_id": "BIBREF10" | |
| }, | |
| { | |
| "start": 400, | |
| "end": 432, | |
| "text": "Kiperwasser and Goldberg (2016b)", | |
| "ref_id": "BIBREF11" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Dependency parsing for Vietnamese has not been actively explored. One main reason is because there is no manually labeled dependency treebank available. Thi et al. (2013) and Nguyen et al. (2014b) propose constituent-to-dependency conversion approaches to automatically translate the manually built constituent treebank for Vietnamese (Nguyen et al., 2009) to dependency treebanks. The converted dependency treebanks are then used in later works on Vietnamese dependency parsing, including Vu-Manh et al. 2015, and . All of the previous research works use either the MSTparser (McDonald et al., 2005) or the Maltparser (Nivre et al., 2007b) for their parsing experiments. Among them, Nguyen et al. (2014b) report the highest results with LAS at 71.66% and UAS at 79.08% obtained by MSTparser. However, MSTparser and Maltparser are no longer considered state-of-the-art parsers.", | |
| "cite_spans": [ | |
| { | |
| "start": 153, | |
| "end": 170, | |
| "text": "Thi et al. (2013)", | |
| "ref_id": "BIBREF28" | |
| }, | |
| { | |
| "start": 175, | |
| "end": 196, | |
| "text": "Nguyen et al. (2014b)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 324, | |
| "end": 356, | |
| "text": "Vietnamese (Nguyen et al., 2009)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 577, | |
| "end": 600, | |
| "text": "(McDonald et al., 2005)", | |
| "ref_id": "BIBREF15" | |
| }, | |
| { | |
| "start": 619, | |
| "end": 640, | |
| "text": "(Nivre et al., 2007b)", | |
| "ref_id": "BIBREF23" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "In this paper, we present an empirical study of Vietnamese dependency parsing. We make comparisons between neural network-based parsers and traditional parsers, and also between graphbased parsers and transition-based parsers. We show that the neural network-based parsers obtain significantly higher scores than the traditional parsers. Specifically, we report the highest up-todate scores for Vietnamese with LAS at 73.53% and UAS at 80.66%. We also examine potential problems specific to parsing Vietnamese, and point out potential solutions for improving the parsing performance.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Dataset: There are two Vietnamese dependency treebanks which are automatically converted from the manually-annotated Vietnamese constituent Table 1 : VnDT statistics by most frequent dependency and part-of-speech (POS) labels, and sentence length (i.e. number of words). \"Rate\" denotes the percentage occurrence in VnDT. Dependency labels:", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 140, | |
| "end": 147, | |
| "text": "Table 1", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Experimental setup", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "adv (adverbial), amod (adjectival modifier), conj (conjunct), coord (coordinating conjunction), dep (unspecified dependency), det (determiner), dob (direct object), loc (location), nmod (noun modifier), pob (object of a preposi- tion), punct (punctuation), sub (subject), tmp (tem- poral), vmod (verb modifier). POS tags: A (Adjec- tive), C (Conjunction), E (Preposition), M (Quan- tity), N (Noun), Nc (Classifier noun), Np (Proper noun), P (Pronoun), R (Adjunct), V (Verb).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Experimental setup", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "treebank (Nguyen et al., 2009) , using conversion approaches proposed by Thi et al. (2013) and Nguyen et al. (2014b) . In Thi et al. (2013)'s conversion approach, it is not clear how the dependency labels are inferred; also, it ignores grammatical information encoded in grammatical function tags. In addition, Thi et al. (2013)'s approach is unable to handle cases of coordination and empty category mappings, which frequently appear in the Vietnamese constituent treebank. Nguyen et al. (2014b) later proposed a new conversion approach to handle those cases, with a better use of existing information in the Vietnamese constituent treebank. So we conduct experiments using VnDT, the high quality Vietnamese dependency treebank produced by Nguyen et al. (2014b). The VnDT treebank consists of 10,200 sentences (about 219K words). Table 1 gives some basic statistics of VnDT. We use the last 1020 sentences of VnDT for testing while the remaining sentences are used for training, resulting in an out-of-vocabulary rate of 3.3%.", | |
| "cite_spans": [ | |
| { | |
| "start": 9, | |
| "end": 30, | |
| "text": "(Nguyen et al., 2009)", | |
| "ref_id": "BIBREF17" | |
| }, | |
| { | |
| "start": 73, | |
| "end": 90, | |
| "text": "Thi et al. (2013)", | |
| "ref_id": "BIBREF28" | |
| }, | |
| { | |
| "start": 95, | |
| "end": 116, | |
| "text": "Nguyen et al. (2014b)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 831, | |
| "end": 838, | |
| "text": "Table 1", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Experimental setup", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "We experiment with four parsers: the graph-based parsers BISTbmstparser 1 (BistG) and MSTparser 2 (MST), and the transition-based parsers BIST-barchybrid 3 (BistT) and Maltparser 4 (Malt). The state-ofthe-art BistG and BistT parsers (Kiperwasser and Goldberg, 2016b ) employ a bidirectional LSTM RNN architecture (Schuster and Paliwal, 1997; Hochreiter and Schmidhuber, 1997) to automatically learn the feature representation. In contrast, the traditional parsers MST (McDonald et al., 2005) and Malt (Nivre et al., 2007b ) use a set of predefined features. For training these parsers, we used the default settings.", | |
| "cite_spans": [ | |
| { | |
| "start": 233, | |
| "end": 265, | |
| "text": "(Kiperwasser and Goldberg, 2016b", | |
| "ref_id": "BIBREF11" | |
| }, | |
| { | |
| "start": 313, | |
| "end": 341, | |
| "text": "(Schuster and Paliwal, 1997;", | |
| "ref_id": "BIBREF26" | |
| }, | |
| { | |
| "start": 342, | |
| "end": 375, | |
| "text": "Hochreiter and Schmidhuber, 1997)", | |
| "ref_id": "BIBREF9" | |
| }, | |
| { | |
| "start": 464, | |
| "end": 491, | |
| "text": "MST (McDonald et al., 2005)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 501, | |
| "end": 521, | |
| "text": "(Nivre et al., 2007b", | |
