acl acl2013 acl2013-372 knowledge-graph by maker-knowledge-mining
Source: pdf
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.
Reference: text
sentIndex sentText sentNum sentScore
1 Using CCG categories to improve Hindi dependency parsing Bharat Ram Ambati Tejaswini Deoskar Mark Steedman Institute for Language, Cognition and Computation School of Informatics, University of Edinburgh bharat . [sent-1, score-0.363]
2 uk @ 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. [sent-7, score-0.364]
3 We first describe a novel way to obtain a CCG lexicon and treebank from an existing dependency treebank, using a CCG parser. [sent-8, score-0.373]
4 We use the output of a supertagger trained on the CCGbank as a feature for a state-of-the-art Hindi dependency parser (Malt). [sent-9, score-0.325]
5 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. [sent-10, score-0.198]
6 1 Introduction As compared to English, many Indian languages including Hindi have a freer word order and are also morphologically richer. [sent-11, score-0.045]
7 Today, the best dependency parsing accuracies for Hindi are obtained by the shift-reduce parser of Nivre et al. [sent-13, score-0.224]
8 It has been observed that Malt is relatively accurate at recovering short distance dependencies, like arguments of a verb, but is less accurate at recovering long distance dependencies like co-ordination, root of the sentence, etc (Mcdonald and Nivre, 2007; Ambati et al. [sent-15, score-0.538]
9 In this work, we show that using CCG lexical categories (Steedman, 2000), which contain subcategorization information and capture long distance dependencies elegantly, can help Malt with those dependencies. [sent-17, score-0.321]
10 Section 2 first shows how we extract a CCG lexicon from an existing Hindi dependency treebank (Bhatt et al. [sent-18, score-0.373]
11 In section 3, we develop a supertagger using the CCGbank and explore different ways of providing CCG categories from the supertagger as features to Malt. [sent-20, score-0.356]
12 Our re- sults show that using CCG categories can help Malt by improving the recovery of long distance relations. [sent-21, score-0.267]
13 (2009) created a CCGbank from an Italian dependency treebank by converting dependency trees into phrase structure trees and then applying an algorithm similar to Hockenmaier and Steedman (2007). [sent-24, score-0.507]
14 In this work, following C ¸akıcı (2005), we first extract a Hindi CCG lexicon from a dependency treebank. [sent-25, score-0.208]
15 We then use a CKY parser based on the CCG formalism to automatically obtain a treebank of CCG derivations from this lexicon, a novel methodology that may be applicable to obtaining CCG treebanks in other languages as well. [sent-26, score-0.401]
16 5) released as part of Coling 2012 Shared Task on parsing (Bharati et al. [sent-29, score-0.058]
17 HDT is a multi-layered dependency treebank (Bhatt et al. [sent-31, score-0.323]
18 , 2009) annotated with morpho-syntactic (morphological, part-of-speech and chunk information) and syntactico-semantic (dependency) information (Bharati et al. [sent-32, score-0.223]
19 Dependency labels are fine-grained, and mark dependencies that are syntactico-semantic in nature, such as agent (usually corresponding to subject), patient (object), and time and place expressions. [sent-35, score-0.224]
20 There are special labels to mark long distance relations like relative clauses, co-ordination etc 604 Proce dingSsof oifa, th Beu 5l1gsarti Aan,An u aglu Mste 4e-ti9n2g 0 o1f3 t. [sent-36, score-0.249]
21 The treebank contains 12,041 training, 1,233 development and 1,828 testing sentences with an average of 22 words per sentence. [sent-41, score-0.165]
22 We used the CoNLL format1 for our purposes, which contains word, lemma, pos-tag, and coarse pos-tag in the WORD, LEMMA, POS, and CPOS fields respectively and morphological features and chunk information in the FEATS column. [sent-42, score-0.328]
23 2 Algorithm We first made a list of argument and adjunct dependency labels in the treebank. [sent-44, score-0.29]
