id
stringlengths 7
12
| sentence1
stringlengths 6
1.27k
| sentence2
stringlengths 6
926
| label
stringclasses 4
values |
|---|---|---|---|
train_1100 | In another strand of work, syntactic annotations are assumed on both sides of the parallel data, and a model is trained to exploit the parallel data at test time as well (Smith and Smith, 2004;Burkett and Klein, 2008). | to this work, our goal is to explore the benefits of multilingual grammar induction in a fully unsupervised setting. | contrasting |
train_1101 | We could rewrite this by explicitly integrating over the yield, context, coupling, Giza-score parameters as well as the alignment trees. | since maximizing this integral directly would be intractable, we resort to standard Markov chain sampling techniques. | contrasting |
train_1102 | These marginals can be efficiently pre-computed and form the "inside" table of the famous Inside-Outside algorithm. | in our setting, trees come in pairs, and their joint probability crucially depends on their alignment. | contrasting |
train_1103 | For Korean we found that the baseline performed well using these values. | on our English and Chinese data, we found that somewhat higher smoothing values worked best, so we utilized values of 20 and 80 for constituent and distituent smoothing counts, respectively. | contrasting |
train_1104 | the fields f , the words w are generated independently: 4 where r(k) and f (k) are the record and field responsible for generating word w k , as determined by the segmentation c. The word choice model p w (w | t, v) specifies a distribution over words given the field type t and field value v. This distribution is a mixture of a global backoff distribution over words and a field-specific distribution which depends on the field type t. Although we designed our word choice model to be relatively general, it is undoubtedly influenced by the three domains. | we can readily extend or replace it with an alternative if desired; this modularity is one principal benefit of probabilistic modeling. | contrasting |
train_1105 | Inspection of the errors revealed the following problem: The alignment task requires us to sometimes align a sentence to multiple redundant records (e.g., play and score) referenced by the same part of the text. | our model generates each part of text from only one record, and thus it can only allow an alignment to one record. | contrasting |
train_1106 | This maximizes the probability of word segmentation w given a string s : This approach often implicitly includes heuristic criteria proposed so far 1 , while having a clear statistical semantics to find the most probable word segmentation that will maximize the probability of the data, here the strings. | they are still naïve with respect to word spellings, and the inference is very slow owing to inefficient Gibbs sampling. | contrasting |
train_1107 | The inference of this model interleaves adding and removing a customer to optimize t hw , d, and θ using MCMC. | in our case "words" are not known a priori: the next section describes how to accomplish this by constructing a nested HPYLM of words and characters, with the associated inference algorithm. | contrasting |
train_1108 | To avoid dependency on ngram order n, we actually used the ∞-gram language model (Mochihashi and Sumita, 2007), a variable order HPYLM, for characters. | for generality we hereafter state that we used the HPYLM. | contrasting |
train_1109 | This quantity is random since the observation string itself is. | we note that the distribution of O n is unaffected if one relabels the words in Ω n . | contrasting |
train_1110 | Correct stress placement is important in textto-speech systems because it affects the accuracy of human word recognition (Tagliapietra and Tabossi, 2005;Arciuli and Cupples, 2006). | the issue has often been ignored in previous letter-to-phoneme (L2P) systems. | contrasting |
train_1111 | In many languages, only two levels of stress are distinguished: stressed and unstressed. | some languages exhibit more than two levels of stress. | contrasting |
train_1112 | Performance is 1-3% lower on letters than on phonemes. | the performance of ORACLESYL drops much less on letters. | contrasting |
train_1113 | Following the above approaches, we can expand the input or output symbols of our L2P system to include stress. | since both decision tree systems and our L2P predictor utilize only local context, they may produce invalid global output. | contrasting |
train_1114 | Numerous studies have demonstrated that active learners can make more efficient use of unlabelled data than do passive learners (Abe and Mamitsuka, 1998;Miller et al., 2004;Culotta and McCallum, 2005). | relatively few researchers have applied active learning techniques to the L2P domain. | contrasting |
train_1115 | 5 The idea of lowering the specificity of letter class questions as the context length increases is due to Kienappel and Kneser (2001), and is intended to avoid overfitting. | their configuration differs from ours in that they use longer context lengths (4 for German and 5 for English) and ask letter class questions at every position. | contrasting |
train_1116 | Oh and Choi (2002) define classes of phonemes and assign various distances between phonemes of different classes. | we make use of phonological descriptors to define the similarity between phonemes in this paper. | contrasting |