| "ref_id": "BIBREF23" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Dependency parsers:", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The metrics are the labeled attachment score (LAS), unlabeled attachment score (UAS) and label accuracy score (LS). LAS is the percentage of words which are correctly assigned both dependency arc and label while UAS is the percentage of words for which the dependency arc is assigned correctly, and LS is the percentage of words for which the dependency label is assigned correctly. Table 2 compares the parsing results obtained by the four parsers. The first four rows report the scores with gold part-of-speech (POS) tags while the last four rows present the scores with automatically predicted POS tags. 5 As expected the neural network-based parsers BistG and BistT perform significantly better than the traditional parsers MST and Malt. 6 Specifically, we find 2 + % absolute improvements in LAS and UAS scores in both graph-and transitionbased types. In most cases, there are no significant differences between the LAS and UAS scores of BistG and BistT, except LAS scored on gold POS Next, Section 3.2 gives a detailed accuracy analysis on gold POS tags without punctuation, and Section 3.3 discusses the source of some errors and possible improvements.", | |
| "cite_spans": [ | |
| { | |
| "start": 607, | |
| "end": 608, | |
| "text": "5", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 383, | |
| "end": 390, | |
| "text": "Table 2", | |
| "ref_id": "TABREF2" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Evaluation metrics:", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Sentence length: Figures 1 and 2 detail LAS and UAS scores by sentence length in bins of length 10. It is not surprising that all parsers produce better results for shorter sentences. For sentences shorter than 10 words, all LAS and UAS scores are around 80% and 85%, respectively. However, the scores drop by 10 + % for sentences longer than 50 words. The Malt parser obtains the lowest LAS and UAS scores across all sentence bins. BistG obtains the highest scores for sentences shorter than 20 words while BistT obtains highest scores for sentences longer than 40 words. BistG, BistT and MST perform similarly on 30to-40-word sentences. For shorter sentences from 20 to 30 words, BistG and BistT produce similar results but higher than obtained by MST. Figures 3 and 4 show F 1 scores in terms of the distance from each dependent word to its head. Similar to English (Choi et al., 2015), we find better predictions for the left dependencies than for the right dependencies. Unlike in English where the lower scores are associated with longer distances, we find a different pattern when predicting the left dependencies 7 The differences are statistically significant at p < 0.02. in Vietnamese. In a distance bin of 3, 4 and 5 words with respect to the left dependencies, three over four parsers including BistG, BistT and Malt generally obtain better predictions for longer distances. Compared to English, Vietnamese is headinitial, so finding a difference with respect to left dependencies is not completely unexpected. In addition, for this distance bin, the transition-based parser does better than the graph-based parser in both neural net-based and traditional categories (i.e. BistT > BistG and Malt > MST). In both those categories, however, the graph-based parser does better than the transition-based parser for 5word-longer distances (i.e. BistG > BistT and MST > Malt), while they produce similar results on dependency distances of 1 or 2 words. Because the dependency distance of 3, 4 or 5 occurs quite frequently in long sentences, so the results here are consistent with the results shown in Figures 1 and 2 where BistT obtains the highest scores for long sentences. Table 3 presents LAS scores for the most frequent dependency labels. The labels with higher than 90% accuracy are adv, det and pob in which surprisingly MST obtains the best results on all these labels and even on both conj and dob labels. BistT obtains the best scores on the two most frequent labels nmod and vmod, and also on the loc label. BistG performs best on the remaining labels. Biggest ranges (> 10%) of obtained scores across parsers associate to labels coord, dep, sub and tmp. Table 3 also shows that the label with the lowest LAS scores (< 50%) across all parsers is dep which is a very general label. Those with LAS scores ranging from 50% to about 60% are coord, loc, tmp and vmod in which coord, loc and tmp are among the least frequent labels, while vmod is the second most frequent label. ", | |
| "cite_spans": [ | |
| { | |
| "start": 1121, | |
| "end": 1122, | |
| "text": "7", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 17, | |
| "end": 32, | |
| "text": "Figures 1 and 2", | |
| "ref_id": "FIGREF0" | |
| }, | |
| { | |
| "start": 755, | |
| "end": 770, | |
| "text": "Figures 3 and 4", | |
| "ref_id": "FIGREF2" | |
| }, | |
| { | |
| "start": 2109, | |
| "end": 2124, | |
| "text": "Figures 1 and 2", | |
| "ref_id": "FIGREF0" | |
| }, | |
| { | |
| "start": 2184, | |
| "end": 2191, | |
| "text": "Table 3", | |
| "ref_id": "TABREF4" | |
| }, | |
| { | |
| "start": 2675, | |
| "end": 2682, | |
| "text": "Table 3", | |
| "ref_id": "TABREF4" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Accuracy analysis", | |
| "sec_num": "3.2" | |
| }, | |
| { | |
| "text": "In Table 4 we analyze the results by the POS tag of the dependent. BistG achieves the highest results on the two most frequent POS tags N and V and also on C and E. BistT achieves the highest scores on the remaining POS tags except M for which MST produces the highest score.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 3, | |
| "end": 10, | |
| "text": "Table 4", | |
| "ref_id": "TABREF5" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "POS tags:", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Linguistic aspects: One surprising characteristic of the results is the poor performance of verbrelated dependencies: vmod accuracy is low, as are scores associated with the second most frequent POS tag V (Verb). For the latter, we find significantly lower scores for verbs in Vietnamese (around 65% as shown in Table 4 ) against scores for verbs (about 80 + %) obtained by MST and Malt parsers on 13 other languages reported in McDonald and Nivre (2011) , and also much worse performance in terms of rank relative to other POS. This may be related to syntactic characteristics of Vietnamese (Thompson, 1987) . First, Vietnamese is described as a copula-drop language. Consider C\u00f4 H\u00e0 c\u00f3 nh\u00e0 \u0111\u1eb9p \"Miss H\u00e0 has a beautiful house\", where the attributive adjective \u0111\u1eb9p Figure 5 : An example of a Vietnamese sentence with copula-drop. The first parsed tree is the gold one while the second parsed tree is the same output produced by all parsers. This sentence is translated to English as \"the land #1 is revived #2 .\", in which the copula \"is\" is dropped in Vietnamese. The subscripts in the English-translated sentence refer to alignments with the word indexes in the Vietnamese sentence.", | |