24 , dependencies with the label k1 and k2 (corresponding to subject and object respectively) are considered to be arguments, while labels like k7p and k7t (corresponding to place and time expressions) are considered to be adjuncts. [sent-47, score-0.198]
25 For readability reasons, we will henceforth refer to dependency labels with their English equivalents (e. [sent-48, score-0.294]
26 , SUBJ, OBJ, PURPOSE, CASE for k1, k2 , rt, lwg psp respectively). [sent-50, score-0.048]
27 Starting from the root of the dependency tree, we traverse each node. [sent-51, score-0.271]
28 The category of a node depends on both its parent and children. [sent-52, score-0.275]
29 If the node is an argument of its parent, we assign the chunk tag of the node (e. [sent-53, score-0.455]
30 Otherwise, we assign it a category of X | X, where X is the parent’s result category and | is directionality (\ or /), ws hreicsuhl depends on tdh |e position noftahliet yno (\de o rw. [sent-56, score-0.392]
31 nTdhse ornes uthlte category ooff a node is the category obtained once its arguments are resolved. [sent-60, score-0.342]
32 For example, S, is the result category for ( S\NP ) \NP. [sent-61, score-0.1]
33 Once we get the partial category foofr a nSo\dNeP P b)a\seNdP on tnhcee wnoede g’est parent information, we traverse through the children of the node. [sent-62, score-0.336]
34 If a child is an argument, we add that child’s chunk tag, with appropriate directionality, to the node’s category. [sent-63, score-0.261]
35 The algorithm is sketched in Figure 1 and an example of a CCG derivation for a simple sentence (marked with chunk tags; NP and VGF are the chunk tags for noun and finite verb chunks respectively. [sent-64, score-0.549]
36 In Type 1, we keep morphological information in noun categories and in Type 2, we don’t. [sent-68, score-0.168]
37 For example, consider a noun chunk ‘raam ne (Ram ERG)’ . [sent-69, score-0.337]
38 In Type 1, CCG categories for ‘raam’ and ‘ne’ are NP and 1http://nextens. [sent-70, score-0.088]
wordName wordTfidf (topN-words)
[('ccg', 0.545), ('hindi', 0.282), ('malt', 0.252), ('chunk', 0.223), ('bharati', 0.223), ('raam', 0.219), ('treebank', 0.165), ('dependency', 0.158), ('steedman', 0.138), ('supertagger', 0.134), ('mohan', 0.126), ('ccgbank', 0.126), ('np', 0.115), ('hdt', 0.109), ('khariidii', 0.109), ('kitaab', 0.109), ('hockenmaier', 0.107), ('treebanks', 0.102), ('ram', 0.101), ('category', 0.1), ('ambati', 0.089), ('bhatt', 0.089), ('dependencies', 0.088), ('categories', 0.088), ('bharat', 0.084), ('ne', 0.082), ('erg', 0.08), ('parent', 0.078), ('directionality', 0.076), ('node', 0.067), ('traverse', 0.065), ('etc', 0.064), ('recovering', 0.061), ('distance', 0.06), ('ak', 0.058), ('lie', 0.054), ('formalism', 0.051), ('long', 0.05), ('lexicon', 0.05), ('derivations', 0.05), ('ke', 0.05), ('morphological', 0.048), ('root', 0.048), ('psp', 0.048), ('foofr', 0.048), ('nso', 0.048), ('labels', 0.048), ('arguments', 0.046), ('argument', 0.046), ('tnhcee', 0.045), ('elegantly', 0.045), ('freer', 0.045), ('nivre', 0.044), ('tse', 0.042), ('sketched', 0.042), ('lemma', 0.04), ('gf', 0.04), ('child', 0.038), ('adjunct', 0.038), ('cky', 0.037), ('equivalents', 0.036), ('obj', 0.036), ('sults', 0.036), ('ed', 0.035), ('subj', 0.035), ('subcategorization', 0.035), ('est', 0.034), ('lexicalised', 0.034), ('parsing', 0.033), ('parser', 0.033), ('tdh', 0.033), ('combinatory', 0.033), ('recovery', 0.033), ('noun', 0.032), ('respectively', 0.032), ('bos', 0.031), ('place', 0.031), ('object', 0.031), ('patient', 0.03), ('pose', 0.03), ('inf', 0.03), ('depends', 0.03), ('indian', 0.03), ('ooff', 0.029), ('curran', 0.029), ('chunks', 0.029), ('today', 0.028), ('categorial', 0.028), ('book', 0.027), ('mark', 0.027), ('ws', 0.027), ('readability', 0.026), ('assign', 0.026), ('henceforth', 0.026), ('converting', 0.026), ('tag', 0.026), ('italian', 0.026), ('clauses', 0.025), ('coarse', 0.025), ('released', 0.025)]
simIndex simValue paperId paperTitle
same-paper 1 1.0000001 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.
