text stringlengths 82 736 | label int64 0 1 |
|---|---|
our system implements a stacked bidirectional gated recurrent units based on a capsule network---our recurrent structure is a two layer stacked bidirectional network with gated recurrent unit cells | 1 |
for the gender identification task , mohammad and yang show that there are marked differences across genders in how they use emotion words in work-place email---mohammad and yang has shown that there are marked differences across genders in how they use emotion words in work-place email | 1 |
it has been shown that word embeddings are able to capture to certain semantic and syntactic aspects of words---the word embeddings can provide word vector representation that captures semantic and syntactic information of words | 1 |
the targetside 4-gram language model was estimated using the srilm toolkit and modified kneser-ney discounting with interpolation---we trained a 4-gram language model on this data with kneser-ney discounting using srilm | 1 |
bengio et al and kumar et al developed training paradigms which are inspired by the learning principle that humans can learn more effectively when training starts with easier concepts and gradually proceeds with more difficult ones---we use stanford corenlp for chinese word segmentation and pos tagging | 0 |
transliteration is the task of converting a word from one alphabetic script to another---transliteration is a process of rewriting a word from a source language to a target language in a different writing system using the word ’ s phonological equivalent | 1 |
relation extraction is a crucial task in the field of natural language processing ( nlp )---relation extraction is the task of automatically detecting occurrences of expressed relations between entities in a text and structuring the detected information in a tabularized form | 1 |
a second goal of these experiments was to show that the hmm taggers offer improved handling of ambiguity compared with the unigram tagger of scherrer and sagot---one goal of the experiments presented here was to validate the pipeline proposed earlier in scherrer and sagot | 1 |
this hypothesis is the basis for our algorithm for distinguishing literal and metaphorical senses---our hypothesis is a generalization of the original hypothesis since it allows a reducible sequence to form several adjacent subtrees | 1 |
we used a 4-gram language model which was trained on the xinhua section of the english gigaword corpus using the srilm 4 toolkit with modified kneser-ney smoothing---we train a 4-gram language model on the xinhua portion of the gigaword corpus using the sri language toolkit with modified kneser-ney smoothing | 1 |
eisner and satta define an oparser for split head automaton grammars which can be used for dependency parsing---eisner and satta give a cubic algorithm for lexicalized phrase structures | 1 |
a pun is a form of wordplay in which a word suggests two or more meanings by exploiting polysemy , homonymy , or phonological similarity to another word , for an intended humorous or rhetorical effect ( cite-p-15-3-1 )---our nnape model is inspired by the mt work of bahdanau et al which is based on bidirectional recurrent neural networks | 0 |
we derive our predicate-argument structures from a semantic parse based on the propbank annotation scheme---we use mateplus for srl which produces predicate-argument structures as per propbank | 1 |
in this paper we presented a word sense disambiguation based system for multilingual lexical substitution---work , we aim to investigate the effect of adding different contextual information | 0 |
the algorithm is similar to those for context-free parsing such as chart parsing and the cky algorithm---the algorithm is similar to those for context free parsing such as chart parsing and the cky algorithm | 1 |
visual question answering ( vqa ) is the task of answering natural-language questions about images---visual question answering ( vqa ) is the task of predicting a suitable answer given an image and a question about it | 1 |
for the word-embedding based classifier , we use the glove pre-trained word embeddings---we use the pre-trained glove 50-dimensional word embeddings to represent words found in the glove dataset | 1 |
semantic applications , such as qa or summarization , typically extract sentence features from a derived intermediate structure---semantic applications typically extract information from intermediate structures derived from sentences , such as dependency | 1 |
researches on the psychology of concepts show that categories in the human mind are not simply sets with clearcut boundaries---we train a kn-smoothed 5-gram language model on the target side of the parallel training data with srilm | 0 |
the word embeddings are pre-trained by skip-gram---table 1 presents the results from the automatic evaluation , in terms of bleu and nist test | 0 |
for all our classification experiments , we used the weka toolkit---we used weka for all our classification experiments | 1 |
representation learning is a promising technique for discovering features that allow supervised classifiers to generalize from a source domain dataset to arbitrary new domains---representation learning is the dominant technique for unsupervised domain adaptation , but existing approaches have two major weaknesses | 1 |
the trigram language model is implemented in the srilm toolkit---li and liu extended the character-level mt model by incorporating the pronunciation information | 0 |
daum茅 proposed a heuristic based non-linear mapping of source and target data to a high dimensional space---daum茅 proposed a feature space transformation method for domain adaptation based on a simple idea of feature augmentation | 1 |
for each production , an svm classifier is trained using a string subsequence kernel---for each of these productions , a supportvector machine classifier is trained using string similarity as the kernel | 1 |