| "cite_spans": [ | |
| { | |
| "start": 429, | |
| "end": 454, | |
| "text": "McDonald and Nivre (2011)", | |
| "ref_id": "BIBREF14" | |
| }, | |
| { | |
| "start": 592, | |
| "end": 608, | |
| "text": "(Thompson, 1987)", | |
| "ref_id": "BIBREF29" | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 312, | |
| "end": 319, | |
| "text": "Table 4", | |
| "ref_id": "TABREF5" | |
| }, | |
| { | |
| "start": 764, | |
| "end": 772, | |
| "text": "Figure 5", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Discussions", | |
| "sec_num": "3.3" | |
| }, | |
| { | |
| "text": "\"beautiful\" postmodifies the noun nh\u00e0 \"house\". Adjectives can also be predicative, where they are conventionally labelled V (Verb), and a copula is absent: with Vietnamese's SVO word order, this is also nh\u00e0 \u0111\u1eb9p \"the house is beautiful.\" Figure 5 presents an example from the treebank: all four parsers produce the incorrect structure, which is what would be expected for the attributive adjectival use in an NP. This construction is quite common in Vietnamese.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 237, | |
| "end": 245, | |
| "text": "Figure 5", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Discussions", | |
| "sec_num": "3.3" | |
| }, | |
| { | |
| "text": "Second, Vietnamese permits verb serialization, as in Figure 6 : gi\u1eadt_m\u00ecnh \"accuses\" should be a vmod dependent of c\u00f3 \"excuses\"; such a construction is analogous to the more familiar nmod in other languages. Verb dependencies in Vietnamese might thus be less predictable than in other languages, with a more varied distribution of dependents.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 53, | |
| "end": 61, | |
| "text": "Figure 6", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Discussions", | |
| "sec_num": "3.3" | |
| }, | |
| { | |
| "text": "Other aspects: Generally, one reason for low overall scores on Vietnamese dependency parsing when compared to the scores obtained on the other languages (McDonald and Nivre, 2011) is probably because of the complex structures of many long sentences in the VnDT treebank (e.g. 45% of the sentences in VnDT consist of more than 20 words). So we can only obtain 60% and 50% for left and right dependency distances larger than 5 as shown in Figure 4 , respectively, while for English both left and right dependencies with distances larger than 5 have greater than 70% accuracy (Choi et al., 2015). Figure 6 : An example of a Vietnamese sentence with verb serialization (and pronoun-dropping), parsed by BistT. The gold parsed tree is when the indexed-3 word is attached to the indexed-1 word by vmod, instead of the indexed-2 word. For this sentence, BistG, MST and Malt attach the indexed-3 word to be dependent to the indexed-2 word by the label nmod. This sentence is translated to English as \"He who excuses #1 himself #2 , accuses #3 himself .\" One simple approach to improve parsing performance for Vietnamese is to separately use the graph-based parser BistG for short sentences and the transition-based parser BistT for longer sentences. Another approach is to use system combination (Nivre and McDonald, 2008; Zhang and Clark, 2008) , e.g. building ensemble systems (Sagae and Tsujii, 2007; Surdeanu and Manning, 2010; Haffari et al., 2011) . Table 5 presents an upper bound of oracle ensemble performance, using the DEPENDABLE toolkit (Choi et al., 2015) . DEPENDABLE assumes that either the best tree or the best arc can be determined by an oracle.", | |
| "cite_spans": [ | |
| { | |
| "start": 1288, | |
| "end": 1314, | |
| "text": "(Nivre and McDonald, 2008;", | |
| "ref_id": "BIBREF21" | |
| }, | |
| { | |
| "start": 1315, | |
| "end": 1337, | |
| "text": "Zhang and Clark, 2008)", | |
| "ref_id": "BIBREF32" | |
| }, | |
| { | |
| "start": 1371, | |
| "end": 1395, | |
| "text": "(Sagae and Tsujii, 2007;", | |
| "ref_id": "BIBREF25" | |
| }, | |
| { | |
| "start": 1396, | |
| "end": 1423, | |
| "text": "Surdeanu and Manning, 2010;", | |
| "ref_id": "BIBREF27" | |
| }, | |
| { | |
| "start": 1424, | |
| "end": 1445, | |
| "text": "Haffari et al., 2011)", | |
| "ref_id": "BIBREF8" | |
| }, | |
| { | |
| "start": 1541, | |
| "end": 1560, | |
| "text": "(Choi et al., 2015)", | |
| "ref_id": "BIBREF6" | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 437, | |
| "end": 445, | |
| "text": "Figure 4", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 594, | |
| "end": 602, | |
| "text": "Figure 6", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 1448, | |
| "end": 1455, | |
| "text": "Table 5", | |
| "ref_id": "TABREF7" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Discussions", | |
| "sec_num": "3.3" | |
| }, | |
| { | |
| "text": "We have presented an empirical comparison for Vietnamese dependency parsing. Experimental results on the Vietnamese dependency treebank VnDT (Nguyen et al., 2014b) show that the neural network-based parsers (Kiperwasser and Goldberg, 2016b) obtain significantly higher scores than the traditional parsers (McDonald et al., 2005; Nivre et al., 2007b) . More specifically, in each graph-or transition-based type, we find a 2% absolute improvement of the neural network-based parser over the traditional one.", | |
| "cite_spans": [ | |
| { | |
| "start": 136, | |
| "end": 163, | |
| "text": "VnDT (Nguyen et al., 2014b)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 207, | |
| "end": 240, | |
| "text": "(Kiperwasser and Goldberg, 2016b)", | |
| "ref_id": "BIBREF11" | |
| }, | |
| { | |
| "start": 305, | |
| "end": 328, | |
| "text": "(McDonald et al., 2005;", | |
| "ref_id": "BIBREF15" | |
| }, | |
| { | |
| "start": 329, | |
| "end": 349, | |
| "text": "Nivre et al., 2007b)", | |
| "ref_id": "BIBREF23" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusions", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "We report the highest performance up to date for Vietnamese dependency parsing with LAS at 73.53% and UAS at 80.66%.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusions", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "http://www.maltparser.org5 We adapted the RDRPOSTagger toolkit(Nguyen et al., 2014a; Nguyen et al., 2016) to automatically assign POS tags to words in the test set with an accuracy of 94.58%.6 Using McNemar's test, the differences are statistically significant at p < 0.001.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [ | |
| { | |
| "text": "The first author is supported by an International Postgraduate Research Scholarship and a NICTA NRPA Top-Up Scholarship.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Acknowledgments", | |
| "sec_num": null | |
| } | |
| ], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "Globally Normalized Transition-Based Neural Networks", | |
| "authors": [ | |
| { | |
| "first": "Daniel", | |
| "middle": [], | |
| "last": "Andor", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Chris", | |
| "middle": [], | |
| "last": "Alberti", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "David", | |
| "middle": [], | |
| "last": "Weiss", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Aliaksei", | |
| "middle": [], | |
| "last": "Severyn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Alessandro", | |
| "middle": [], | |
| "last": "Presta", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Kuzman", | |
| "middle": [], | |
| "last": "Ganchev", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Slav", | |
| "middle": [], | |
| "last": "Petrov", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Michael", | |
| "middle": [], | |
| "last": "Collins", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "2442--2452", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Daniel Andor, Chris Alberti, David Weiss, Aliaksei Severyn, Alessandro Presta, Kuzman Ganchev, Slav Petrov, and Michael Collins. 2016. Globally Normalized Transition-Based Neural Networks. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pages 2442-2452.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Very high accuracy and fast dependency parsing is not a contradiction", | |
| "authors": [ | |
| { | |
| "first": "Bernd", | |
| "middle": [], | |
| "last": "Bohnet", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2010, | |
| "venue": "Proceedings of the 23rd International Conference on Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "89--97", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Bernd Bohnet. 2010. Very high accuracy and fast dependency parsing is not a contradiction. In Proceedings of the 23rd International Conference on Computational Linguistics, pages 89-97.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "CoNLL-X shared task on multilingual dependency parsing", | |
| "authors": [ | |
| { | |
| "first": "Sabine", | |
| "middle": [], | |
| "last": "Buchholz", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Erwin", | |
| "middle": [], | |
| "last": "Marsi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2006, | |
| "venue": "Proceedings of the Tenth Conference on Computational Natural Language Learning", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "149--164", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Sabine Buchholz and Erwin Marsi. 2006. CoNLL-X shared task on multilingual dependency parsing. In Proceedings of the Tenth Conference on Computational Natural Language Learning, CoNLL-X, pages 149-164.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "A fast and accurate dependency parser using neural networks", | |
| "authors": [ | |
| { | |
| "first": "Danqi", | |
| "middle": [], | |
| "last": "Chen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Christopher", | |
| "middle": [], | |
| "last": "Manning", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "Proceedings of the", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Danqi Chen and Christopher Manning. 2014. A fast and accurate dependency parser using neural networks. In Proceedings of the 2014", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Conference on Empirical Methods in Natural Language Processing", | |
| "authors": [], | |
| "year": null, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "740--750", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Conference on Empirical Methods in Natural Language Processing, pages 740-750.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Transition-based dependency parsing with selectional branching", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Jinho", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Andrew", | |
| "middle": [], | |
| "last": "Choi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Mccallum", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "1052--1062", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Jinho D. Choi and Andrew McCallum. 2013. Transition-based dependency parsing with selectional branching. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, pages 1052-1062.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "It Depends: Dependency Parser Comparison Using A Web-based Evaluation Tool", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Jinho", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Joel", | |
| "middle": [], | |
| "last": "Choi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Amanda", | |
| "middle": [], | |
| "last": "Tetreault", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Stent", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "387--396", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Jinho D. Choi, Joel Tetreault, and Amanda Stent. 2015. It Depends: Dependency Parser Comparison Using A Web-based Evaluation Tool. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 387-396.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "Transition-Based Dependency Parsing with Stack Long Short-Term Memory", | |
| "authors": [ | |
| { | |
| "first": "Chris", | |
| "middle": [], | |
| "last": "Dyer", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Miguel", | |
| "middle": [], | |
| "last": "Ballesteros", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Wang", | |
| "middle": [], | |
| "last": "Ling", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Austin", | |
| "middle": [], | |
| "last": "Matthews", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Noah", | |
| "middle": [ | |
| "A" | |
| ], | |
| "last": "Smith", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "334--343", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Chris Dyer, Miguel Ballesteros, Wang Ling, Austin Matthews, and Noah A. Smith. 2015. Transition- Based Dependency Parsing with Stack Long Short-Term Memory. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 334-343.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "An ensemble model that combines syntactic and semantic clustering for discriminative dependency parsing", | |