2 0.45695505 199 acl-2013-Integrating Multiple Dependency Corpora for Inducing Wide-coverage Japanese CCG Resources
Author: Sumire Uematsu ; Takuya Matsuzaki ; Hiroki Hanaoka ; Yusuke Miyao ; Hideki Mima
Abstract: This paper describes a method of inducing wide-coverage CCG resources for Japanese. While deep parsers with corpusinduced grammars have been emerging for some languages, those for Japanese have not been widely studied, mainly because most Japanese syntactic resources are dependency-based. Our method first integrates multiple dependency-based corpora into phrase structure trees and then converts the trees into CCG derivations. The method is empirically evaluated in terms of the coverage of the obtained lexi- con and the accuracy of parsing.
3 0.31071728 357 acl-2013-Transfer Learning for Constituency-Based Grammars
Author: Yuan Zhang ; Regina Barzilay ; Amir Globerson
Abstract: In this paper, we consider the problem of cross-formalism transfer in parsing. We are interested in parsing constituencybased grammars such as HPSG and CCG using a small amount of data specific for the target formalism, and a large quantity of coarse CFG annotations from the Penn Treebank. While all of the target formalisms share a similar basic syntactic structure with Penn Treebank CFG, they also encode additional constraints and semantic features. To handle this apparent discrepancy, we design a probabilistic model that jointly generates CFG and target formalism parses. The model includes features of both parses, allowing trans- fer between the formalisms, while preserving parsing efficiency. We evaluate our approach on three constituency-based grammars CCG, HPSG, and LFG, augmented with the Penn Treebank-1. Our experiments show that across all three formalisms, the target parsers significantly benefit from the coarse annotations.1 —
4 0.26923689 347 acl-2013-The Role of Syntax in Vector Space Models of Compositional Semantics
Author: Karl Moritz Hermann ; Phil Blunsom
Abstract: Modelling the compositional process by which the meaning of an utterance arises from the meaning of its parts is a fundamental task of Natural Language Processing. In this paper we draw upon recent advances in the learning of vector space representations of sentential semantics and the transparent interface between syntax and semantics provided by Combinatory Categorial Grammar to introduce Combinatory Categorial Autoencoders. This model leverages the CCG combinatory operators to guide a non-linear transformation of meaning within a sentence. We use this model to learn high dimensional embeddings for sentences and evaluate them in a range of tasks, demonstrating that the incorporation of syntax allows a concise model to learn representations that are both effective and general.
5 0.12544009 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.
