acl acl2013 acl2013-189 knowledge-graph by maker-knowledge-mining
Source: pdf
Author: Egoitz Laparra ; German Rigau
Abstract: This paper presents a novel deterministic algorithm for implicit Semantic Role Labeling. The system exploits a very simple but relevant discursive property, the argument coherence over different instances of a predicate. The algorithm solves the implicit arguments sequentially, exploiting not only explicit but also the implicit arguments previously solved. In addition, we empirically demonstrate that the algorithm obtains very competitive and robust performances with respect to supervised approaches that require large amounts of costly training data.
Reference: text
sentIndex sentText sentNum sentScore
1 e s Abstract This paper presents a novel deterministic algorithm for implicit Semantic Role Labeling. [sent-3, score-0.437]
2 The system exploits a very simple but relevant discursive property, the argument coherence over different instances of a predicate. [sent-4, score-0.296]
3 The algorithm solves the implicit arguments sequentially, exploiting not only explicit but also the implicit arguments previously solved. [sent-5, score-1.154]
4 1 Introduction Traditionally, Semantic Role Labeling (SRL) systems have focused in searching the fillers of those explicit roles appearing within sentence boundaries (Gildea and Jurafsky, 2000, 2002; Carreras and M `arquez, 2005; Surdeanu et al. [sent-7, score-0.274]
5 However, when the participants of a predicate are implicit this approach obtains incomplete predicative structures with null arguments. [sent-11, score-0.673]
6 The previous analysis includes annotations for the nominal predicate loss based on the NomBank structure (Meyers et al. [sent-14, score-0.411]
7 Traditional SRL systems facing this type of examples are not able to fill the arguments of a predicate because their fillers are not in the same sentence of the predicate. [sent-20, score-0.579]
8 For the predicate plan in the previous sentence, a traditional SRL process only returns the filler for the argument arg1, the theme of the plan. [sent-22, score-0.721]
9 In fact, Gerber and Chai (2010) pointed out that implicit arguments can increase the coverage of argument structures in NomBank by 71%. [sent-25, score-0.704]
10 In this work, we study the coherence of the predicate and argument realization in discourse. [sent-31, score-0.521]
11 (1987) who filled the arguments of anaphoric mentions of nominal predicates using previous mentions of the same predicate. [sent-35, score-0.342]
12 We present an extension of this idea assuming that in a coherent document the different ocurrences of a predicate, including both verbal and nominal forms, tend to be mentions of the same event, and thus, they share the same argument fillers. [sent-36, score-0.329]
13 That is, our system can be applied to any predicate without training data. [sent-38, score-0.27]
14 The main contributions of this work are the following: • We empirically prove that there exists a strong mdipsicroicuarllsye relationship t be thtwereeen e txhiset ism aplicit and explicit argument fillers ofthe same predicates. [sent-39, score-0.481]
15 • • • We propose a deterministic approach that exploits tohpios sdeis ac doeutresrimvei property irno oarchde trh taot oexb-tain the fillers of implicit arguments. [sent-40, score-0.584]
16 We adapt to the implicit SRL problem a classWice algorithm hfoer i pronoun R reLso plruotbiolne. [sent-41, score-0.382]
17 Section 4 describes our algorithm for implicit argument resolution. [sent-47, score-0.589]
18 2 Related Work The first attempt for the automatic annotation of implicit semantic roles was proposed by Palmer et al. [sent-51, score-0.391]
19 (1991) also attempted to solve implicit 1http : / / adimen . [sent-55, score-0.342]
20 (2010) presented a task in SemEval-2010 that included an implicit argument identification challenge based on FrameNet (Baker et al. [sent-64, score-0.549]
21 , 2010) is a supervised system that extended an existing semantic role labeler to enlarge the search window to other sentences, replacing the features defined for regular arguments with two new semantic features. [sent-72, score-0.295]
22 On the other hand, Gerber and Chai (2010, 2012) studied the implicit argument resolution on NomBank. [sent-83, score-0.605]
23 This works represent the deepest study so far of the features that characterizes the implicit arguments 2. [sent-91, score-0.497]
24 However, many of the most important features are lexically dependent on the predicate and cannot been generalized. [sent-92, score-0.27]
25 Thus, specific annotations are required for each new predicate to be analyzed. [sent-93, score-0.322]
26 All the works presented in this section agree that implicit arguments must be modeled as a particular case of coreference together with features that include lexical-semantic information, to build selectional preferences. [sent-94, score-0.545]
27 Another common point is the fact that these works try to solve each instance of the implicit arguments independently, without taking into account the previous realizations of the same implicit argument in the document. [sent-95, score-1.08]
28 We propose that these realizations, together with the explicit ones, must maintain a certain coherence along the document and, in consequence, the filler of an argument remains the same along the following instances of that argument until a stronger evidence indicates a change. [sent-96, score-0.862]
29 This dataset (hereinafter BNB which stands for ”Beyond NomBank”) extends existing predicate annotations for NomBank and ProbBank. [sent-99, score-0.322]
30 BNB presented the first annotation work of implicit arguments based on PropBank and NomBank frames. [sent-100, score-0.497]
31 These predicates are all nominalizations of other verbal predicates, without sense ambiguity, that appear frequently in the corpus and tend to have implicit arguments associated with their instances. [sent-104, score-0.64]
32 These constraints allowed them to model enough occurrences of each implicit argument in order to cover adequately all the possible cases appearing in a test document. [sent-105, score-0.549]
33 For each missing argument position they went over all the preceding sentences and annotated all mentions of the filler of that argument. [sent-106, score-0.471]
34 1 Discoursive coherence of predicates Exploring the training dataset of BNB, we observed a very strong discourse effect on the implicit and explicit argument fillers of the predicates. [sent-112, score-0.943]
35 That is, if several instances of the same predicate appear in a well-written discourse, it is very likely that they maintain the same argument fillers. [sent-113, score-0.563]
36 This property holds when joining the different parts-of-speech of the predicates (nominal or verbal) and the explicit or implicit realizations of the argument fillers. [sent-114, score-0.812]
37 For instance, we observed that 46% of all implicit arguments share the same filler with the previous instance of the same predicate while only 14% of them have a different filler. [sent-115, score-0.965]
38 The remaining 40% of all implicit arguments correspond to first occurrences of their predicates. [sent-116, score-0.497]
39 They refer to previous predicate instances assuming that the reader already recalls the involved participants. [sent-120, score-0.315]
40 That is, the filler of the different instances of a predicate argument maintain a certain discourse coherence. [sent-121, score-0.761]
41 For instance, in example (1), all the argument positions of the second occurrence of the 1182 predicate loss are missing, but they can be easily inferred from the previous instance of the same predicate. [sent-122, score-0.511]
42 Therefore, we propose to exploit this property in order to capture correctly how the fillers of all predicate arguments evolve through a document. [sent-125, score-0.612]
43 Our algorithm, ImpAr, processes the documents sentence by sentence, assuming that sequences of the same predicate (in its nominal or verbal form) share the same argument fillers (ex- plicit or implicit)3. [sent-126, score-0.725]
44 Thus, for every core argument argn of a predicate, ImpAr stores its previous known filler as a default value. [sent-127, score-0.564]
45 If the arguments of a predicate are explicit, they always replace default fillers previously captured. [sent-128, score-0.67]
46 When there is no antecedent for a particular implicit argument argn, the algorithm tries to find in the surrounding context which participant is the most likely to be the filler according to some salience factors (see Section 4. [sent-129, score-1.028]
47 For the following instances, without an explicit filler for a particular argument position, the algorithm repeats the same selection process and compares the new implicit candidate with the default one. [sent-131, score-1.036]
48 That is, the default implicit argument of a predicate with no antecedent can change every time the algorithm finds a filler with a greater salience. [sent-132, score-1.182]
49 A damping factor is applied to reduce the salience of distant predicates. [sent-133, score-0.394]
50 2 Filling arguments without explicit antecedents Filling the implicit arguments of a predicate has been identified as a particular case of corefer- ence, very close to pronoun resolution (Silberer and Frank, 2012). [sent-135, score-1.15]
51 Consequently, for those implicit arguments that have not explicit antecedents, we propose an adaptation of a classic algorithm for deterministic pronoun resolution. [sent-136, score-0.712]
52 When our algorithm needs to fill an implicit predicate argument without an explicit antecedent it considers a set of candidates within a window formed by the sentence of the predicate and the two previous sentences. [sent-138, score-1.326]
53 Apply two constraints to the candidate list: (a) All candidates that are already explicit arguments of the predicate are ruled out. [sent-141, score-0.671]
54 (b) All candidates commanded by the predicate in the dependency tree are ruled out. [sent-142, score-0.392]