| "authors": [ | |
| { | |
| "first": "Gholamreza", | |
| "middle": [], | |
| "last": "Haffari", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Marzieh", | |
| "middle": [], | |
| "last": "Razavi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Anoop", | |
| "middle": [], | |
| "last": "Sarkar", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "710--714", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Gholamreza Haffari, Marzieh Razavi, and Anoop Sarkar. 2011. An ensemble model that combines syntactic and semantic clustering for discriminative dependency parsing. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, pages 710-714.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "Long short-term memory", | |
| "authors": [ | |
| { | |
| "first": "Sepp", | |
| "middle": [], | |
| "last": "Hochreiter", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J\u00fcrgen", | |
| "middle": [], | |
| "last": "Schmidhuber", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1997, | |
| "venue": "Neural Computation", | |
| "volume": "9", | |
| "issue": "8", | |
| "pages": "1735--1780", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Sepp Hochreiter and J\u00fcrgen Schmidhuber. 1997. Long short-term memory. Neural Computation, 9(8):1735-1780.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "Easy-first dependency parsing with hierarchical tree lstms", | |
| "authors": [ | |
| { | |
| "first": "Eliyahu", | |
| "middle": [], | |
| "last": "Kiperwasser", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yoav", | |
| "middle": [], | |
| "last": "Goldberg", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Transactions of the Association for Computational Linguistics", | |
| "volume": "4", | |
| "issue": "", | |
| "pages": "445--461", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Eliyahu Kiperwasser and Yoav Goldberg. 2016a. Easy-first dependency parsing with hierarchical tree lstms. Transactions of the Association for Computational Linguistics, 4:445-461.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "Simple and Accurate Dependency Parsing Using Bidirectional LSTM Feature Representations", | |
| "authors": [ | |
| { | |
| "first": "Eliyahu", | |
| "middle": [], | |
| "last": "Kiperwasser", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yoav", | |
| "middle": [], | |
| "last": "Goldberg", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Transactions of the Association for Computational Linguistics", | |
| "volume": "4", | |
| "issue": "", | |
| "pages": "313--327", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Eliyahu Kiperwasser and Yoav Goldberg. 2016b. Simple and Accurate Dependency Parsing Using Bidirectional LSTM Feature Representations. Transactions of the Association for Computational Linguistics, 4:313-327.", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "Fast dependency parsing using distributed word representations", | |
| "authors": [ | |
| { | |
| "first": "Phuong", | |
| "middle": [], | |
| "last": "Le-Hong", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Thi-Minh-Huyen", | |
| "middle": [], | |
| "last": "Nguyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Thi-Luong", | |
| "middle": [], | |
| "last": "Nguyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "My-Linh", | |
| "middle": [], | |
| "last": "Ha", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the PAKDD 2015 Workshops", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "261--272", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Phuong Le-Hong, Thi-Minh-Huyen Nguyen, Thi- Luong Nguyen, and My-Linh Ha. 2015. Fast dependency parsing using distributed word representations. In Proceedings of the PAKDD 2015 Workshops, pages 261-272.", | |
| "links": null | |
| }, | |
| "BIBREF13": { | |
| "ref_id": "b13", | |
| "title": "Turning on the Turbo: Fast Third-Order Non-Projective Turbo Parsers", | |
| "authors": [ | |
| { | |
| "first": "Andre", | |
| "middle": [], | |
| "last": "Martins", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Miguel", | |
| "middle": [], | |
| "last": "Almeida", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Noah", | |
| "middle": [ | |
| "A" | |
| ], | |
| "last": "Smith", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics", | |
| "volume": "2", | |
| "issue": "", | |
| "pages": "617--622", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Andre Martins, Miguel Almeida, and Noah A. Smith. 2013. Turning on the Turbo: Fast Third-Order Non-Projective Turbo Parsers. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 617-622.", | |
| "links": null | |
| }, | |
| "BIBREF14": { | |
| "ref_id": "b14", | |
| "title": "Analyzing and integrating dependency parsers", | |
| "authors": [ | |
| { | |
| "first": "Ryan", | |
| "middle": [], | |
| "last": "Mcdonald", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Joakim", | |
| "middle": [], | |
| "last": "Nivre", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "Computational Linguistics", | |
| "volume": "37", | |
| "issue": "1", | |
| "pages": "197--230", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Ryan McDonald and Joakim Nivre. 2011. Analyzing and integrating dependency parsers. Computational Linguistics, 37(1):197-230.", | |
| "links": null | |
| }, | |
| "BIBREF15": { | |
| "ref_id": "b15", | |
| "title": "Online large-margin training of dependency parsers", | |
| "authors": [ | |
| { | |
| "first": "Ryan", | |
| "middle": [], | |
| "last": "Mcdonald", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Koby", | |
| "middle": [], | |
| "last": "Crammer", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Fernando", | |
| "middle": [], | |
| "last": "Pereira", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2005, | |
| "venue": "Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "91--98", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Ryan McDonald, Koby Crammer, and Fernando Pereira. 2005. Online large-margin training of dependency parsers. In Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics, pages 91-98.", | |
| "links": null | |
| }, | |
| "BIBREF16": { | |
| "ref_id": "b16", | |
| "title": "Error Analysis for Vietnamese Dependency Parsing", | |
| "authors": [ | |
| { | |
| "first": "Kiet", | |
| "middle": [], | |
| "last": "Van Nguyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ngan", | |
| "middle": [], | |
| "last": "Luu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "-Thuy", | |
| "middle": [], | |