train_1117 | and very low lemmatization quality. | the experiments with the tree structure were very successful. | contrasting |
train_1118 | By doing so, we can guarantee to use at least all the top n best parse trees in the forest. | please note that even after pruning there is still exponential number of additional trees embedded in the forest because of the sharing structure of forest. | contrasting |
train_1119 | Although a forest can cover much more phrases than a single tree does, there are still many non-syntactic phrases that cannot be captured by a forest due to structure divergence issue. | tree sequence is a good solution to non-syntactic translation equivalence modeling. | contrasting |
train_1120 | 3) In Table 3, the fully-lexicalized rules are the major part (around 60%), followed by the partially-lexicalized (around 35%) and unlexicalized (around 15%). | in Table 2, partially-lexicalized rules extracted from training corpus are the major part (more than 70%). | contrasting |
train_1121 | Another option for the contribution of OOV phrases is to take log of expected probability ratio (LEPR): It is not difficult to prove that there is no difference between Model 1 and Model 2 when ELPR scoring is used for sentence selection. | the situation is different for LEPR scoring: the two models produce different sentence rankings in this case. | contrasting |
train_1122 | (Callison-Burch and Osborne, 2003b;Callison-Burch and Osborne, 2003a) provide a co-training approach to MT, where one language pair creates data for another language pair. | our co-training approach uses consensus translations and our setting for active learning is very different from their semi-supervised setting. | contrasting |
train_1123 | As properties in infobox are not complete sentences and do not present relevant arguments, it is inappropriate to concatenate them as a summary. | they are good indicators for summary generation. | contrasting |
train_1124 | This strategy is commonly used in multi-document summarization (Barzilay et al., 1999;Goldstein et al., 2000;Radev et al., 2000), where the combination step eliminates the redundancy across selected excerpts. | separating the two steps may not be optimal for this task -the balance between coverage and redundancy is harder to achieve when a multi-paragraph summary is generated. | contrasting |
train_1125 | The NLG process (Reiter and Dale, 2000) is often viewed as a pipeline consisting of content planning (selecting and structuring the story's content), microplanning (sentence ag-gregation, generation of referring expressions, lexical choice), and surface realization (agreement, verb-subject ordering). | story generation systems typically operate in two phases: (a) creating a plot for the story and (b) transforming it into text (often by means of template-based NLG). | contrasting |
train_1126 | While debating, participants often refer to and acknowledge the viewpoints of the opposing side. | they do not endorse this rival opinion. | contrasting |
train_1127 | An aspect may be associated with both of the debate topics, but not, by itself, be discriminative between stances toward the topics. | opinions toward that aspect might discriminate between them. | contrasting |
train_1128 | This results in Recall and Accuracy being the same. | all of the systems do not classify a post if the post does not contain the information it needs. | contrasting |
train_1129 | Because the iPhone and the Blackberry are both phones, the word "phone" does not have any distinguishing power in debates. | the PMI measure of "e-mail" suggests that it is not closely related to the debate topics, though it is, in fact, a desirable feature for smart phone users, even more so with Blackberry users. | contrasting |
train_1130 | In the second sentence, the lexicon does hit the word "worth," and, using syntactic rules, we can determine it is negated. | the opinion-target pairing system only tells us that the opinion is tied to the "it." | contrasting |
train_1131 | Theirs is a supervised bag-of-words system using unigrams, bigrams, and trigrams as features. | our approach is unsupervised, and exploits different types of information. | contrasting |
train_1132 | Several researchers have recognized the important role discourse plays in opinion analysis (Polanyi and Zaenen, 2005;Snyder and Barzilay, 2007;Somasundaran et al., 2008;Asher et al., 2008;Sadamitsu et al., 2008). | previous work did not account for concessions in determining whether an opinion supports one side or the other. | contrasting |
train_1133 | The lack of Chinese sentiment corpora limits the research progress on Chinese sentiment classification. | there are many freely available English sentiment corpora on the Web. | contrasting |
train_1134 | Actually, TSVM(ENCN2) is similar to CoTrain because CoTrain also combines the results of two classifiers in the same way. | the co-training approach can train two more effective classifiers, and the accuracy values of the component English and Chinese classifiers are 0.775 and 0.790, respectively, which are higher than the corresponding TSVM classifiers. | contrasting |
train_1135 | Though, their techniques for proving algorithm convergence and correctness can be readily adapted for our models, (Li et al., 2008) do not incorporate dual supervision as we do. | while (Sindhwani et al., 2008) do incorporate dual supervision in a non-linear kernelbased setting, they do not enforce non-negativity or orthogonality -aspects of matrix factorization models that have shown benefits in prior empirical studies, see e.g., (Ding et al., 2006). | contrasting |