55 Select those candidates that are semantically consistent with the semantic category of the implicit argument. [sent-144, score-0.434]
56 Sort the candidates by their proximity to the predicate of the implicit argument. [sent-148, score-0.655]
57 As a result, the candidate with the highest salience value is selected as the filler of the implicit argument. [sent-151, score-0.785]
58 Thus, this filler with its corresponding salience weight will be also considered in subsequent instances of the same predicate. [sent-152, score-0.45]
59 In this example, arg0 is missing for the predicate plan: (2) Quest Medical Inc said it adopted [arg1 a shareholders’ rights] [np plan] in which rights to purchase shares of common stock will be distributed as a dividend to shareholders of record as of Oct 23. [sent-154, score-0.597]
60 In the first step, the algorithm filters out the candidates that are actual explicit arguments of the predicate or have a syntactic dependency with the predicate, and therefore, they are in the search space of a traditional SRL system. [sent-156, score-0.628]
61 To determine the semantic coherence between the potential candidates and a predicate argument argn, we have exploited the selectional preferences in the same way as in previous SRL and implicit argument resolution works. [sent-159, score-1.218]
62 Second, we have semi-automatically assigned one of them to every predicate argument argn in PropBank and NomBank. [sent-161, score-0.545]
63 Then, we have acquired the most common categories of each predicate argument argn. [sent-165, score-0.477]
64 ImpAr algorithm also uses the SuperSenseTagger over the documents to be processed from BNB to check if the candidate belongs to the expected semantic category of the implicit argument to be filled. [sent-166, score-0.676]
65 As the semantic category for the implicit argument arg0 for the predicate plan has been recognized to be also COGNITIVE, [Quest Medical Inc] remains in the list of candidates as a possible filler. [sent-169, score-0.957]
66 In this process, the algorithm assigns to each candidate a set of salience factors that scores its prominence. [sent-208, score-0.285]
67 The subject, direct object, indirect object and nonadverbial factors weight the salience of the candidate depending on the syntactic role they belong to. [sent-210, score-0.287]
68 In the example, candidate [Quest Medical Inc] is in the same sentence as the predicate plan, it 4Lexicographic files according to WordNet terminology. [sent-213, score-0.308]
69 In order to reduce the impact of these errors we have included a damping factor that is applied sentence by sentence to the salience value of the default candidate. [sent-219, score-0.485]
70 It assumes that, independently of the initial salience assigned, 100 points of the salience score came from the sentence recency factor. [sent-221, score-0.444]
71 So, given a salience score s, the value of the score in a following sentence, s0, is: s0 = s 100 + 100 · r Obviously, the value of r must be defined without harming excessively those cases where the default candidate has been correctly identified. [sent-223, score-0.336]
72 For this, we studied in the training dataset the cases of implicit arguments filled with the default candidate. [sent-224, score-0.588]
73 Figure 1 shows that the influence of the default filler is much higher in near sentences that in more distance ones. [sent-225, score-0.289]
74 In this way, if the filler of the implicit argument is wrongly identified, the error only spreads to the nearest instances. [sent-228, score-0.747]
75 The distribution shown in figure 1 follows an exponential decay, therefore we have described the damping factor as a curve like the following, where α must be a value within 0 and 1: 1184 Figure 1: Distances between the implicit argument and the default candidate. [sent-230, score-0.827]
76 5d This value maintains the influence of the default fillers with high salience in near sentences. [sent-235, score-0.452]
77 At two sentence distance, the resulting score for the default filler is 260 − 100 + 100 · 0. [sent-246, score-0.289]
78 n cIen of the default filler of arg0 becomes smaller. [sent-249, score-0.289]
79 For every argument position in the goldstandard the scorer expects a single predicted constituent to fill in. [sent-251, score-0.303]
80 As shown above, previous works in implicit argument resolution proposed a metric that involves the correct identification of the whole span of the filler. [sent-262, score-0.64]
81 This dataset contains 437 predicate instances but just 246 argument positions are implicitly filled. [sent-267, score-0.522]
82 Interestingly, both systems present very different performances predicate by predicate. [sent-273, score-0.306]
83 Instead, the supervised approach would need a large amount of manual annotations for every predicate to be processed. [sent-456, score-0.322]
84 • Exp2: Only explicit fillers are considered as cEaxnpd2i:da Otensl6y. [sent-463, score-0.274]
85 Additionally, the highest loss appears when the default fillers are ruled out (Exp3). [sent-466, score-0.324]
86 As Exp1 also includes instances with explicit antecedents, and for these cases the damping factor component has no effect, we have designed two additional experiments: • • Exp4: Full system for the cases not solved by explicit aulnltse ycsedteemntfs o. [sent-515, score-0.472]
87 2 Correct span of the fillers As explained in Section 5, our algorithm works with syntactic dependencies and its predictions only return the head token of the filler. [sent-519, score-0.282]
88 In this work we have applied a simple heuristic that returns all the descendant 6That is, implicit arguments without explicit antecedents are not filled. [sent-521, score-0.669]
89 For example, from the following sentence our system gives the following prediction for an implicit arg1 of an instance of the predicate sale: Ports of Call Inc. [sent-524, score-0.612]
90 7 Conclusions and Future Work In this work we have presented a robust deterministic approach for implicit Semantic Role Labeling. [sent-530, score-0.397]
91 The method exploits a very simple but relevant discoursive coherence property that holds over explicit and implicit arguments of closely related nominal and verbal predicates. [sent-531, score-0.822]
92 This property states that if several instances of the same predicate appear in a well-written discourse, it is very likely that they maintain the same argument fillers. [sent-532, score-0.596]
93 We have shown the importance of this phenomenon for recovering the implicit information about semantic roles. [sent-533, score-0.391]
94 These annotations can be easily obtained from plain text using available tools7, what makes this algorithm the first effective tool available for implicit SRL. [sent-539, score-0.434]
95 1187 to test alternative approaches to solve the arguments without explicit antecedents. [sent-544, score-0.275]
96 Semantic role labeling of implicit arguments for nominal predicates. [sent-604, score-0.594]
97 Beyond nombank: a study of implicit arguments for nominal predicates. [sent-610, score-0.552]
98 Exploiting explicit annotations and semantic types for implicit argument resolution. [sent-653, score-0.77]
99 Predicate-specific annotations for implicit role binding: Corpus annotation, data analysis and evaluation experiments. [sent-683, score-0.436]
100 Casting implicit role linking as an anaphora resolution task. [sent-735, score-0.492]
wordName wordTfidf (topN-words)
[('impar', 0.423), ('implicit', 0.342), ('predicate', 0.27), ('gerber', 0.234), ('chai', 0.217), ('argument', 0.207), ('salience', 0.207), ('filler', 0.198), ('arguments', 0.155), ('fillers', 0.154), ('damping', 0.136), ('explicit', 0.12), ('nombank', 0.11), ('quest', 0.104), ('shareholders', 0.101), ('default', 0.091), ('srl', 0.087), ('inc', 0.086), ('rigau', 0.083), ('predicates', 0.076), ('bnb', 0.075), ('rights', 0.074), ('argn', 0.068), ('supersensetagger', 0.068), ('ruppenhofer', 0.066), ('laparra', 0.065), ('medical', 0.061), ('obtains', 0.061), ('delmonte', 0.06), ('resolution', 0.056), ('deterministic', 0.055), ('nominal', 0.055), ('head', 0.053), ('constituent', 0.052), ('annotations', 0.052), ('antecedents', 0.052), ('tonelli', 0.052), ('anaphora', 0.052), ('dividend', 0.051), ('gorinski', 0.051), ('leass', 0.051), ('oct', 0.051), ('factor', 0.051), ('arquez', 0.049), ('losses', 0.049), ('semantic', 0.049), ('coreference', 0.048), ('propbank', 0.048), ('plan', 0.046), ('usa', 0.046), ('buyers', 0.045), ('ehu', 0.045), ('bid', 0.045), ('dahl', 0.045), ('ruled', 0.045), ('silberer', 0.045), ('instances', 0.045), ('coherence', 0.044), ('scorer', 0.044), ('candidates', 0.043), ('role', 0.042), ('ciaramita', 0.042), ('maintain', 0.041), ('algorithm', 0.04), ('palmer', 0.04), ('altun', 0.039), ('lappin', 0.039), ('verbal', 0.039), ('meyers', 0.039), ('candidate', 0.038), ('missing', 0.038), ('performances', 0.036), ('potsdam', 0.036), ('span', 0.035), ('realizations', 0.034), ('purchase', 0.034), ('antecedent', 0.034), ('loss', 0.034), ('commanded', 0.034), ('discoursive', 0.034), ('moor', 0.034), ('venses', 0.034), ('whittemore', 0.034), ('surdeanu', 0.034), ('sporleder', 0.033), ('property', 0.033), ('california', 0.033), ('iwcs', 0.032), ('np', 0.03), ('recency', 0.03), ('investor', 0.03), ('widen', 0.03), ('stock', 0.029), ('gildea', 0.028), ('baker', 0.028), ('mentions', 0.028), ('framenet', 0.028), ('loan', 0.028), ('nominalizations', 0.028), ('ixa', 0.028)]
simIndex simValue paperId paperTitle
same-paper 1 0.99999952 189 acl-2013-ImpAr: A Deterministic Algorithm for Implicit Semantic Role Labelling
Author: Egoitz Laparra ; German Rigau
Abstract: This paper presents a novel deterministic algorithm for implicit Semantic Role Labeling. The system exploits a very simple but relevant discursive property, the argument coherence over different instances of a predicate. The algorithm solves the implicit arguments sequentially, exploiting not only explicit but also the implicit arguments previously solved. In addition, we empirically demonstrate that the algorithm obtains very competitive and robust performances with respect to supervised approaches that require large amounts of costly training data.