sentence compression is a paraphrasing task where the goal is to generate sentences shorter than given while preserving the essential content---sentence compression is the task of compressing long sentences into short and concise ones by deleting words | 1 |
using these representations as features , bansal et al obtained improvements in dependency recovery in the mst parser---in bansal et al , better word embeddings for dependency parsing were obtained by using a corpus created to capture dependency context | 1 |
word alignment , which can be defined as an object for indicating the corresponding words in a parallel text , was first introduced as an intermediate result of statistical translation models ( cite-p-13-1-2 )---word alignment is a crucial early step in the training of most statistical machine translation ( smt ) systems , in which the estimated alignments are used for constraining the set of candidates in phrase/grammar extraction ( cite-p-9-3-5 , cite-p-9-1-4 , cite-p-9-3-0 ) | 1 |
in all cases , we propose a simple , unsupervised n-gram based model whose parameters are estimated using web counts---we found that simple , unsupervised models perform significantly better when n-gram frequencies are obtained from the web | 1 |
kilicoglu and bergler apply a combination of lexical and syntactic methods , improving on previous results and showing that quantifying the strength of a hedge can be beneficial for classification of speculative sentences---in our large-scale speech-to-speech translation system under development , the usrate is estimated to be under 20 % , i . e . , over80 % of unificationsare estimated to be failures | 0 |
we use skip-gram representation for the training of word2vec tool---to evaluate the performance of our proposed method , we use the semeval-2010 task 8 dataset | 0 |
the findings point out an array of issues that future qa research may need to solve---experiments point out an array of issues that future qa systems may need to solve | 1 |
event extraction is a task in information extraction where mentions of predefined events are extracted from texts---event extraction is the task of detecting certain specified types of events that are mentioned in the source language data | 1 |
particle swarm optimization is an evolutionary technique , inspired by the social behavior of birds---particle swarm optimization ( pso ) is a meta-heuristic intelligent technique inspired by social behavior of the swarm | 1 |
the first part of this proposal is concerned with the efficient discovery of publications in the web for a particular domain---dependency parsing is a simpler task than constituent parsing , since dependency trees do not have extra non-terminal nodes and there is no need for a grammar to generate them | 0 |
we report bleu scores to compare translation results---to minimize the objective , we use the diagonal variant of adagrad with minibatches | 0 |
on the remaining tweets , we trained a 10-gram word length model , and a 5-gram language model , using srilm with kneyser-ney smoothing---we used the sri language modeling toolkit to train a fivegram model with modified kneser-ney smoothing | 1 |
relation extraction ( re ) is the task of assigning a semantic relationship between a pair of arguments---relation extraction ( re ) is the task of recognizing relationships between entities mentioned in text | 1 |
also , while in previous approaches , the features are collected from corpora , those we make use of are retrieved from the lexicon entries---on the basis of features retrieved from the corpus , we make use of features retrieved from the lexicon | 1 |
we use binary crossentropy loss and the adam optimizer for training the nil-detection models---we use a binary cross-entropy loss function , and the adam optimizer | 1 |
we present an algorithm that uses the same knowledge sources to disambiguate different words---we trained the l1-regularized logistic regression classifier implemented in liblinear | 0 |
the weights of the different feature functions were optimised by means of minimum error rate training on the 2013 wmt test set---the nnlm weights are optimized as the other feature weights using minimum error rate training | 1 |
the current results thus are mostly in line with the findings of correa and sureka who found that deleted questions have a higher number of characters in the question body than closed questions---recently , bert , a pre-trained deep neural network , based on the transformer , has improved the state of the art for many natural language processing tasks | 0 |
the reordering approach improved the bleu score for the moses system from 28.52 to 30.86 on the nist 2006 evaluation data---4 submitted to the official nist 2006 mt evaluation , the reordered system also improved the bleu score substantially ( by 1 . 34 on nist 2006 data ) | 1 |
surdeanu and manning also show that the diversity of parsers is important for performance improvement when integrating different parsers in the supervised track---translation quality can be measured in terms of the bleu metric | 0 |
thus , we propose a new approach based on the expectation-maximization algorithm---hence we use the expectation maximization algorithm for parameter learning | 1 |
in addition , we use an english corpus of roughly 227 million words to build a target-side 5-gram language model with srilm in combination with kenlm---all evaluated systems use the same surface trigram language model , trained on approximately 340 million words from the english gigaword corpus using the srilm toolkit | 1 |
a morphological analysis consists of a part-of-speech tag ( pos ) , possibly other morphological features , and a lemma ( basic form ) corresponding to this tag and features combination ( see table 1 for examples )---to encode the original sentences we used word2vec embeddings pre-trained on google news | 0 |