train_1136 | Ideally, we would like to evaluate a random sample of the more than 1,000 languages represented in PANDICTIONARY. | 5 a high-quality evaluation of translation between two languages requires a person who is fluent in both languages. | contrasting |
train_1137 | Recent studies Kozareva et al., 2008) show that if the size of a corpus, such as the Web, is nearly unlimited, a pattern has a higher chance to explicitly appear in the corpus. | corpus size is often not that large; hence the problem still exists. | contrasting |
train_1138 | The annotated PropBank corpus, and therefore implicitly its role labels inventory, has been largely adopted in NLP because of its exhaustiveness and because it is coupled with syntactic annotation, properties that make it very attractive for the automatic learning of these roles and their further applications to NLP tasks. | the labelling choices made by PropBank have recently come under scrutiny (Zapirain et al., 2008;. | contrasting |
train_1139 | Arguments receiving labels A0-A5 or AA do not express consistent semantic roles and are specific to a verb, while arguments receiving an AM-X label are supposed to be adjuncts and the respective roles they express are consistent across all verbs. | among argument labels, A0 and A1 are assigned attempting to capture Proto-Agent and Proto-Patient properties (Dowty, 1991). | contrasting |
train_1140 | For example, Agent and Instrumental Cause are often subjects (of verbs selecting animate and inanimate subjects respectively); Patients/Themes can be direct objects of transitive verbs and subjects of change of state verbs; Goal and Beneficiary can be passivised and undergo the dative alternation; Instrument and Comitative are expressed by the same preposition in many languages (see Levin and Rappaport Hovav (2005).) | most annotation schemes in NLP and linguistics assume that semantic role labels are atomic. | contrasting |
train_1141 | The resulting scores are cheap and objective. | studies such as Callison-Burch et al. | contrasting |
train_1142 | This matches well with our intuitions: We see some noise-induced degradation for the entailment features, but not much. | surface-based features are better at detecting bad translations than at discriminating among good ones. | contrasting |
train_1143 | These automatic evaluation metrics allow developers to optimize their systems without the need for expensive human assessments for each of their possible system configurations. | estimating the system output quality according to its similarity to human references is not a trivial task. | contrasting |
train_1144 | In order to tackle language variability in the context of Machine Translation, a considerable effort has also been made to include deeper linguistic information in automatic evaluation metrics, both syntactic and semantic (see Section 2 for details). | the most commonly used metrics are still based on n-gram matching. | contrasting |
train_1145 | Therefore, we need additional meta-evaluation criteria in order to clarify the behavior of linguistic metrics as compared to n-gram based metrics. | there are some exceptions. | contrasting |
train_1146 | This result supports the intuition in (Banerjee and Lavie, 2005) that correlation at segment level is necessary to ensure the reliability of metrics in different situations. | the correlation values of metrics at segment level have also drawbacks related to their interpretability. | contrasting |
train_1147 | In general, the linguistic metrics do not improve the ability to capture wrong translations (horizontal axis in the figure). | again, the combining metric ULC achieves the same reliability as the best n-gram based metric. | contrasting |
train_1148 | To better leverage syntactic constraint yet still allow non-syntactic translations, Chiang (2005) introduces a count for each hypothesis and accumulates it whenever the hypothesis exactly matches syntactic boundaries on the source side. | and Cherry (2008) accumulate a count whenever hypotheses violate constituent boundaries. | contrasting |
train_1149 | Similar to previous methods, our SDB model is integrated into the decoder's log-linear model as a feature so that we can inherit the idea of soft constraints. | to the constituent matching/violation counting (CMVC) (Chiang, 2005;Cherry, 2008), our SDB model has 2 Here we expand the definition of phrase to include both syntactic and non-syntactic phrases. | contrasting |
train_1150 | Interestingly, hierarchical phrase-based models provide this bene t without making any linguistic commitments beyond the structure of the model. | the system's lack of linguistic commitment is also responsible for one of its great-est drawbacks. | contrasting |
train_1151 | 1 into English using the following set of rules: To correctly translate the sentence, a hierarchical phrase-based system needs to model the subject noun phrase, object noun phrase and copula constructions; these are captured by rules X a , X d and X b respectively, so this set of rules represents a hierarchical phrase-based system that can be used to correctly translate the Chinese sentence. | note that the Chinese word order is correctly preserved in the subject (X a ) as well as copula constructions (X b ), and correctly inverted in the object construc- although it can generate the correct translation in Fig Figure 2: The derivation that leads to the correct translation The computation of the dominance relationship using this alignment information will be discussed in detail in the next section. | contrasting |