2 0.2127644 306 acl-2013-SPred: Large-scale Harvesting of Semantic Predicates
Author: Tiziano Flati ; Roberto Navigli
Abstract: We present SPred, a novel method for the creation of large repositories of semantic predicates. We start from existing collocations to form lexical predicates (e.g., break ∗) and learn the semantic classes that best f∗it) tahned ∗ argument. Taon idco this, we extract failtl thhee ∗ occurrences ion Wikipedia ewxthraiccht match the predicate and abstract its arguments to general semantic classes (e.g., break BODY PART, break AGREEMENT, etc.). Our experiments show that we are able to create a large collection of semantic predicates from the Oxford Advanced Learner’s Dictionary with high precision and recall, and perform well against the most similar approach.
3 0.18929875 314 acl-2013-Semantic Roles for String to Tree Machine Translation
Author: Marzieh Bazrafshan ; Daniel Gildea
Abstract: We experiment with adding semantic role information to a string-to-tree machine translation system based on the rule extraction procedure of Galley et al. (2004). We compare methods based on augmenting the set of nonterminals by adding semantic role labels, and altering the rule extraction process to produce a separate set of rules for each predicate that encompass its entire predicate-argument structure. Our results demonstrate that the second approach is effective in increasing the quality of translations.
4 0.18791287 229 acl-2013-Leveraging Synthetic Discourse Data via Multi-task Learning for Implicit Discourse Relation Recognition
Author: Man Lan ; Yu Xu ; Zhengyu Niu
Abstract: To overcome the shortage of labeled data for implicit discourse relation recognition, previous works attempted to automatically generate training data by removing explicit discourse connectives from sentences and then built models on these synthetic implicit examples. However, a previous study (Sporleder and Lascarides, 2008) showed that models trained on these synthetic data do not generalize very well to natural (i.e. genuine) implicit discourse data. In this work we revisit this issue and present a multi-task learning based system which can effectively use synthetic data for implicit discourse relation recognition. Results on PDTB data show that under the multi-task learning framework our models with the use of the prediction of explicit discourse connectives as auxiliary learning tasks, can achieve an averaged F1 improvement of 5.86% over baseline models.
5 0.16582637 98 acl-2013-Cross-lingual Transfer of Semantic Role Labeling Models
Author: Mikhail Kozhevnikov ; Ivan Titov
Abstract: Semantic Role Labeling (SRL) has become one of the standard tasks of natural language processing and proven useful as a source of information for a number of other applications. We address the problem of transferring an SRL model from one language to another using a shared feature representation. This approach is then evaluated on three language pairs, demonstrating competitive performance as compared to a state-of-the-art unsupervised SRL system and a cross-lingual annotation projection baseline. We also consider the contribution of different aspects of the feature representation to the performance of the model and discuss practical applicability of this method. 1 Background and Motivation Semantic role labeling has proven useful in many natural language processing tasks, such as question answering (Shen and Lapata, 2007; Kaisser and Webber, 2007), textual entailment (Sammons et al., 2009), machine translation (Wu and Fung, 2009; Liu and Gildea, 2010; Gao and Vogel, 2011) and dialogue systems (Basili et al., 2009; van der Plas et al., 2009). Multiple models have been designed to automatically predict semantic roles, and a considerable amount of data has been annotated to train these models, if only for a few more popular languages. As the annotation is costly, one would like to leverage existing resources to minimize the human effort required to construct a model for a new language. A number of approaches to the construction of semantic role labeling models for new languages have been proposed. On one end of the scale is unsupervised SRL, such as Grenager and Manning (2006), which requires some expert knowledge, but no labeled data. It clusters together arguments that should bear the same semantic role, but does not assign a particular role to each cluster. On the other end is annotating a new dataset from scratch. There are also intermediate options, which often make use of similarities between languages. This way, if an accurate model exists for one language, it should help simplify the construction of a model for another, related language. The approaches in this third group often use parallel data to bridge the gap between languages. Cross-lingual annotation projection systems (Pad o´ and Lapata, 2009), for example, propagate information directly via word alignment links. However, they are very sensitive to the quality of parallel data, as well as the accuracy of a sourcelanguage model on it. An alternative approach, known as cross-lingual model transfer, or cross-lingual model adaptation, consists of modifying a source-language model to make it directly applicable to a new language. This usually involves constructing a shared feature representation across the two languages. McDonald et al. (201 1) successfully apply this idea to the transfer of dependency parsers, using part-of- speech tags as the shared representation of words. A later extension of T ¨ackstr o¨m et al. (2012) enriches this representation with cross-lingual word clusters, considerably improving the performance. In the case of SRL, a shared representation that is purely syntactic is likely to be insufficient, since structures with different semantics may be realized by the same syntactic construct, for example “in August” vs “in Britain”. However with the help of recently introduced cross-lingual word represen1190 Proce dingsS o f ita h,e B 5u1lgsta Arinan,u Aaulg Musete 4ti-n9g 2 o0f1 t3h.e ? Ac s2s0o1ci3a Atiosnso fcoirat Cio nm foprut Caotimonpaulta Lti nognuails Lti cnsg,u piasgteics 1 90–120 , tations, such as the cross-lingual clustering mentioned above or cross-lingual distributed word representations of Klementiev et al. (2012), we may be able to transfer models of shallow semantics in a similar fashion. In this work we construct a shared feature representation for a pair of languages, employing crosslingual representations of syntactic and lexical information, train a semantic role labeling model on one language and apply it to the other one. This approach yields an SRL model for a new language at a very low cost, effectively requiring only a source language model and parallel data. We evaluate on five (directed) language pairs EN-ZH, ZH-EN, EN-CZ, CZ-EN and EN-FR, where EN, FR, CZ and ZH denote English, French, Czech and Chinese, respectively. The transferred model is compared against two baselines: an unsupervised SRL system and a model trained on the output of a cross-lingual annotation projection system. In the next section we will describe our setup, then in section 3 present the shared feature representation we use, discuss the evaluation data and other technical aspects in section 4, present the results and conclude with an overview of related work. – 2 Setup The purpose of the study is not to develop a yet another semantic role labeling system any existing SRL system can (after some modification) be used in this setup but to assess the practical applicability of cross-lingual model transfer to this – – problem, compare it against the alternatives and identify its strong/weak points depending on a particular setup. 2.1 Semantic Role Labeling Model We consider the dependency-based version of semantic role labeling as described in Haji cˇ et al. (2009) and transfer an SRL model from one language to another. We only consider verbal predicates and ignore the predicate disambiguation stage. We also assume that the predicate identification information is available in most languages it can be obtained using a relatively simple heuristic based on part-of-speech tags. The model performs argument identification and classification (Johansson and Nugues, 2008) separately in a pipeline first each candidate is classified as being or not being a head of an argument phrase with respect to the predicate in question and then each of the arguments is assigned a role from a given inventory. The model is factorized over arguments the decisions regarding the classification of different arguments are made in– – – dependently of each other. With respect to the use of syntactic annotation we consider two options: using an existing dependency parser for the target language and obtaining one by means of cross-lingual transfer (see section 4.2). Following McDonald et al. (201 1), we assume that a part-of-speech tagger is available for the target language. 2.2 SRL in the Low-resource Setting Several approaches have been proposed to obtain an SRL model for a new language with little or no manual annotation. Unsupervised SRL models (Lang and Lapata, 2010) cluster the arguments of predicates in a given corpus according to their semantic roles. The performance of such models can be impressive, especially for those languages where semantic roles correlate strongly with syntactic relation of the argument to its predicate. However, assigning meaningful role labels to the resulting clusters requires additional effort and the model’s parameters generally need some adjustment for every language. If the necessary resources are already available for a closely related language, they can be utilized to facilitate the construction of a model for the target language. This can be achieved either by means of cross-lingual annotation projection (Yarowsky et al., 2001) or by cross-lingual model transfer (Zeman and Resnik, 2008). This last approach is the one we are considering in this work, and the other two options are treated as baselines. The unsupervised model will be further referred to as UNSUP and the projection baseline as PROJ. 2.3 Evaluation Measures We use the F1 measure as a metric for the argument identification stage and accuracy as an aggregate measure of argument classification performance. When comparing to the unsupervised SRL system the clustering evaluation measures are used instead. These are purity and collocation 1191 N1Ximajx|Gj∩ Ci| CO =N1Xjmiax|Gj∩ Ci|, PU = where Ci is the set of arguments in the i-th induced cluster, Gj is the set of arguments in the jth gold cluster and N is the total number of arguments. We report the harmonic mean ofthe two (Lang and Lapata, 2011) and denote it F1c to avoid confusing it with the supervised metric. 