| "last": "Nguyen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the 2015 Seventh International Conference on Knowledge and Systems Engineering", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "79--84", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Kiet Van Nguyen and Ngan Luu-Thuy Nguyen. 2015. Error Analysis for Vietnamese Dependency Parsing. In Proceedings of the 2015 Seventh International Conference on Knowledge and Systems Engineering, pages 79-84.", | |
| "links": null | |
| }, | |
| "BIBREF17": { | |
| "ref_id": "b17", | |
| "title": "Building a Large Syntactically-Annotated Corpus of Vietnamese", | |
| "authors": [ | |
| { | |
| "first": "Phuong", | |
| "middle": [], | |
| "last": "Thai Nguyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Xuan", | |
| "middle": [ | |
| "Luong" | |
| ], | |
| "last": "Vu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Thi", | |
| "middle": [], | |
| "last": "Minh Huyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Nguyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Hong", | |
| "middle": [ | |
| "Phuong" | |
| ], | |
| "last": "Van Hiep Nguyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Le", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2009, | |
| "venue": "Proceedings of the Third Linguistic Annotation Workshop", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "182--185", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Phuong Thai Nguyen, Xuan Luong Vu, Thi Minh Huyen Nguyen, Van Hiep Nguyen, and Hong Phuong Le. 2009. Building a Large Syntactically-Annotated Corpus of Vietnamese. In Proceedings of the Third Linguistic Annotation Workshop, pages 182-185.", | |
| "links": null | |
| }, | |
| "BIBREF18": { | |
| "ref_id": "b18", | |
| "title": "RDRPOSTagger: A Ripple Down Rules-based Part-Of-Speech Tagger", | |
| "authors": [], | |
| "year": 2014, | |
| "venue": "Proceedings of the Demonstrations at the 14th Conference of the European Chapter of the Association for Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "17--20", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Dang Duc Pham, and Son Bao Pham. 2014a. RDRPOSTagger: A Ripple Down Rules-based Part-Of-Speech Tagger. In Proceedings of the Demonstrations at the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 17-20.", | |
| "links": null | |
| }, | |
| "BIBREF19": { | |
| "ref_id": "b19", | |
| "title": "From Treebank Conversion to Automatic Dependency Parsing for Vietnamese", | |
| "authors": [], | |
| "year": 2014, | |
| "venue": "Proceedings of 19th International Conference on Application of Natural Language to Information Systems", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "196--207", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Son Bao Pham, Phuong-Thai Nguyen, and Minh Le Nguyen. 2014b. From Treebank Conversion to Automatic Dependency Parsing for Vietnamese. In Proceedings of 19th International Conference on Application of Natural Language to Information Systems, pages 196-207.", | |
| "links": null | |
| }, | |
| "BIBREF20": { | |
| "ref_id": "b20", | |
| "title": "A robust transformationbased learning approach using ripple down rules for part-of-speech tagging", | |
| "authors": [], | |
| "year": 2016, | |
| "venue": "AI Communications", | |
| "volume": "29", | |
| "issue": "3", | |
| "pages": "409--422", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Dat Quoc Nguyen, Dai Quoc Nguyen, Dang Duc Pham, and Son Bao Pham. 2016. A robust transformation- based learning approach using ripple down rules for part-of-speech tagging. AI Communications, 29(3):409-422.", | |
| "links": null | |
| }, | |
| "BIBREF21": { | |
| "ref_id": "b21", | |
| "title": "Integrating Graph-Based and Transition-Based Dependency Parsers", | |
| "authors": [ | |
| { | |
| "first": "Joakim", | |
| "middle": [], | |
| "last": "Nivre", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ryan", | |
| "middle": [], | |
| "last": "Mcdonald", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "Proceedings of ACL-08: HLT", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "950--958", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Joakim Nivre and Ryan McDonald. 2008. Integrating Graph-Based and Transition-Based Dependency Parsers. In Proceedings of ACL-08: HLT, pages 950-958.", | |
| "links": null | |
| }, | |
| "BIBREF22": { | |
| "ref_id": "b22", | |
| "title": "The CoNLL 2007 Shared Task on Dependency Parsing", | |
| "authors": [ | |
| { | |
| "first": "Joakim", | |
| "middle": [], | |
| "last": "Nivre", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Johan", | |
| "middle": [], | |
| "last": "Hall", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Sandra", | |
| "middle": [], | |
| "last": "K\u00fcbler", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ryan", | |
| "middle": [], | |
| "last": "Mcdonald", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Jens", | |
| "middle": [], | |
| "last": "Nilsson", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Sebastian", | |
| "middle": [], | |
| "last": "Riedel", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Deniz", | |
| "middle": [], | |
| "last": "Yuret", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2007, | |
| "venue": "Proceedings of the CoNLL Shared Task Session of EMNLP-CoNLL", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "915--932", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Joakim Nivre, Johan Hall, Sandra K\u00fcbler, Ryan McDonald, Jens Nilsson, Sebastian Riedel, and Deniz Yuret. 2007a. The CoNLL 2007 Shared Task on Dependency Parsing. In Proceedings of the CoNLL Shared Task Session of EMNLP-CoNLL 2007, pages 915-932.", | |
| "links": null | |
| }, | |
| "BIBREF23": { | |
| "ref_id": "b23", | |
| "title": "MaltParser: A language-independent system for datadriven dependency parsing", | |
| "authors": [ | |
| { | |
| "first": "Joakim", | |
| "middle": [], | |
| "last": "Nivre", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Johan", | |
| "middle": [], | |
| "last": "Hall", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Jens", | |
| "middle": [], | |
| "last": "Nilsson", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Atanas", | |
| "middle": [], | |
| "last": "Chanev", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "G\u00fclsen", | |
| "middle": [], | |
| "last": "Eryigit", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Sandra", | |
| "middle": [], | |
| "last": "K\u00fcbler", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Svetoslav", | |
| "middle": [], | |
| "last": "Marinov", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Erwin", | |
| "middle": [], | |