train_1152 | Successful phrase-based systems typically employ language models of order higher than two. | our models so far have the following important "Markovian" property: the cost of a path is additive relative to the costs of transitions. | contrasting |
train_1153 | While this problem was manageable for the artificial monolingual word re-ordering (which had only one possible translation for each source word), it becomes unwieldy for the real translation experiments, which is why in this paper we only considered bigram LMs for these experiments. | we know how to handle this problem in principle, and we now describe a method that we plan to experiment with in the future. | contrasting |
train_1154 | The Minkowski-Weyl theorem (Rockafellar, 1970) ensures that Z(x) has a representation of the form Z(x) = {z ∈ R |A| | Az ≤ b}, for some p-by-|A| matrix A and some vector b in R p . | it is not easy to obtain a compact representation (where p grows polynomially with the number of words n). | contrasting |
train_1155 | Note that conditions 1-2-3 are equivalent to 1-2-3 , in the sense that both define the same set Y(x). | as we will see, the latter set of conditions is more convenient. | contrasting |
train_1156 | Thus, transition-based parsers normally run in linear or quadratic time, using greedy deterministic search or fixed-width beam search Attardi, 2006;Johansson and Nugues, 2007;Titov and Henderson, 2007), and graph-based models support exact inference in at most cubic time, which is efficient enough to make global discriminative training practically feasible (McDonald et al., 2005a;McDonald et al., 2005b). | one problem that still has not found a satisfactory solution in data-driven dependency parsing is the treatment of discontinuous syntactic constructions, usually modeled by non-projective dependency trees, as illustrated in Figure 1. | contrasting |
train_1157 | Additionally prototype-driven grammar induction needs to be used in conjunction with other unsupervised methods (distributional similarity and CCM (Klein and Manning, 2004)) to attain reasonable accuracy, and is only evaluated on length 10 or less sentences with no lexical information. | gE uses only the provided constraints and unparsed sentences, and is used to train a feature-rich discriminative model. | contrasting |
train_1158 | The development and tuning of the above methods constitute the encoding of prior domain knowledge about the desired syntactic structure. | our framework provides a straightforward and explicit method for incorporating prior knowledge. | contrasting |
train_1159 | If there are constraint functions G for all model feature functions F j , and the target expectations G are estimated from labeled data, then the globally optimal parameter setting under the GE objective function is equivalent to the maximum likelihood solution. | gE does not require such a one-to-one correspondence between constraint functions and model feature functions. | contrasting |
train_1160 | Because these features consider multiple edges, including them in the CRF model would make exact inference intractable (McDonald and Satta, 2007). | the CRF may consider the distance between head and child, whereas DMV does not model distance. | contrasting |
train_1161 | We can see that the name chunk usually has the largest mutual information. | the name chunk always needs to be transliterated, and transliteration is often more difficult than translation by lexicon. | contrasting |
train_1162 | We exploit this wide variation with bagging, sampling from automatically extracted seeds to reduce semantic drift. | semantic drift still occurs in later iterations. | contrasting |
train_1163 | This second criterion aims to increase recall. | the selected instances are highly likely to introduce drift. | contrasting |
train_1164 | A drift(t, n, m) of 0.2 corresponds to a 20% difference in average similarity between L 1...n and L (N −m)...N for term t. Drift can be used as a post-processing step to filter terms that are a possible consequence of drift. | our main proposal is to incorporate the drift measure directly within the WMEB bootstrapping algorithm, to detect and then prevent drift occuring. | contrasting |
train_1165 | These types of relations have mostly been identified in isolation by event pairwise comparison. | this approach neglects logical constraints between temporal relations of different types that we believe to be helpful. | contrasting |
train_1166 | 3 Only for Task B we were unable to reach the performance of a rule-based entry to the challenge. | we do perform better than all pure machine 1 It is clearly possible to incorporate weighted constraints into ILPs, but how to learn the corresponding weights is not obvious. | contrasting |
train_1167 | The intuition behind the previous formula can also be captured using a local classifier. | 6 Markov Logic also allows us to say more: beforeDCT (e 1 ) ∧ ¬beforeDCT (e 2 ) ⇒ before (e 1 , e 2 ) (2) In this case, we made a statement about more global properties of a temporal ordering that cannot be captured with local classifiers. | contrasting |