automatic text generation is the process of converting non-linguistic data into coherent and comprehensible text---the model parameters in word embedding are pretrained using glove | 0 |
segmentation is the task of dividing a stream of data ( text or other media ) into coherent units---recent work by munteanu and marcu uses a bilingual lexicon to translate some of the words of the source sentence | 0 |
to train the model , we adopt the averaged perceptron algorithm with early update , following huang and sagae---several authors investigate neural network models that learn a vector of latent variables to represent each word | 0 |
our machine translation system is a phrase-based system using the moses toolkit---for language modeling , we computed 5-gram models using irstlm 7 and queried the model with kenlm | 0 |
these embedding vectors have been shown to improve a variety of language tasks including named entity recognition , phrase chunking , relation extraction , and part of speech induction---collobert et al propose a multi-task learning framework with dnn for various nlp tasks , including part-of-speech tagging , chunking , named entity recognition , and semantic role labelling | 0 |
empirical evaluations based on a large collection of opinionated review documents confirm that the proposed method effectively models personal opinions---extensive experimental results on a large collection of amazon reviews confirm our method significantly outperformed a user-independent generic opinion model | 1 |
esposito and radicioni proposed an algorithm which opens necessary nodes in a lattice in searching the best sequence---to resolve this , esposito and radicioni have proposed a carpediem algorithm which opens only necessary nodes in searching the best sequence | 1 |
the gricean maxim of brevity , applied to nlg in , suggests a preference for the second , shorter realization---we note that the gricean maxim of brevity , applied to nlg in , suggests a preference for the second , shorter realization | 1 |
callison-burch et al and marton et al augmented the translation phrase table with paraphrases to translate unknown phrases---marton et al use a monolingual text on the source side to find paraphrases to oov words for which the translations are available | 1 |
we propose a hierarchical siamese network with an attention mechanism at both word level in order to select the textual mention which better describes the relation---we use the attention mechanism with a context vector proposed in to reward such words which are important to the meaning of a relation and then aggregate their information in the sentence representation | 1 |
in this paper , we focus on identifying the nature of interactions among user pairs---in this paper , we present such an application , which aims to perform fine-grained analysis of user-interactions | 1 |
we used a support vector machine classifier with radial basis function kernels to classify the data---in this paper , we propose a method for keyword extraction using term-domain interdependence | 0 |
baroni et al show that word embeddings are able to outperform count based word vectors on a variety of nlp tasks---baroni et al argues that predict models such as word2vec outperform count based models on a wide range of lexical semantic tasks | 1 |
we use scikitlearn as machine learning library---relation extraction ( re ) is a task of identifying typed relations between known entity mentions in a sentence | 0 |
we described a graph-based approach for finding the optimal permutation---we experimentally evaluated the graph-based approach | 1 |
named entity recognition ( ner ) is a key technique for ie and other natural language processing tasks---named entity recognition ( ner ) is the task of identifying named entities in free text—typically personal names , organizations , gene-protein entities , and so on | 1 |
resnik uses selectional preferences of predicates for word sense disambiguation---in this work , we present an ann architecture that combines the effectiveness of typical ann models to classify sentences | 0 |
experimental results on two language pairs and three different sizes of training data show significant improvements of up to 4 bleu points over a conventionally trained smt model---sentence-pair training sets ( along with a development set for tuning ) show that stacking can yield an improvement of up to 4 bleu points over conventionally trained smt | 1 |
the task is to classify whether each sentence provides the answer to the query---and it is hard for users to interpret a topic only based on the multinomial distribution ( cite-p-21-1-16 ) | 0 |
raina et al use a logic form transformation derived from dependency parses and named entities---raina et al used a logical representation , and accordingly defined the set of operations by a commonly used theorem proving method | 1 |
advances in this field in the past 20 years , along with greater access to training data , make the application of such techniques to readability quite timely---in the past 20 years , along with greater access to training data , make the application of such techniques to readability quite timely | 1 |
interestingly , park and cardie conclude on the worthlessness of word pair features , given the existence of such resources---interestingly , park and cardie concluded on the worthlessness of word-based features , as long as hand-crafted linguistic features were used | 1 |
the feature weights 位 i are trained in concert with the lm weight via minimum error rate training---instead of directly training an lm on these corpora , we extracted from them in-domain sentences using the moore-lewis filtering method , more specifically its implementation in xenc | 0 |