| "last": "Marsi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2007, | |
| "venue": "Natural Language Engineering", | |
| "volume": "13", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Joakim Nivre, Johan Hall, Jens Nilsson, Atanas Chanev, G\u00fclsen Eryigit, Sandra K\u00fcbler, Svetoslav Marinov, and Erwin Marsi. 2007b. MaltParser: A language-independent system for data- driven dependency parsing. Natural Language Engineering, 13(January):1.", | |
| "links": null | |
| }, | |
| "BIBREF24": { | |
| "ref_id": "b24", | |
| "title": "An effective neural network model for graphbased dependency parsing", | |
| "authors": [ | |
| { | |
| "first": "Wenzhe", | |
| "middle": [], | |
| "last": "Pei", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Tao", | |
| "middle": [], | |
| "last": "Ge", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Baobao", | |
| "middle": [], | |
| "last": "Chang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "313--322", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Wenzhe Pei, Tao Ge, and Baobao Chang. 2015. An effective neural network model for graph- based dependency parsing. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 313-322.", | |
| "links": null | |
| }, | |
| "BIBREF25": { | |
| "ref_id": "b25", | |
| "title": "Dependency parsing and domain adaptation with LR models and parser ensembles", | |
| "authors": [ | |
| { | |
| "first": "Kenji", | |
| "middle": [], | |
| "last": "Sagae", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Jun'ichi", | |
| "middle": [], | |
| "last": "Tsujii", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2007, | |
| "venue": "Proceedings of the CoNLL Shared Task Session of EMNLP-CoNLL", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "1044--1050", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Kenji Sagae and Jun'ichi Tsujii. 2007. Dependency parsing and domain adaptation with LR models and parser ensembles. In Proceedings of the CoNLL Shared Task Session of EMNLP-CoNLL 2007, pages 1044-1050.", | |
| "links": null | |
| }, | |
| "BIBREF26": { | |
| "ref_id": "b26", | |
| "title": "Bidirectional recurrent neural networks", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Schuster", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [ | |
| "K" | |
| ], | |
| "last": "Paliwal", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1997, | |
| "venue": "IEEE Transactions on Signal Processing", | |
| "volume": "45", | |
| "issue": "11", | |
| "pages": "2673--2681", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "M. Schuster and K.K. Paliwal. 1997. Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing, 45(11):2673-2681.", | |
| "links": null | |
| }, | |
| "BIBREF27": { | |
| "ref_id": "b27", | |
| "title": "Ensemble models for dependency parsing: Cheap and good?", | |
| "authors": [ | |
| { | |
| "first": "Mihai", | |
| "middle": [], | |
| "last": "Surdeanu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Christopher", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Manning", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2010, | |
| "venue": "Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "649--652", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Mihai Surdeanu and Christopher D. Manning. 2010. Ensemble models for dependency parsing: Cheap and good? In Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 649-652.", | |
| "links": null | |
| }, | |
| "BIBREF28": { | |
| "ref_id": "b28", | |
| "title": "Building a treebank for Vietnamese dependency parsing", | |
| "authors": [ | |
| { | |
| "first": "Linh", | |
| "middle": [ | |
| "Ha" | |
| ], | |
| "last": "Luong Nguyen Thi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Hung", | |
| "middle": [ | |
| "Nguyen" | |
| ], | |
| "last": "My", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Huyen", | |
| "middle": [], | |
| "last": "Viet", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Thi", | |
| "middle": [], | |
| "last": "Nguyen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Phuong", | |
| "middle": [ | |
| "Le" | |
| ], | |
| "last": "Minh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Hong", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of 2013 IEEE RIVF International Conference on Computing and Communication Technologies, Research, Innovation, and Vision for the Future", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "147--151", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Luong Nguyen Thi, Linh Ha My, Hung Nguyen Viet, Huyen Nguyen Thi Minh, and Phuong Le Hong. 2013. Building a treebank for Vietnamese dependency parsing. In Proceedings of 2013 IEEE RIVF International Conference on Computing and Communication Technologies, Research, Innovation, and Vision for the Future, pages 147-151.", | |
| "links": null | |
| }, | |
| "BIBREF29": { | |
| "ref_id": "b29", | |
| "title": "A Vietnamese Reference Grammar", | |
| "authors": [ | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Laurence", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Thompson", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1987, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Laurence C. Thompson. 1987. A Vietnamese Reference Grammar. University of Hawaii Press.", | |
| "links": null | |
| }, | |
| "BIBREF30": { | |
| "ref_id": "b30", | |
| "title": "Improving vietnamese dependency parsing using distributed word representations", | |
| "authors": [ | |
| { | |
| "first": "Cam", | |
| "middle": [], | |
| "last": "Vu-Manh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Anh", | |
| "middle": [ | |
| "Tuan" | |
| ], | |
| "last": "Luong", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Phuong", | |
| "middle": [], | |
| "last": "Le-Hong", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the Sixth International Symposium on Information and Communication Technology", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "54--60", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Cam Vu-Manh, Anh Tuan Luong, and Phuong Le- Hong. 2015. Improving vietnamese dependency parsing using distributed word representations. In Proceedings of the Sixth International Symposium on Information and Communication Technology, pages 54-60.", | |
| "links": null | |
| }, | |
| "BIBREF31": { | |
| "ref_id": "b31", | |
| "title": "Structured training for neural network transition-based parsing", | |
| "authors": [ | |
| { | |
| "first": "David", | |
| "middle": [], | |
| "last": "Weiss", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Chris", | |
| "middle": [], | |