train_1168 | One way is to pick a task and then choose formulae that increase the accuracy for this task on DEV. | our primary goal is to improve the performance of all the tasks together. | contrasting |
train_1169 | We have presented a profile-based Cross Document Coreference (CDC) approach based on a novel fuzzy relational clustering algorithm KARC. | to traditional hard clustering methods, KARC produces fuzzy sets of identities which better reflect the intrinsic uncertainty of the CDC problem. | contrasting |
train_1170 | Although MT may be used in solving this task, it is only used by the algorithms -the final evaluation is done in the source language. | in many real-life situations, such as global business, international tourism, or intelligence work, users may not be able to read the source language. | contrasting |
train_1171 | Since this annotation task was a 10-way selection, with multiple selections possible, there were some disagreements. | if categorized broadly into 5W System errors only, MT errors only, and both 5W System and MT errors, then the annotators had a substantial level of agreement (κ=0.75 for error type, on sentences where both annotators indicated an error). | contrasting |
train_1172 | English-LF also had more Partial answers on the What question: 66% Correct and 12% Partial, versus 75% Correct and 1% Partial for English-function. | englishfunction was more likely to return answers that contained incorrect extra information, such as another 5W or a second predicate. | contrasting |
train_1173 | The results show that, on average, ASIA performs better. | we should emphasize that for the three classes: movie, person, and video game, ASIA did not initially converge to the correct instance list given the most natural concept name. | contrasting |
train_1174 | Indeed, the commonly used hyponymy, synonymy and some cases of the meronymy relations are special cases of lexical reference. | lexical reference is a broader relation. | contrasting |
train_1175 | However this is a rather loose notion, which only indicates that terms are semantically "related" and are likely to co-occur with each other. | lexical reference is a special case of lexical association, which specifies concretely that a reference to the meaning of one term may be inferred from the other. | contrasting |
train_1176 | It is worth noting that in our sample 57% of All-N errors, 62% These three rows replace the last row of Table 2 Places were extracted by the All-N bottom method and thus may be identified as less reliable. | this split was not observed to improve performance in the application oriented evaluations of Section 6. | contrasting |
train_1177 | Following our error analysis, future research is needed for addressing each specific type of error. | during the analysis we observed that all types of erroneous rules tend to relate terms that are rather unlikely to co-occur together. | contrasting |
train_1178 | 6 We also examined another filtering score, the cosine similarity between the vectors representing the two rule sides in LSA (Latent Semantic Analysis) space (Deerwester et al., 1990). | as the results with this filter resemble those for Dice we present results only for the simpler Dice filter. | contrasting |
train_1179 | We make two contributions in the paper: On one hand, we find an effective way of constructing high-quality semantic classes in the patternbased category which deals with multimembership. | we demonstrate, for the first time, that topic modeling can be utilized to help mining the peer relationship among words. | contrasting |
train_1180 | Wu's model can be understood as a strict hierarchical maximum-alignment method. | our alignments are soft (we sum over them), and we do not require strictly isomorphic syntactic structures. | contrasting |
train_1181 | It should be noted that exponential decay is not a good choice from a theoretical point of view, because it does not satisfy one of the necessary con-ditions for convergence-the sum of the learning rates must diverge to infinity (Spall, 2005). | this is probably not a big issue for practitioners because normally the training has to be terminated at a certain number of iterations in practice. | contrasting |
train_1182 | Finally, context may be of some help, but "test" is ambiguous between a noun and verb, and "gasolines" is only seen once in the training data, so there is no guarantee that context is sufficient to make a correct judgment. | some of the other contexts in which "reformulated" appears in the test set, such as "testing of reformulated gasolines," provide strong evidence that it can start a NP, since "of" is a highly reliable indicator that a NP is to follow. | contrasting |
train_1183 | will not encounter any previously unseen words. | to speed up training during our experiments and, in some cases, to avoid running out of memory, we replaced words appearing twice or fewer times in the data with the special symbol * UNKNOWN * . | contrasting |
train_1184 | Our unlabeled sample complexity results show that even with access to a small amount of unlabeled text, 6000 sentences more than what appears in the training and test sets, smoothing using the HMM yields 0.78 F1 on rare chunks. | the smoothed system requires 25,000 more sentences before it outperforms the baseline system on all chunks. | contrasting |