6 0.11310245 204 acl-2013-Iterative Transformation of Annotation Guidelines for Constituency Parsing
7 0.110602 368 acl-2013-Universal Dependency Annotation for Multilingual Parsing
8 0.10833611 136 acl-2013-Enhanced and Portable Dependency Projection Algorithms Using Interlinear Glossed Text
9 0.095692508 208 acl-2013-Joint Inference for Heterogeneous Dependency Parsing
10 0.095197141 94 acl-2013-Coordination Structures in Dependency Treebanks
11 0.090373799 343 acl-2013-The Effect of Higher-Order Dependency Features in Discriminative Phrase-Structure Parsing
12 0.089372925 313 acl-2013-Semantic Parsing with Combinatory Categorial Grammars
13 0.073954724 280 acl-2013-Plurality, Negation, and Quantification:Towards Comprehensive Quantifier Scope Disambiguation
14 0.068367876 270 acl-2013-ParGramBank: The ParGram Parallel Treebank
15 0.067399591 358 acl-2013-Transition-based Dependency Parsing with Selectional Branching
16 0.065080225 70 acl-2013-Bilingually-Guided Monolingual Dependency Grammar Induction
17 0.063893445 26 acl-2013-A Transition-Based Dependency Parser Using a Dynamic Parsing Strategy
18 0.063666545 155 acl-2013-Fast and Accurate Shift-Reduce Constituent Parsing
19 0.063382246 28 acl-2013-A Unified Morpho-Syntactic Scheme of Stanford Dependencies
20 0.062659547 44 acl-2013-An Empirical Examination of Challenges in Chinese Parsing
topicId topicWeight
[(0, 0.129), (1, -0.078), (2, -0.172), (3, -0.011), (4, -0.234), (5, -0.017), (6, 0.049), (7, -0.003), (8, 0.145), (9, -0.096), (10, 0.041), (11, 0.02), (12, 0.207), (13, 0.035), (14, -0.108), (15, 0.045), (16, 0.066), (17, 0.008), (18, -0.254), (19, 0.096), (20, 0.039), (21, -0.238), (22, -0.246), (23, -0.037), (24, 0.183), (25, 0.058), (26, -0.058), (27, 0.068), (28, -0.006), (29, -0.086), (30, 0.147), (31, -0.073), (32, -0.021), (33, 0.153), (34, -0.025), (35, -0.024), (36, -0.052), (37, 0.057), (38, -0.068), (39, 0.043), (40, 0.019), (41, 0.037), (42, -0.089), (43, 0.014), (44, -0.026), (45, -0.078), (46, 0.023), (47, -0.012), (48, 0.035), (49, 0.016)]
simIndex simValue paperId paperTitle
same-paper 1 0.95410544 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.
2 0.94717652 199 acl-2013-Integrating Multiple Dependency Corpora for Inducing Wide-coverage Japanese CCG Resources
Author: Sumire Uematsu ; Takuya Matsuzaki ; Hiroki Hanaoka ; Yusuke Miyao ; Hideki Mima
Abstract: This paper describes a method of inducing wide-coverage CCG resources for Japanese. While deep parsers with corpusinduced grammars have been emerging for some languages, those for Japanese have not been widely studied, mainly because most Japanese syntactic resources are dependency-based. Our method first integrates multiple dependency-based corpora into phrase structure trees and then converts the trees into CCG derivations. The method is empirically evaluated in terms of the coverage of the obtained lexi- con and the accuracy of parsing.
3 0.77440596 357 acl-2013-Transfer Learning for Constituency-Based Grammars
Author: Yuan Zhang ; Regina Barzilay ; Amir Globerson
Abstract: In this paper, we consider the problem of cross-formalism transfer in parsing. We are interested in parsing constituencybased grammars such as HPSG and CCG using a small amount of data specific for the target formalism, and a large quantity of coarse CFG annotations from the Penn Treebank. While all of the target formalisms share a similar basic syntactic structure with Penn Treebank CFG, they also encode additional constraints and semantic features. To handle this apparent discrepancy, we design a probabilistic model that jointly generates CFG and target formalism parses. The model includes features of both parses, allowing trans- fer between the formalisms, while preserving parsing efficiency. We evaluate our approach on three constituency-based grammars CCG, HPSG, and LFG, augmented with the Penn Treebank-1. Our experiments show that across all three formalisms, the target parsers significantly benefit from the coarse annotations.1 —
4 0.59463 347 acl-2013-The Role of Syntax in Vector Space Models of Compositional Semantics
Author: Karl Moritz Hermann ; Phil Blunsom
Abstract: Modelling the compositional process by which the meaning of an utterance arises from the meaning of its parts is a fundamental task of Natural Language Processing. In this paper we draw upon recent advances in the learning of vector space representations of sentential semantics and the transparent interface between syntax and semantics provided by Combinatory Categorial Grammar to introduce Combinatory Categorial Autoencoders. This model leverages the CCG combinatory operators to guide a non-linear transformation of meaning within a sentence. We use this model to learn high dimensional embeddings for sentences and evaluate them in a range of tasks, demonstrating that the incorporation of syntax allows a concise model to learn representations that are both effective and general.