3 Model Transfer The idea of this work is to abstract the model away from the particular source language and apply it to a new one. This setup requires that we use the same feature representation for both languages, for example part-of-speech tags and dependency relation labels should be from the same inventory. Some features are not applicable to certain lan- guages because the corresponding phenomena are absent in them. For example, consider a strongly inflected language and an analytic one. While the latter can usually convey the information encoded in the word form in the former one (number, gender, etc.), finding a shared feature representation for such information is non-trivial. In this study we will confine ourselves to those features that are applicable to all languages in question, namely: part-of-speech tags, syntactic dependency structures and representations of the word’s identity. 3.1 Lexical Information We train a model on one language and apply it to a different one. In order for this to work, the words of the two languages have to be mapped into a common feature space. It is also desirable that closely related words from both languages have similar representations in this space. Word mapping. The first option is simply to use the source language words as the shared representation. Here every source language word would have itself as its representation and every target word would map into a source word that corresponds to it. In other words, we supply the model with a gloss of the target sentence. The mapping (bilingual dictionary) we use is derived from a word-aligned parallel corpus, by identifying, for each word in the target language, the word in the source language it is most often aligned to. Cross-lingual clusters. There is no guarantee that each of the words in the evaluation data is present in our dictionary, nor that the corresponding source-language word is present in the training data, so the model would benefit from the ability to generalize over closely related words. This can, for example, be achieved by using cross-lingual word clusters induced in T ¨ackstr o¨m et al. (2012). We incorporate these clusters as features into our model. 3.2 Syntactic Information Part-of-speech Tags. We map part-of-speech tags into the universal tagset following Petrov et al. (2012). This may have a negative effect on the performance of a monolingual model, since most part-of-speech tagsets are more fine-grained than the universal POS tags considered here. For example Penn Treebank inventory contains 36 tags and the universal POS tagset only 12. Since the finergrained POS tags often reflect more languagespecific phenomena, however, they would only be useful for very closely related languages in the cross-lingual setting. The universal part-of-speech tags used in evaluation are derived from gold-standard annotation for all languages except French, where predicted ones had to be used instead. Dependency Structure. Another important aspect of syntactic information is the dependency structure. Most dependency relation inventories are language-specific, and finding a shared representation for them is a challenging problem. One could map dependency relations into a simplified form that would be shared between languages, as it is done for part-of-speech tags in Petrov et al. (2012). The extent to which this would be useful, however, depends on the similarity of syntactic-semantic in– terfaces of the languages in question. In this work we discard the dependency relation labels where the inventories do not match and only consider the unlabeled syntactic dependency graph. Some discrepancies, such as variations in attachment order, may be present even there, but this does not appear to be the case with the datasets we use for evaluation. If a target language is poor in resources, one can obtain a dependency parser for the target language by means of cross-lingual model transfer (Zeman and Resnik, 2008). We 1192 take this into account and evaluate both using the original dependency structures and the ones obtained by means of cross-lingual model transfer. 3.3 The Model The model we use is based on that of Bj ¨orkelund et al. (2009). It is comprised of a set of linear classifiers trained using Liblinear (Fan et al., 2008). The feature model was modified to accommodate the cross-lingual cluster features and the reranker component was not used. We do not model the interaction between different argument roles in the same predicate. While this has been found useful, in the cross-lingual setup one has to be careful with the assumptions made. For example, modeling the sequence of roles using a Markov chain (Thompson et al., 2003) may not work well in the present setting, especially between distant languages, as the order or arguments is not necessarily preserved. Most constraints that prove useful for SRL (Chang et al., 2007) also require customization when applied to a new language, and some rely on languagespecific resources, such as a valency lexicon. Taking into account the interaction between different arguments of a predicate is likely to improve the performance of the transferred model, but this is outside the scope of this work. 3.4 Feature Selection Compatibility of feature representations is necessary but not sufficient for successful model transfer. We have to make sure that the features we use are predictive of similar outcomes in the two languages as well. Depending on the pair of languages in question, different aspects of the feature representation will retain or lose their predictive power. We can be reasonably certain that the identity of an argument word is predictive of its semantic role in any language, but it might or might not be true of, for example, the word directly preceding the argument word. It is therefore important to pre- SCPDGylOespoSntreslTabunc1lra:obsFel-daitnguplrdoaeusntpagd-elronwfu-dcsopeyrnsd c.eylafguhtorsia mepgnrhs vent the model from capturing overly specific aspects of the source language, which we do by confining the model to first-order features. We also avoid feature selection, which, performed on the source language, is unlikely to help the model to better generalize to the target one. The experiments confirm that feature selection and the use of second-order features degrade the performance of the transferred model. 3.5 Feature Groups For each word, we use its part-of-speech tag, cross-lingual cluster id, word identity (glossed, when evaluating on the target language) and its dependency relation to its parent. Features associated with an argument word include the attributes of the predicate word, the argument word, its parent, siblings and children, and the words directly preceding and following it. Also included are the sequences of part-of-speech tags and dependency relations on the path between the predicate and the argument. Since we are also interested in the impact of different aspects of the feature representation, we divide the features into groups as summarized in table 1 and evaluate their respective contributions to the performance of the model. If a feature group is enabled the model has access to the corre– sponding source of information. For example, if only POS group is enabled, the model relies on the part-of-speech tags of the argument, the predicate and the words to the right and left of the argument word. If Synt is enabled too, it also uses the POS tags of the argument’s parent, children and siblings. Word order information constitutes an implicit group that is always available. It includes the Pos it ion feature, which indicates whether the argument is located to the left or to the right of the predicate, and allows the model to look up the attributes of the words directly preceding and following the argument word. The model we compare against the baselines uses all applicable feature groups (Deprel is only used in EN-CZ and CZ-EN experiments with original syntax). 4 Evaluation 4.1 Datasets and Preprocessing Evaluation of the cross-lingual model transfer requires a rather specific kind of dataset. Namely, the data in both languages has to be annotated 1193 with the same set of semantic roles following the same (or compatible) guidelines, which is seldom the case. We have identified three language pairs for which such resources are available: EnglishChinese, English-Czech and English-French. The evaluation datasets for English and Chinese are those from the CoNLL Shared Task 2009 (Haji ˇc et al., 2009) (henceforth CoNLL-ST). Their annotation in the CoNLL-ST is not identical, but the guidelines for “core” semantic roles are similar (Kingsbury et al., 2004), so we evaluate only on core roles here. The data for the second language pair is drawn from the Prague Czech-English Dependency Treebank 2.0 (Haji ˇc et al., 2012), which we converted to a format similar to that of CoNLL-ST1 . The original annotation uses the tectogrammatical representation (Haji ˇc, 2002) and an inventory of semantic roles (or functors), most of which are interpretable across various predicates. Also note that the syntactic anno- tation of English and Czech in PCEDT 2.0 is quite similar (to the extent permitted by the difference in the structure of the two languages) and we can use the dependency relations in our experiments. For English-French, the English CoNLL-ST dataset was used as a source and the model was evaluated on the manually annotated dataset from van der Plas et al. (201 1). The latter contains one thousand sentences