train_1185 | The entire process is fully automated and yields better performance than any existing state-of-the-art system, even though our models were not provided with any additional linguistic knowledge (for example, explicit syntactic constraints to avoid certain tag combinations such as "V V", etc.). | it is easy to model some of these linguistic constraints (both at the local and global levels) directly using integer programming, and this may result in further improvements and lead to new possibilities for future research. | contrasting |
train_1186 | Nakagawa (2004) described a training method based on a word-based Markov model and a character-based maximum entropy model that can be completed in a reasonable time. | this training method is limited by the generatively-trained Markov model in which informative features are hard to exploit. | contrasting |
train_1187 | Conditional random fields (CRFs) (Lafferty et al., 2001) further improve the performance (Kudo et al., 2004;Shi and Wang, 2007) by performing whole-sequence normalization to avoid label-bias and length-bias problems. | cRF-based algorithms typically require longer training times, and we observed an infeasible convergence time for our hybrid model. | contrasting |
train_1188 | Due to the obvious improvement brought by annotation adaptation to both word segmentation and Joint S&T, we can safely conclude that the knowledge can be effectively transferred from on an- Parsing F 1 % gold-standard segmentation 82.35 baseline segmentation 80.28 adapted segmentation 81.07 Table 4 we find that out of 30 word clusters appeared in the developing set of CTB, 13 clusters benefit from the annotation adaptation strategy, while 4 clusters suffer from it. | the compositive error rate of Recall for all word clusters is reduced by 20.66%, such a fact invalidates the effectivity of annotation adaptation. | contrasting |
train_1189 | proposed a method based on patterns of a sequence of morphemes (Monma et al., 2003). | the target of the research is closed-captions of Japanese TV shows, in which less than or equal to 2 lines text is displayed on a screen and the text all switches to other text at a time. | contrasting |
train_1190 | This ratio is less than half of that for all the bunsetsu boundaries. | when the bunsetsu boundary right after the bunsetsu which does not depend on the next bunsetsu, the ratio of linefeed insertion was 52.7%. | contrasting |
train_1191 | The general method is to let the two classifiers predict the class for a given sample, and if they agree, the hypothesized label is used. | when this agreement-based approach is used for prosodic event detection, we notice that there is not only difference in the labeling accuracy between positive and negative samples, but also an imbalance of the self-labeled positive and negative examples (details in Section 6). | contrasting |
train_1192 | Adding these samples has a negative impact on the classifier's performance. | our confidence-based approach balances the number of positive and negative samples and significantly reduces the error rates for the negative samples as well, thus leading to performance improvement. | contrasting |
train_1193 | Specific instances of this general algorithm have recently been proposed for two linear similarity measures (Tromble et al., 2008;Zhang and Gildea, 2008). | the sentence similarity measures we want to optimize in MT are not linear functions, and so this fast algorithm for MBR does not apply. | contrasting |
train_1194 | Computing MBR even with simple non-linear measures such as BLEU, NIST or bagof-words F1 seems to require O(k 2 ) computation time. | these measures are all functions of features of e . | contrasting |
train_1195 | For example, taking a source parse as input, a tree-to-string decoder (e.g., ) pattern-matches the source parse with treeto-string rules and produces a string on the target side. | a string-to-tree decoder (e.g., (Galley et al., 2006;Shen et al., 2008)) is a parser that applies string-to-tree rules to obtain a target parse for the source string. | contrasting |
train_1196 | A common property of all the work mentioned above is that the combination models work on the basis of n-best translation lists (full hypotheses) of existing machine translation systems. | the n-best list only presents a very small portion of the entire search space of a Statistical Machine Translation (SMT) model while a majority of the space, within which there are many potentially good translations, is pruned away in decoding. | contrasting |
train_1197 | This may be regarded as favoring n-grams that are likely to appear in the reference translation (because they are likely in the derivation forest). | in order to score well on the BLEU metric for MT evaluation (Papineni et al., 2001), which gives partial credit, we would also like to favor lower-order ngrams that are likely to appear in the reference, even if this means picking some less-likely highorder n-grams. | contrasting |
train_1198 | Few types of arguments are shared between the chain and (fly X). | (charge X) shares many arguments with (accuse X), (search X) and (suspect X) (e.g., criminal and suspect). | contrasting |
train_1199 | Police pull over cars, but this schema does not have a chain involving cars. | (Y search) scores well with the 'police' chain and (search X) scores well in the 'defendant' chain too. | contrasting |
Subsets and Splits