5 0.51458079 313 acl-2013-Semantic Parsing with Combinatory Categorial Grammars
Author: Yoav Artzi ; Nicholas FitzGerald ; Luke Zettlemoyer
Abstract: unkown-abstract
6 0.41830036 335 acl-2013-Survey on parsing three dependency representations for English
7 0.39570722 270 acl-2013-ParGramBank: The ParGram Parallel Treebank
8 0.38790822 94 acl-2013-Coordination Structures in Dependency Treebanks
9 0.33850938 368 acl-2013-Universal Dependency Annotation for Multilingual Parsing
10 0.3339197 208 acl-2013-Joint Inference for Heterogeneous Dependency Parsing
11 0.33051443 204 acl-2013-Iterative Transformation of Annotation Guidelines for Constituency Parsing
12 0.30276585 28 acl-2013-A Unified Morpho-Syntactic Scheme of Stanford Dependencies
13 0.28172144 343 acl-2013-The Effect of Higher-Order Dependency Features in Discriminative Phrase-Structure Parsing
14 0.27569443 331 acl-2013-Stop-probability estimates computed on a large corpus improve Unsupervised Dependency Parsing
15 0.26326984 280 acl-2013-Plurality, Negation, and Quantification:Towards Comprehensive Quantifier Scope Disambiguation
16 0.25967619 215 acl-2013-Large-scale Semantic Parsing via Schema Matching and Lexicon Extension
17 0.24701604 13 acl-2013-A New Syntactic Metric for Evaluation of Machine Translation
18 0.24537204 176 acl-2013-Grounded Unsupervised Semantic Parsing
19 0.23913474 311 acl-2013-Semantic Neighborhoods as Hypergraphs
20 0.23749755 367 acl-2013-Universal Conceptual Cognitive Annotation (UCCA)
topicId topicWeight
[(0, 0.01), (11, 0.076), (14, 0.02), (24, 0.02), (26, 0.026), (35, 0.052), (42, 0.617), (48, 0.017), (70, 0.013), (95, 0.028)]
simIndex simValue paperId paperTitle
Author: Sina Zarriess ; Jonas Kuhn
Abstract: We suggest a generation task that integrates discourse-level referring expression generation and sentence-level surface realization. We present a data set of German articles annotated with deep syntax and referents, including some types of implicit referents. Our experiments compare several architectures varying the order of a set of trainable modules. The results suggest that a revision-based pipeline, with intermediate linearization, significantly outperforms standard pipelines or a parallel architecture.
same-paper 2 0.97601533 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.
3 0.96461612 125 acl-2013-Distortion Model Considering Rich Context for Statistical Machine Translation
Author: Isao Goto ; Masao Utiyama ; Eiichiro Sumita ; Akihiro Tamura ; Sadao Kurohashi
Abstract: This paper proposes new distortion models for phrase-based SMT. In decoding, a distortion model estimates the source word position to be translated next (NP) given the last translated source word position (CP). We propose a distortion model that can consider the word at the CP, a word at an NP candidate, and the context of the CP and the NP candidate simultaneously. Moreover, we propose a further improved model that considers richer context by discriminating label sequences that specify spans from the CP to NP candidates. It enables our model to learn the effect of relative word order among NP candidates as well as to learn the effect of distances from the training data. In our experiments, our model improved 2.9 BLEU points for Japanese-English and 2.6 BLEU points for Chinese-English translation compared to the lexical reordering models.