from the French part ofthe Europarl (Koehn, 2005) corpus, annotated with semantic roles following an adapted version of PropBank (Palmer et al., 2005) guidelines. The authors perform annotation projection from English to French, using a joint model of syntax and semantics and employing heuristics for filtering. We use a model trained on the output of this projection system as one of the baselines. The evaluation dataset is relatively small in this case, so we perform the transfer only one-way, from English to French. The part-of-speech tags in all datasets were replaced with the universal POS tags of Petrov et al. (2012). For Czech, we have augmented the map- pings to account for the tags that were not present in the datasets from which the original mappings were derived. Namely, tag “t” is mapped to “VERB” and “Y” to “PRON”. We use parallel data to construct a bilingual dictionary used in word mapping, as well as in the projection baseline. For English-Czech – 1see http://www.ml4nlp.de/code-and-data/treex2conll and English-French, the data is drawn from Europarl (Koehn, 2005), for English-Chinese from MultiUN (Eisele and Chen, 2010). The word alignments were obtained using GIZA++ (Och and Ney, 2003) and the intersection heuristic. – 4.2 Syntactic Transfer In the low-resource setting, we cannot always rely on the availability of an accurate dependency parser for the target language. If one is not available, the natural solution would be to use crosslingual model transfer to obtain it. Unfortunately, the models presented in the previous work, such as Zeman and Resnik (2008), McDonald et al. (201 1) and T ¨ackstr o¨m et al. (2012), were not made available, so we reproduced the direct transfer algorithm of McDonald et al. (201 1), using Malt parser (Nivre, 2008) and the same set of features. We did not reimplement the projected transfer algorithm, however, and used the default training procedure instead of perceptron-based learning. The dependency structure thus obtained is, of course, only a rough approximation even a much more sophisticated algorithm may not perform well when transferring syntax between such languages as Czech and English, given the inherent difference in their structure. The scores are shown in table 2. We will henceforth refer to the syntactic annotations that were provided with the datasets as original, as opposed to the annotations obtained by means of syntactic transfer. – 4.3 Baselines Unsupervised Baseline: We are using a version of the unsupervised semantic role induction system of Titov and Klementiev (2012a) adapted to SetupUAS, % Table2:SyntaciE C ZcN HNt- rE ZaCFnN HZRsfer34 692567acuracy,unlabe dat- tachment score (percent). Note that in case of French we evaluate against the output of a supervised system, since manual annotation is not available for this dataset. This score does not reflect the true performance of syntactic transfer. 1194 the shared feature representation considered in order to make the scores comparable with those of the transfer model and, more importantly, to enable evaluation on transferred syntax. Note that the original system, tailored to a more expressive language-specific syntactic representation and equipped with heuristics to identify active/passive voice and other phenomena, achieves higher scores than those we report here. Projection Baseline: The projection baseline we use for English-Czech and English-Chinese is a straightforward one: we label the source side of a parallel corpus using the source-language model, then identify those verbs on the target side that are aligned to a predicate, mark them as predicates and propagate the argument roles in the same fashion. A model is then trained on the resulting training data and applied to the test set. For English-French we instead use the output of a fully featured projection model of van der Plas et al. (201 1), published in the CLASSiC project. 5 Results In order to ensure that the results are consistent, the test sets, except for the French one, were partitioned into five equal parts (of 5 to 10 thousand sentences each, depending on the dataset) and the evaluation performed separately on each one. All evaluation figures for English, Czech or Chinese below are the average values over the five subsets. In case of French, the evaluation dataset is too small to split it further, so instead we ran the evaluation five times on a randomly selected 80% sample of the evaluation data and averaged over those. In both cases the results are consistent over the subsets, the standard deviation does not exceed 0.5% for the transfer system and projection baseline and 1% for the unsupervised system. 5.1 Argument Identification We summarize the results in table 3. Argument identification is known to rely heavily on syntactic information, so it is unsurprising that it proves inaccurate when transferred syntax is used. Our simple projection baseline suffers from the same problem. Even with original syntactic information available, the performance of argument identification is moderate. Note that the model of (van der Plas et al., 2011), though relying on more expressive syntax, only outperforms the transferred system by 3% (F1) on this task. SetupSyntaxTRANSPROJ ZEC NH Z- EFCZNRHt r a n s 3462 1. 536 142 35. 4269 Table3EZ C:N H- CFEZANHZRrgumeon rt ig identf56 7ic13 a. t27903ion,21569t10ra. 3976nsferd model vs. projection baseline, F1. Most unsupervised SRL approaches assume that the argument identification is performed by some external means, for example heuristically (Lang and Lapata, 2011). Such heuristics or unsupervised approaches to argument identification (Abend et al., 2009) can also be used in the present setup. 5.2 Argument Classification In the following tables, TRANS column contains the results for the transferred system, UNSUP for the unsupervised baseline and PROJ for projection baseline. We highlight in bold the higher score where the difference exceeds twice the maximum of the standard deviation estimates of the two results. Table 4 presents the unsupervised evaluation results. Note that the unsupervised model performs as well as the transferred one or better where the – – SetupSyntaxTRANSUNSUP ZEC NH Z- EFCZNRHt r a n s 768 93648. 34627 6 5873. 1769 TableEZ C4NHZ:- FCEZANHZRrgumoe nr itg clasi78 fi94 3c. a25136tion,8 7 r9a4263n. 07 sferd model vs. unsupervised baseline in terms of the clustering metric F1c (see section 2.3). 1195 SetupSyntaxTRANSPROJ ZEC NH Z- EFCZNRHt r a n s 657 053. 1 36456419. 372 Table5EZ C:N H- CFEZANHZRrgumeon rt ig clasif657ic1936a. t170 ion,65 9t3804ra. 20847nsferd model vs. projection baseline, accuracy. original syntactic dependencies are available. In the more realistic scenario with transferred syn- tax, however, the transferred model proves more accurate. In table 5 we compare the transferred system with the projection baseline. It is easy to see that the scores vary strongly depending on the language pair, due to both the difference in the annotation scheme used and the degree of relatedness between the languages. The drop in performance when transferring the model to another language is large in every case, though, see table 6. SetupTargetSource Table6:MoCEZdHeNZ l- FECaZNRcH urac67 y53169o. 017nthes87 o25670u. r1245ceandtrge language using original syntax. The source language scores for English vary between language pairs because of the difference in syntactic annotation and role subset used. We also include the individual F1 scores for the top-10 most frequent labels for EN-CZ transfer with original syntax in table 7. The model provides meaningful predictions here, despite low overall accuracy. Most of the labels2 are self-explanatory: Patient (PAT), Actor (ACT), Time (TWHEN), Effect (EFF), Location (LOC), Manner (MANN), Addressee (ADDR), Extent (EXT). CPHR marks the 2http://ufal.mff.cuni.cz/∼toman/pcedt/en/functors.html LabelFreq.F1Re.Pr. recall and precision for the top-10 most frequent roles. nominal part of a complex predicate, as in “to have [a plan]CPHR”, and DIR3 indicates destination. 5.3 Additional Experiments We now evaluate the contribution of different aspects of the feature representation to the performance of the model. Table 8 contains the results for English-French. FeaturesOrigTrans ferent feature subsets, using original and transferred syntactic information. The fact that the model performs slightly better with transferred syntax may be explained by two factors. Firstly, as we already mentioned, the original syntactic annotation is also produced automatically. Secondly, in the model transfer setup it is more important how closely the syntacticsemantic interface on the target side resembles that on the source side than how well it matches the “true” structure of the target language, and in this respect a transferred dependency parser may have an advantage over one trained on target-language data. The high impact of the Glos s features here 1196 may be partly attributed to the fact that the mapping is derived from the same corpus as the evaluation data Europarl (Koehn, 2005) and partly by the similarity between English and French in terms of word order, usage of articles and prepositions. The moderate contribution of the crosslingual cluster features are likely due to the insufficient granularity of the clustering for this task. For more distant language pairs, the contributions of individual feature groups are less interpretable, so we only highlight a few observations. First of all, both EN-CZ and CZ-EN benefit noticeably from the use of the original syntactic annotation, including dependency relations, but not from the transferred syntax, most likely due to the low syntactic transfer performance. Both perform better when lexical information is available, although – – the improvement is not as significant as in the case of French only up to 5%. The situation with Chinese is somewhat complicated in that adding lexical information here fails to yield an improvement in terms of the metric considered. This is likely due to the fact that we consider only the core roles, which can usually be predicted with high