| "last": "Alberti", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Michael", | |
| "middle": [], | |
| "last": "Collins", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Slav", | |
| "middle": [], | |
| "last": "Petrov", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "323--333", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "David Weiss, Chris Alberti, Michael Collins, and Slav Petrov. 2015. Structured training for neural network transition-based parsing. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 323-333.", | |
| "links": null | |
| }, | |
| "BIBREF32": { | |
| "ref_id": "b32", | |
| "title": "A Tale of Two Parsers: Investigating and Combining Graphbased and Transition-based Dependency Parsing", | |
| "authors": [ | |
| { | |
| "first": "Yue", | |
| "middle": [], | |
| "last": "Zhang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Stephen", | |
| "middle": [], | |
| "last": "Clark", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "562--571", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Yue Zhang and Stephen Clark. 2008. A Tale of Two Parsers: Investigating and Combining Graph- based and Transition-based Dependency Parsing. In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing, pages 562-571.", | |
| "links": null | |
| }, | |
| "BIBREF33": { | |
| "ref_id": "b33", | |
| "title": "Transition-based dependency parsing with rich non-local features", | |
| "authors": [ | |
| { | |
| "first": "Yue", | |
| "middle": [], | |
| "last": "Zhang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Joakim", | |
| "middle": [], | |
| "last": "Nivre", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "188--193", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Yue Zhang and Joakim Nivre. 2011. Transition-based dependency parsing with rich non-local features. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, pages 188-193.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "num": null, | |
| "type_str": "figure", | |
| "uris": null, | |
| "text": "LAS by sentence length." | |
| }, | |
| "FIGREF1": { | |
| "num": null, | |
| "type_str": "figure", | |
| "uris": null, | |
| "text": "UAS by sentence length" | |
| }, | |
| "FIGREF2": { | |
| "num": null, | |
| "type_str": "figure", | |
| "uris": null, | |
| "text": "F 1 scores by dependency distance for labeled attachmentFigure 4: F 1 scores by dependency distance for unlabeled attachment" | |
| }, | |
| "TABREF1": { | |
| "content": "<table><tr><td/><td colspan=\"2\">With punctuation</td><td/><td colspan=\"2\">Without punctuation</td></tr><tr><td>System</td><td>Overall</td><td colspan=\"2\">Exact match</td><td>Overall</td><td>Exact match</td></tr><tr><td/><td>LAS UAS LS</td><td colspan=\"2\">LAS UAS LS</td><td>LAS UAS LS</td><td>LAS UAS LS</td></tr><tr><td colspan=\"3\">Gold POS BistG 73.Auto POS BistG 68.40 76.28 80.56 9.12 BistT 68.22 76.56 80.22 9.80 MST 65.99 73.94 79.78 6.86 Malt 64.94 72.32 78.43 7.35</td><td colspan=\"3\">16.18 11.76 68.50 77.55 77.65 9.71 16.27 13.24 68.31 77.91 77.27 10.00 17.06 13.24 17.25 11.76 10.78 10.88 66.99 76.12 76.75 7.84 13.33 10.88 12.25 10.20 65.88 74.36 75.56 7.55 14.02 10.20</td></tr></table>", | |
| "text": "https://github.com/elikip/bist-parser/tree/master/bmstparser 2 http://www.seas.upenn.edu/~strctlrn/MSTParser/MSTParser.html 3 https://github.com/elikip/bist-parser/tree/master/barchybrid 15.20 73.53 80.66 81.86 11.96 20.88 15.20 BistT 72.53 79.33 83.71 11.27 19.41 16.18 72.91 80.73 81.29 11.67 20.29 16.18 MST 70.29 76.47 83.23 8.43 12.94 14.02 71.61 78.71 80.72 9.80 16.37 14.02 Malt 69.10 74.91 81.72 9.22 14.80 13.92 70.39 77.08 79.33 9.71 17.16 13.92", | |
| "html": null, | |
| "type_str": "table", | |
| "num": null | |
| }, | |
| "TABREF2": { | |
| "content": "<table/>", | |
| "text": "Parsing results. \"Without punctuation\" denotes parsing results where the punctuation and other symbols are excluded from evaluation. \"Exact match\" denotes the proportion of sentences whose predicted dependency trees are entirely correct. tags (73.17% against 72.53%, and 73.53% against 72.91%).", | |
| "html": null, | |
| "type_str": "table", | |
| "num": null | |
| }, | |
| "TABREF3": { | |
| "content": "<table><tr><td>Type</td><td>BistG BistT MST</td><td>Malt</td><td>Avg.</td></tr><tr><td>adv</td><td colspan=\"3\">92.09 56</td></tr><tr><td>dob</td><td colspan=\"3\">73.01 70.81 78.62 76.35 74.70</td></tr><tr><td>loc</td><td colspan=\"3\">52.54 53.86 51.43 50.77 52.15</td></tr><tr><td colspan=\"4\">nmod 79.34 79.51 78.10 76.67 78.41</td></tr><tr><td>pob</td><td colspan=\"3\">94.35 95.27 96.18 95.85 95.41</td></tr><tr><td>root</td><td colspan=\"3\">85.69 82.55 82.06 74.41 81.18</td></tr><tr><td>sub</td><td colspan=\"3\">73.34 72.61 66.49 62.67 68.78</td></tr><tr><td>tmp</td><td colspan=\"3\">60.68 57.05 44.66 41.45 50.96</td></tr><tr><td colspan=\"4\">vmod 61.51 62.02 60.79 60.23 61.14</td></tr></table>", | |
| "text": "92.33 92.31 amod 77.30 73.89 76.11 73.21 75.13 conj 74.82 73.11 78 .00 71.64 74.39 coord 57.49 49.52 46.14 52.66 51.45 dep 47.83 46 .00 32.54 42.08 42.11 det 94.15 94.30 95.27 94.52 94.", | |
| "html": null, | |
| "type_str": "table", | |
| "num": null | |
| }, | |
| "TABREF4": { | |
| "content": "<table><tr><td>POS</td><td>LAS BistG BistT MST Malt BistG BistT MST Malt UAS</td></tr><tr><td>A</td><td>68.32 70.31 69.89 66.83 73.01 75.50 74.86 70.88</td></tr><tr><td>C</td><td>55.90 50.00 44.94 50.00 61.33 56.87 50.60 54.94</td></tr><tr><td>E</td><td>55.47 53.87 50.91 49.96 72.27 71.54 68.86 64.45</td></tr><tr><td>M</td><td>92.11 91.05 93.03 91.18 93.42 93.16 94.21 91.71</td></tr><tr><td>N</td><td>74.37 73.58 73.77 71.30 83.95 83.86 82.58 80.48</td></tr><tr><td colspan=\"2\">Nc 69.86 72.02 68.49 67.12 78.47 79.26 76.13 74.17</td></tr><tr><td colspan=\"2\">Np 84.69 84.47 84.80 82.84 88.49 88.49 88.06 86.43</td></tr><tr><td>P</td><td>79.34 80.16 79.69 77.23 85.45 85.92 84.62 82.39</td></tr><tr><td>R</td><td>91.94 93.08 92.42 92.60 92.90 93.87 92.96 93.27</td></tr><tr><td>V</td><td>68.01 66.49 63.78 63.13 75.05 74.95 71.83 70.49</td></tr></table>", | |
| "text": "LAS by most frequent dependency labels. \"Avg.\" denotes the averaged score of four parsers.", | |
| "html": null, | |
| "type_str": "table", | |
| "num": null | |
| }, | |
| "TABREF5": { | |
| "content": "<table/>", | |
| "text": "Results by most frequent POS tags.", | |
| "html": null, | |
| "type_str": "table", | |
| "num": null | |
| }, | |
| "TABREF7": { | |
| "content": "<table/>", | |
| "text": "Upper bound of ensemble performance.", | |
| "html": null, | |
| "type_str": "table", | |
| "num": null | |
| } | |
| } | |
| } | |
| } |