4 0.91868615 64 acl-2013-Automatically Predicting Sentence Translation Difficulty
Author: Abhijit Mishra ; Pushpak Bhattacharyya ; Michael Carl
Abstract: In this paper we introduce Translation Difficulty Index (TDI), a measure of difficulty in text translation. We first define and quantify translation difficulty in terms of TDI. We realize that any measure of TDI based on direct input by translators is fraught with subjectivity and adhocism. We, rather, rely on cognitive evidences from eye tracking. TDI is measured as the sum of fixation (gaze) and saccade (rapid eye movement) times of the eye. We then establish that TDI is correlated with three properties of the input sentence, viz. length (L), degree of polysemy (DP) and structural complexity (SC). We train a Support Vector Regression (SVR) system to predict TDIs for new sentences using these features as input. The prediction done by our framework is well correlated with the empirical gold standard data, which is a repository of < L, DP, SC > and TDI pairs for a set of sentences. The primary use of our work is a way of “binning” sentences (to be translated) in “easy”, “medium” and “hard” categories as per their predicted TDI. This can decide pricing of any translation task, especially useful in a scenario where parallel corpora for Machine Translation are built through translation crowdsourcing/outsourcing. This can also provide a way of monitoring progress of second language learners.
5 0.90923113 11 acl-2013-A Multi-Domain Translation Model Framework for Statistical Machine Translation
Author: Rico Sennrich ; Holger Schwenk ; Walid Aransa
Abstract: While domain adaptation techniques for SMT have proven to be effective at improving translation quality, their practicality for a multi-domain environment is often limited because of the computational and human costs of developing and maintaining multiple systems adapted to different domains. We present an architecture that delays the computation of translation model features until decoding, allowing for the application of mixture-modeling techniques at decoding time. We also de- scribe a method for unsupervised adaptation with development and test data from multiple domains. Experimental results on two language pairs demonstrate the effectiveness of both our translation model architecture and automatic clustering, with gains of up to 1BLEU over unadapted systems and single-domain adaptation.
6 0.89293015 302 acl-2013-Robust Automated Natural Language Processing with Multiword Expressions and Collocations
7 0.88583517 206 acl-2013-Joint Event Extraction via Structured Prediction with Global Features
8 0.87497175 40 acl-2013-Advancements in Reordering Models for Statistical Machine Translation
9 0.67712295 166 acl-2013-Generalized Reordering Rules for Improved SMT
10 0.67076117 281 acl-2013-Post-Retrieval Clustering Using Third-Order Similarity Measures
11 0.66437203 77 acl-2013-Can Markov Models Over Minimal Translation Units Help Phrase-Based SMT?
12 0.65578765 56 acl-2013-Argument Inference from Relevant Event Mentions in Chinese Argument Extraction
13 0.64431995 199 acl-2013-Integrating Multiple Dependency Corpora for Inducing Wide-coverage Japanese CCG Resources
14 0.63117385 38 acl-2013-Additive Neural Networks for Statistical Machine Translation
15 0.61457515 69 acl-2013-Bilingual Lexical Cohesion Trigger Model for Document-Level Machine Translation
16 0.59891194 127 acl-2013-Docent: A Document-Level Decoder for Phrase-Based Statistical Machine Translation
17 0.59092295 68 acl-2013-Bilingual Data Cleaning for SMT using Graph-based Random Walk
18 0.58638853 363 acl-2013-Two-Neighbor Orientation Model with Cross-Boundary Global Contexts
19 0.58221054 208 acl-2013-Joint Inference for Heterogeneous Dependency Parsing
20 0.57862282 181 acl-2013-Hierarchical Phrase Table Combination for Machine Translation