accuracy based on syntactic information alone. – 6 Related Work Development of robust statistical models for core NLP tasks is a challenging problem, and adaptation of existing models to new languages presents a viable alternative to exhaustive annotation for each language. Although the models thus obtained are generally imperfect, they can be further refined for a particular language and domain using techniques such as active learning (Settles, 2010; Chen et al., 2011). Cross-lingual annotation projection (Yarowsky et al., 2001) approaches have been applied ex- tensively to a variety of tasks, including POS tagging (Xi and Hwa, 2005; Das and Petrov, 2011), morphology segmentation (Snyder and Barzilay, 2008), verb classification (Merlo et al., 2002), mention detection (Zitouni and Florian, 2008), LFG parsing (Wr o´blewska and Frank, 2009), information extraction (Kim et al., 2010), SRL (Pad o´ and Lapata, 2009; van der Plas et al., 2011; Annesi and Basili, 2010; Tonelli and Pianta, 2008), dependency parsing (Naseem et al., 2012; Ganchev et al., 2009; Smith and Eisner, 2009; Hwa et al., 2005) or temporal relation prediction (Spreyer and Frank, 2008). Interestingly, it has also been used to propagate morphosyntactic information between old and modern versions of the same language (Meyer, 2011). Cross-lingual model transfer methods (McDonald et al., 2011; Zeman and Resnik, 2008; Durrett et al., 2012; Søgaard, 2011; Lopez et al., 2008) have also been receiving much attention recently. The basic idea behind model transfer is similar to that of cross-lingual annotation projection, as we can see from the way parallel data is used in, for example, McDonald et al. (201 1). A crucial component of direct transfer approaches is the unified feature representation. There are at least two such representations of lexical information (Klementiev et al., 2012; T ¨ackstr o¨m et al., 2012), but both work on word level. This makes it hard to account for phenomena that are expressed differently in the languages considered, for example the syntactic function of a certain word may be indicated by a preposition, inflection or word order, depending on the language. Accurate representation of such information would require an extra level of abstraction (Haji ˇc, 2002). A side-effect ofusing adaptation methods is that we are forced to use the same annotation scheme for the task in question (SRL, in our case), which in turn simplifies the development of cross-lingual tools for downstream tasks. Such representations are also likely to be useful in machine translation. Unsupervised semantic role labeling methods (Lang and Lapata, 2010; Lang and Lapata, 2011; Titov and Klementiev, 2012a; Lorenzo and Cerisara, 2012) also constitute an alternative to cross-lingual model transfer. For an overview of of semi-supervised approaches we refer the reader to Titov and Klementiev (2012b). 7 Conclusion We have considered the cross-lingual model transfer approach as applied to the task of semantic role labeling and observed that for closely related languages it performs comparably to annotation projection approaches. It allows one to quickly construct an SRL model for a new language without manual annotation or language-specific heuristics, provided an accurate model is available for one of the related languages along with a certain amount of parallel data for the two languages. While an1197 notation projection approaches require sentenceand word-aligned parallel data and crucially depend on the accuracy of the syntactic parsing and SRL on the source side of the parallel corpus, cross-lingual model transfer can be performed using only a bilingual dictionary. Unsupervised SRL approaches have their advantages, in particular when no annotated data is available for any of the related languages and there is a syntactic parser available for the target one, but the annotation they produce is not always sufficient. In applications such as Information Retrieval it is preferable to have precise labels, rather than just clusters of arguments, for example. Also note that when applying cross-lingual model transfer in practice, one can improve upon the performance of the simplistic model we use for evaluation, for example by picking the features manually, taking into account the properties of the target language. Domain adaptation techniques can also be employed to adjust the model to the target language. Acknowledgments The authors would like to thank Alexandre Klementiev and Ryan McDonald for useful suggestions and T ¨ackstr o¨m et al. (2012) for sharing the cross-lingual word representations. This research is supported by the MMCI Cluster of Excellence. References Omri Abend, Roi Reichart, and Ari Rappoport. 2009. Unsupervised argument identification for semantic role labeling. In Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP, ACL ’09, pages 28–36, Stroudsburg, PA, USA. Association for Computational Linguistics. Paolo Annesi and Roberto Basili. 2010. Cross-lingual alignment of FrameNet annotations through hidden Markov models. In Proceedings of the 11th international conference on Computational Linguistics and Intelligent Text Processing, CICLing’ 10, pages 12– 25, Berlin, Heidelberg. Springer-Verlag. Roberto Basili, Diego De Cao, Danilo Croce, Bonaventura Coppola, and Alessandro Moschitti. 2009. Cross-language frame semantics transfer in bilingual corpora. In Alexander F. Gelbukh, editor, Proceedings of the 10th International Conference on Computational Linguistics and Intelligent Text Pro- cessing, pages 332–345. Anders Bj ¨orkelund, Love Hafdell, and Pierre Nugues. 2009. Multilingual semantic role labeling. In Proceedings of the Thirteenth Conference on Computational Natural Language Learning (CoNLL 2009): Shared Task, pages 43–48, Boulder, Colorado, June. Association for Computational Linguistics. Ming-Wei Chang, Lev Ratinov, and Dan Roth. 2007. Guiding semi-supervision with constraint-driven learning. In ACL. Chenhua Chen, Alexis Palmer, and Caroline Sporleder. 2011. Enhancing active learning for semantic role labeling via compressed dependency trees. In Proceedings of 5th International Joint Conference on Natural Language Processing, pages 183–191, Chiang Mai, Thailand, November. Asian Federation of Natural Language Processing. Dipanjan Das and Slav Petrov. 2011. Unsupervised part-of-speech tagging with bilingual graph-based projections. Proceedings of the Association for Computational Linguistics. Greg Durrett, Adam Pauls, and Dan Klein. 2012. Syntactic transfer using a bilingual lexicon. In Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, pages 1–1 1, Jeju Island, Korea, July. Association for Computational Linguistics. Andreas Eisele and Yu Chen. 2010. MultiUN: A multilingual corpus from United Nation documents. In Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC’10). European Language Resources Association (ELRA). Rong-En Fan, Kai-Wei Chang, Cho-Jui Hsieh, XiangRui Wang, and Chih-Jen Lin. 2008. LIBLINEAR: A library for large linear classification. Journal of Machine Learning Research, 9: 1871–1874. Kuzman Ganchev, Jennifer Gillenwater, and Ben Taskar. 2009. Dependency grammar induction via bitext projection constraints. In Proceedings of the 47th Annual Meeting of the ACL, pages 369–377, Stroudsburg, PA, USA. Association for Computational Linguistics. Qin Gao and Stephan Vogel. 2011. Corpus expansion for statistical machine translation with semantic role label substitution rules. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, pages 294–298, Portland, Oregon, USA. Trond Grenager and Christopher D. Manning. 2006. Unsupervised discovery of a statistical verb lexicon. In Proceedings of EMNLP. Jan Haji cˇ. 2002. Tectogrammatical representation: Towards a minimal transfer in machine translation. In Robert Frank, editor, Proceedings of the 6th International Workshop on Tree Adjoining Grammars 1198 and Related Frameworks (TAG+6), pages 216— 226, Venezia. Universita di Venezia. Jan Haji cˇ, Massimiliano Ciaramita, Richard Johansson, Daisuke Kawahara, Maria Ant o`nia Mart ı´, Llu ı´s M `arquez, Adam Meyers, Joakim Nivre, Sebastian Pad o´, Jan Sˇt eˇp a´nek, Pavel Stra nˇ a´k, Mihai Surdeanu, Nianwen Xue, and Yi Zhang. 2009. The CoNLL2009 shared task: Syntactic and semantic dependencies in multiple languages. In Proceedings of the Thirteenth Conference on Computational Natural Language Learning (CoNLL 2009): Shared Task, pages 1–18, Boulder, Colorado. Jan Haji cˇ, Eva Haji cˇov a´, Jarmila Panevov a´, Petr Sgall, Ond ˇrej Bojar, Silvie Cinkov´ a, Eva Fuˇ c ´ıkov a´, Marie Mikulov a´, Petr Pajas, Jan Popelka, Ji ˇr´ ı Semeck´ y, Jana Sˇindlerov a´, Jan Sˇt eˇp a´nek, Josef Toman, Zde nˇka Ure sˇov a´, and Zden eˇk Zˇabokrtsk y´. 2012. Announcing Prague Czech-English dependency treebank 2.0. In Nicoletta Calzolari (Conference Chair), Khalid Choukri, Thierry Declerck, Mehmet U gˇur Doˇ gan, Bente Maegaard, Joseph Mariani, Jan Odijk, and Stelios Piperidis, editors, Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC’12), Istanbul, Turkey, May. European Language Resources Association (ELRA). Rebecca Hwa, Philip Resnik, Amy Weinberg, Clara Cabezas, and Okan Kolak. 2005. Bootstrapping parsers via syntactic projection across parallel text. Natural Language Engineering, 11(3):3 11–325. Richard Johansson and Pierre Nugues. 2008. Dependency-based semantic role labeling of PropBank. In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing, pages 69–78, Honolulu, Hawaii. Michael Kaisser and Bonnie Webber. 2007. Question answering based on semantic roles. In ACL Workshop on Deep Linguistic Processing. Seokhwan Kim, Minwoo Jeong, Jonghoon Lee, and Gary Geunbae Lee. 2010. A cross-lingual annotation projection approach for relation detection. In Proceedings of the 23rd International Conference on Computational Linguistics, COLING ’ 10, pages 564–571, Stroudsburg, PA, USA. Association for Computational Linguistics. Paul Kingsbury, Nianwen Xue, and Martha Palmer. 2004. Propbanking in parallel. In In Proceedings of the Workshop on the Amazing Utility of Parallel and Comparable Corpora, in conjunction with LREC’04. Alexandre Klementiev, Ivan Titov, and Binod Bhattarai. 2012. Inducing crosslingual distributed representations of words. In Proceedings of the International Conference on Computational Linguistics (COLING), Bombay, India. Philipp Koehn. 2005. Europarl: A parallel corpus for statistical machine translation. In Conference Proceedings: the tenth Machine Translation Summit, pages 79–86, Phuket, Thailand. AAMT. Joel Lang and Mirella Lapata. 2010. Unsupervised induction of semantic roles. In Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 939–947, Los Angeles, California, June. Association for Computational Linguistics. Joel Lang and Mirella Lapata. 2011. Unsupervised semantic role induction via split-merge clustering. In Proc. of Annual Meeting of the Association for Computational Linguistics (ACL). Ding Liu and Daniel Gildea. 2010. Semantic role features for machine translation. In Proceedings of the 23rd International Conference on Computational Linguistics (Coling 2010), Beijing, China. Adam Lopez, Daniel Zeman, Michael Nossal, Philip Resnik, and Rebecca Hwa. 2008. Cross-language parser adaptation between related languages. In IJCNLP-08 Workshop on NLP for Less Privileged Languages, pages 35–42, Hyderabad, India, January. Alejandra Lorenzo and Christophe Cerisara. 2012. Unsupervised frame based semantic role induction: application to French and English. In Proceedings of the ACL 2012 Joint Workshop on Statistical Parsing and Semantic Processing of Morphologically Rich Languages, pages 30–35, Jeju, Republic of Korea, July. Association for Computational Linguistics. Ryan McDonald, Slav Petrov, and Keith Hall. 2011. Multi-source transfer of delexicalized dependency parsers. In Proceedings of the Conference on Empirical Methods in Natural Language Processing, EMNLP ’ 11, pages 62–72, Stroudsburg, PA, USA. Association for Computational Linguistics. Paola Merlo, Suzanne Stevenson, Vivian Tsang, and Gianluca Allaria. 2002. A multi-lingual paradigm for automatic verb classification. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL’02), pages 207– 214, Philadelphia, PA. Roland Meyer. 2011. New wine in old wineskins?– Tagging old Russian via annotation projection from modern translations. Russian Linguistics, 35(2):267(15). Tahira Naseem, Regina Barzilay, and Amir Globerson. 2012. Selective sharing for multilingual dependency parsing. In Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics, pages 629–637, Jeju Island, Korea, July. Association for Computational Linguistics. Joakim Nivre. 2008. Algorithms for deterministic incremental dependency parsing. Comput. Linguist., 34(4):513–553, December. 1199 Franz Josef Och and Hermann Ney. 2003. A systematic comparison of various statistical alignment models. Computational Linguistics, 29(1). Sebastian Pad o´ and Mirella Lapata. 2009. Crosslingual annotation projection for semantic roles. Journal of Artificial Intelligence Research, 36:307– 340. Martha Palmer, Daniel Gildea, and Paul Kingsbury. 2005. The Proposition Bank: An annotated corpus of semantic roles. Computational Linguistics, 31:71–105. Slav Petrov, Dipanjan Das, and Ryan McDonald. 2012. A universal part-of-speech tagset. In Proceedings of LREC, May. Mark Sammons, Vinod Vydiswaran, Tim Vieira, Nikhil Johri, Ming wei Chang, Dan Goldwasser, Vivek Srikumar, Gourab Kundu, Yuancheng Tu, Kevin Small, Joshua Rule, Quang Do, and Dan Roth. 2009. Relation alignment for textual entailment recognition. In Text Analysis Conference (TAC). Burr Settles. 2010. Active learning literature survey. Computer Sciences Technical Report, 1648. Dan Shen and Mirella Lapata. 2007. Using semantic roles to improve question answering. In EMNLP. David A Smith and Jason Eisner. 2009. Parser adaptation and projection with quasi-synchronous grammar features. In Proceedings of the 2009 Confer- ence on Empirical Methods in Natural Language Processing, pages 822–831. Association for Computational Linguistics. Benjamin Snyder and Regina Barzilay. 2008. Crosslingual propagation for morphological analysis. In Proceedings of the 23rd national conference on Artificial intelligence. Anders Søgaard. 2011. Data point selection for crosslanguage adaptation of dependency parsers. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, volume 2 of HLT ’11, pages 682–686, Stroudsburg, PA, USA. Association for Computational Linguistics. Kathrin Spreyer and Anette Frank. 2008. Projectionbased acquisition of a temporal labeller. Proceedings of IJCNLP 2008. Oscar T¨ ackstr o¨m, Ryan McDonald, and Jakob Uszkoreit. 2012. Cross-lingual word clusters for direct transfer of linguistic structure. In Proc. of the Annual Meeting of the North American Association of Computational Linguistics (NAACL), pages 477– 487, Montr ´eal, Canada. Cynthia A. Thompson, Roger Levy, and Christopher D. Manning. 2003. A generative model for seman- tic role labeling. In Proceedings of the 14th European Conference on Machine Learning, ECML 2003, pages 397–408, Dubrovnik, Croatia. Ivan Titov and Alexandre Klementiev. 2012a. A Bayesian approach to unsupervised semantic role induction. In Proc. of European Chapter of the Association for Computational Linguistics (EACL). Ivan Titov and Alexandre Klementiev. 2012b. Semisupervised semantic role labeling: Approaching from an unsupervised perspective. In Proceedings of the International Conference on Computational Linguistics (COLING), Bombay, India, December. Sara Tonelli and Emanuele Pianta. 2008. Frame information transfer from English to Italian. In Proceedings of LREC 2008. Lonneke van der Plas, James Henderson, and Paola Merlo. 2009. Domain adaptation with artificial data for semantic parsing of speech. In Proc. 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 125–128, Boulder, Colorado. Lonneke van der Plas, Paola Merlo, and James Henderson. 2011. Scaling up automatic cross-lingual semantic role annotation. In Proceedings of the 49th Annual Meeting of the Association for Computa- tional Linguistics: Human Language Technologies, HLT ’ 11, pages 299–304, Stroudsburg, PA, USA. Association for Computational Linguistics. Alina Wr o´blewska and Anette Frank. 2009. Crosslingual projection of LFG F-structures: Building an F-structure bank for Polish. In Eighth International Workshop on Treebanks and Linguistic Theories, page 209. Dekai Wu and Pascale Fung. 2009. Can semantic role labeling improve SMT? In Proceedings of 13th Annual Conference of the European Association for Machine Translation (EAMT 2009), Barcelona. Chenhai Xi and Rebecca Hwa. 2005. A backoff model for bootstrapping resources for non-English languages. In Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing, pages 85 1–858, Stroudsburg, PA, USA. David Yarowsky, Grace Ngai, and Ricahrd Wicentowski. 2001. Inducing multilingual text analysis tools via robust projection across aligned corpora. In Proceedings of Human Language Technology Conference. Daniel Zeman and Philip Resnik. 2008. Crosslanguage parser adaptation between related lan- guages. In Proceedings of the IJCNLP-08 Workshop on NLP for Less Privileged Languages, pages 35– 42, Hyderabad, India, January. Asian Federation of Natural Language Processing. Imed Zitouni and Radu Florian. 2008. Mention detection crossing the language barrier. In Proceedings of the Conference on Empirical Methods in Natural Language Processing. 1200
6 0.1557761 267 acl-2013-PARMA: A Predicate Argument Aligner
7 0.14671804 27 acl-2013-A Two Level Model for Context Sensitive Inference Rules
8 0.13891743 56 acl-2013-Argument Inference from Relevant Event Mentions in Chinese Argument Extraction
9 0.13767877 41 acl-2013-Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
10 0.13003753 17 acl-2013-A Random Walk Approach to Selectional Preferences Based on Preference Ranking and Propagation
11 0.12758645 376 acl-2013-Using Lexical Expansion to Learn Inference Rules from Sparse Data
12 0.11526078 206 acl-2013-Joint Event Extraction via Structured Prediction with Global Features
13 0.089747973 280 acl-2013-Plurality, Negation, and Quantification:Towards Comprehensive Quantifier Scope Disambiguation
14 0.089601219 252 acl-2013-Multigraph Clustering for Unsupervised Coreference Resolution
15 0.086213753 57 acl-2013-Arguments and Modifiers from the Learner's Perspective
16 0.077025093 207 acl-2013-Joint Inference for Fine-grained Opinion Extraction
17 0.068893246 130 acl-2013-Domain-Specific Coreference Resolution with Lexicalized Features
18 0.066094145 245 acl-2013-Modeling Human Inference Process for Textual Entailment Recognition
19 0.064859226 144 acl-2013-Explicit and Implicit Syntactic Features for Text Classification
20 0.064714827 289 acl-2013-QuEst - A translation quality estimation framework
topicId topicWeight
[(0, 0.175), (1, 0.037), (2, -0.057), (3, -0.124), (4, -0.09), (5, 0.15), (6, -0.049), (7, 0.125), (8, -0.016), (9, 0.093), (10, 0.065), (11, 0.047), (12, -0.014), (13, 0.104), (14, -0.006), (15, -0.071), (16, 0.082), (17, -0.018), (18, 0.12), (19, -0.018), (20, 0.044), (21, 0.102), (22, -0.126), (23, -0.042), (24, 0.063), (25, 0.037), (26, 0.104), (27, 0.008), (28, 0.152), (29, -0.01), (30, -0.053), (31, -0.07), (32, 0.077), (33, 0.095), (34, 0.041), (35, -0.041), (36, -0.019), (37, -0.055), (38, 0.016), (39, 0.106), (40, -0.049), (41, -0.055), (42, 0.03), (43, 0.013), (44, 0.048), (45, 0.015), (46, 0.083), (47, -0.061), (48, 0.049), (49, -0.103)]
simIndex simValue paperId paperTitle
same-paper 1 0.95748943 189 acl-2013-ImpAr: A Deterministic Algorithm for Implicit Semantic Role Labelling
Author: Egoitz Laparra ; German Rigau
Abstract: This paper presents a novel deterministic algorithm for implicit Semantic Role Labeling. The system exploits a very simple but relevant discursive property, the argument coherence over different instances of a predicate. The algorithm solves the implicit arguments sequentially, exploiting not only explicit but also the implicit arguments previously solved. In addition, we empirically demonstrate that the algorithm obtains very competitive and robust performances with respect to supervised approaches that require large amounts of costly training data.
2 0.70241797 306 acl-2013-SPred: Large-scale Harvesting of Semantic Predicates
Author: Tiziano Flati ; Roberto Navigli
Abstract: We present SPred, a novel method for the creation of large repositories of semantic predicates. We start from existing collocations to form lexical predicates (e.g., break ∗) and learn the semantic classes that best f∗it) tahned ∗ argument. Taon idco this, we extract failtl thhee ∗ occurrences ion Wikipedia ewxthraiccht match the predicate and abstract its arguments to general semantic classes (e.g., break BODY PART, break AGREEMENT, etc.). Our experiments show that we are able to create a large collection of semantic predicates from the Oxford Advanced Learner’s Dictionary with high precision and recall, and perform well against the most similar approach.
3 0.70193809 17 acl-2013-A Random Walk Approach to Selectional Preferences Based on Preference Ranking and Propagation
Author: Zhenhua Tian ; Hengheng Xiang ; Ziqi Liu ; Qinghua Zheng
Abstract: This paper presents an unsupervised random walk approach to alleviate data sparsity for selectional preferences. Based on the measure of preferences between predicates and arguments, the model aggregates all the transitions from a given predicate to its nearby predicates, and propagates their argument preferences as the given predicate’s smoothed preferences. Experimental results show that this approach outperforms several state-of-the-art methods on the pseudo-disambiguation task, and it better correlates with human plausibility judgements.
4 0.61649221 267 acl-2013-PARMA: A Predicate Argument Aligner
Author: Travis Wolfe ; Benjamin Van Durme ; Mark Dredze ; Nicholas Andrews ; Charley Beller ; Chris Callison-Burch ; Jay DeYoung ; Justin Snyder ; Jonathan Weese ; Tan Xu ; Xuchen Yao
Abstract: We introduce PARMA, a system for crossdocument, semantic predicate and argument alignment. Our system combines a number of linguistic resources familiar to researchers in areas such as recognizing textual entailment and question answering, integrating them into a simple discriminative model. PARMA achieves state of the art results on an existing and a new dataset. We suggest that previous efforts have focussed on data that is biased and too easy, and we provide a more difficult dataset based on translation data with a low baseline which we beat by 17% F1.
5 0.57832825 314 acl-2013-Semantic Roles for String to Tree Machine Translation
Author: Marzieh Bazrafshan ; Daniel Gildea
Abstract: We experiment with adding semantic role information to a string-to-tree machine translation system based on the rule extraction procedure of Galley et al. (2004). We compare methods based on augmenting the set of nonterminals by adding semantic role labels, and altering the rule extraction process to produce a separate set of rules for each predicate that encompass its entire predicate-argument structure. Our results demonstrate that the second approach is effective in increasing the quality of translations.
6 0.54791921 376 acl-2013-Using Lexical Expansion to Learn Inference Rules from Sparse Data
7 0.53115553 344 acl-2013-The Effects of Lexical Resource Quality on Preference Violation Detection
8 0.52547032 41 acl-2013-Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation
9 0.51052058 229 acl-2013-Leveraging Synthetic Discourse Data via Multi-task Learning for Implicit Discourse Relation Recognition
10 0.49950835 98 acl-2013-Cross-lingual Transfer of Semantic Role Labeling Models
11 0.49679124 280 acl-2013-Plurality, Negation, and Quantification:Towards Comprehensive Quantifier Scope Disambiguation
12 0.48152637 27 acl-2013-A Two Level Model for Context Sensitive Inference Rules
13 0.47773281 367 acl-2013-Universal Conceptual Cognitive Annotation (UCCA)
14 0.46899691 228 acl-2013-Leveraging Domain-Independent Information in Semantic Parsing
15 0.42435238 56 acl-2013-Argument Inference from Relevant Event Mentions in Chinese Argument Extraction
16 0.39382809 184 acl-2013-Identification of Speakers in Novels
17 0.38878596 206 acl-2013-Joint Event Extraction via Structured Prediction with Global Features
18 0.38843295 85 acl-2013-Combining Intra- and Multi-sentential Rhetorical Parsing for Document-level Discourse Analysis
19 0.37737778 324 acl-2013-Smatch: an Evaluation Metric for Semantic Feature Structures
20 0.37498268 192 acl-2013-Improved Lexical Acquisition through DPP-based Verb Clustering
topicId topicWeight
[(0, 0.049), (6, 0.035), (11, 0.055), (14, 0.016), (15, 0.016), (24, 0.071), (26, 0.031), (35, 0.057), (42, 0.074), (48, 0.07), (64, 0.01), (70, 0.035), (71, 0.014), (88, 0.045), (90, 0.028), (95, 0.066), (99, 0.249)]
simIndex simValue paperId paperTitle
1 0.81923914 375 acl-2013-Using Integer Linear Programming in Concept-to-Text Generation to Produce More Compact Texts
Author: Gerasimos Lampouras ; Ion Androutsopoulos
Abstract: We present an ILP model of concept-totext generation. Unlike pipeline architectures, our model jointly considers the choices in content selection, lexicalization, and aggregation to avoid greedy decisions and produce more compact texts.
same-paper 2 0.80085206 189 acl-2013-ImpAr: A Deterministic Algorithm for Implicit Semantic Role Labelling
Author: Egoitz Laparra ; German Rigau
Abstract: This paper presents a novel deterministic algorithm for implicit Semantic Role Labeling. The system exploits a very simple but relevant discursive property, the argument coherence over different instances of a predicate. The algorithm solves the implicit arguments sequentially, exploiting not only explicit but also the implicit arguments previously solved. In addition, we empirically demonstrate that the algorithm obtains very competitive and robust performances with respect to supervised approaches that require large amounts of costly training data.
Author: Trevor Cohn ; Lucia Specia
Abstract: Annotating linguistic data is often a complex, time consuming and expensive endeavour. Even with strict annotation guidelines, human subjects often deviate in their analyses, each bringing different biases, interpretations of the task and levels of consistency. We present novel techniques for learning from the outputs of multiple annotators while accounting for annotator specific behaviour. These techniques use multi-task Gaussian Processes to learn jointly a series of annotator and metadata specific models, while explicitly representing correlations between models which can be learned directly from data. Our experiments on two machine translation quality estimation datasets show uniform significant accuracy gains from multi-task learning, and consistently outperform strong baselines.
4 0.66495776 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 0.59069026 83 acl-2013-Collective Annotation of Linguistic Resources: Basic Principles and a Formal Model
Author: Ulle Endriss ; Raquel Fernandez
Abstract: Crowdsourcing, which offers new ways of cheaply and quickly gathering large amounts of information contributed by volunteers online, has revolutionised the collection of labelled data. Yet, to create annotated linguistic resources from this data, we face the challenge of having to combine the judgements of a potentially large group of annotators. In this paper we investigate how to aggregate individual annotations into a single collective annotation, taking inspiration from the field of social choice theory. We formulate a general formal model for collective annotation and propose several aggregation methods that go beyond the commonly used majority rule. We test some of our methods on data from a crowdsourcing experiment on textual entailment annotation.
6 0.56664425 82 acl-2013-Co-regularizing character-based and word-based models for semi-supervised Chinese word segmentation
7 0.56518239 123 acl-2013-Discriminative Learning with Natural Annotations: Word Segmentation as a Case Study
8 0.56192219 185 acl-2013-Identifying Bad Semantic Neighbors for Improving Distributional Thesauri
9 0.56124222 62 acl-2013-Automatic Term Ambiguity Detection
10 0.55971569 47 acl-2013-An Information Theoretic Approach to Bilingual Word Clustering
11 0.55780995 389 acl-2013-Word Association Profiles and their Use for Automated Scoring of Essays
12 0.557374 164 acl-2013-FudanNLP: A Toolkit for Chinese Natural Language Processing
13 0.55727381 56 acl-2013-Argument Inference from Relevant Event Mentions in Chinese Argument Extraction
14 0.55701602 225 acl-2013-Learning to Order Natural Language Texts
15 0.55625707 223 acl-2013-Learning a Phrase-based Translation Model from Monolingual Data with Application to Domain Adaptation
16 0.55526 172 acl-2013-Graph-based Local Coherence Modeling
17 0.55522537 267 acl-2013-PARMA: A Predicate Argument Aligner
18 0.55460525 68 acl-2013-Bilingual Data Cleaning for SMT using Graph-based Random Walk
19 0.55396599 99 acl-2013-Crowd Prefers the Middle Path: A New IAA Metric for Crowdsourcing Reveals Turker Biases in Query Segmentation
20 0.55351102 155 acl-2013-Fast and Accurate Shift-Reduce Constituent Parsing