using this approach , the joint distribution of all variables is described by only the most systematic variable interactions , thereby limiting the number of parameters to be estimated , supporting computational efficiency , and providing an understanding of the data---the most important systematic interactions among variables limits the number of parameters to be estimated , supports computational efficiency , and provides an understanding of the data | 1 |
therefore , dependency parsing is a potential “ sweet spot ” that deserves investigation---dependency parsing is a topic that has engendered increasing interest in recent years | 1 |
part-of-speech ( pos ) tagging is a well studied problem in these fields---part-of-speech ( pos ) tagging is a fundamental task in natural language processing | 1 |
we separately test the feasibility of our approach against the data set published by durrett and denero , five data sets over three languages---we also compare our results to those obtained using the system of durrett and denero on the same test data | 1 |
for example , mikolov et al identify phrases using a monolingual point-wise mutual information criterion with discounting---mikolov et al identify phrases using a monolingual point-wise mutual information criterion with discounting | 1 |
bannard and callison-burch introduced the pivot approach to extracting paraphrase phrases from bilingual parallel corpora---bannard and callison-burch first presented the method to learn paraphrase phrases from a bilingual phrase table | 1 |
we use the srilm toolkit to compute our language models---we train a trigram language model with the srilm toolkit | 1 |
we use word embeddings of dimension 100 pretrained using word2vec on the training dataset---we pretrain word vectors with the word2vec tool on the news dataset released by ding et al , which are fine-tuned during training | 1 |
recently , a recurrent neural network architecture was proposed for language modelling---neural language modeling has since demonstrated powerful capabilities at the word level and character level | 1 |
for both languages , we used the srilm toolkit to train a 5-gram language model using all monolingual data provided---semantic role labeling ( srl ) is the process of extracting simple event structures , i.e. , “ who ” did “ what ” to “ whom ” , “ when ” and “ where ” | 0 |
srilm was employed to train a 5-gram language models with all japanese corpus in cj corpus and ej corpus---experimental results on duc2004 data sets and some expanded data demonstrate the good quality of our summaries | 0 |
we use the europarl parallel corpus 3 for all language pairs except for vietnamese-english---we use the europarl english-french parallel corpus plus around 1m segments of symantec translation memory | 1 |
we obtained distributed word representations using word2vec 4 with skip-gram---in this paper , we adopt continuous bag-of-word in word2vec as our context-based embedding model | 1 |
this approach was first suggested in , where parameterized heuristic rules are combined with a genetic algorithm into a system for keyphrase extraction that automatically identifies keywords in a document---the dialogs were further annotated using the anvil tool to create a set of target referring expressions | 0 |
we used the penn wall street journal treebank as training and test data---we trained and tested the model on data from the penn treebank | 1 |
we used srilm to build a 4-gram language model with interpolated kneser-ney discounting---here is to deterministically choose a shift or reduce action | 0 |
the much higher accuracy of our system on the noisy dataset shows that our meaning-based approach understands the meaning of each quantity more---on the noisy dataset shows that our meaning-based approach understands the meaning of each quantity | 1 |
okazaki et al proposed using a logistic regression model for approximate dictionary matching---okazaki et al utilized substring substitution rules and incorporated the rules into a l 1 -regularized logistic regression model | 1 |
our machine translation system is a phrase-based system using the moses toolkit---as a baseline system for our experiments we use the syntax-based component of the moses toolkit | 1 |
we use state-of-the-art word embedding methods , namely continuous bag of words and global vectors---our normalization approach is based on continuous distributed word vector representations , namely the state-of-the-art method word2vec | 1 |
in recent years , corpus based approaches to machine translation have become predominant , with phrase based statistical machine translation being the most actively progressing area---a trigram language model with modified kneser-ney discounting and interpolation was used as produced by the srilm toolkit | 0 |
recaps not only help the audience absorb the essence of previous episodes , but also grab people ’ s attention with upcoming plots---shows help the audience absorb the essence of previous episodes , and grab their attention with upcoming plots | 1 |
mikolov et al proposed a computationally efficient method for learning distributed word representation such that words with similar meanings will map to similar vectors---mikolov et al proposed a faster skip-gram model word2vec 5 which tries to maximize classification of a word based on another word in the same sentence | 1 |
similarly , lazaridou et al improve the word representations of derivationally related words by composing vector space representations of stems and derivational suffixes---lazaridou et al apply compositional methods by having the stem and affix representations in order to estimate the distributional representation of morphologically complex words | 1 |
we use pretrained 300-dimensional english word embeddings---we use the word2vec tool with the skip-gram learning scheme | 1 |
language models were trained with the kenlm toolkit---the 5-gram target language model was trained using kenlm | 1 |
haghighi , berg-kirkpatrick , and klein proposed a generative model for inducing a bilingual lexicon from monolingual text by exploiting orthographic and contextual similarities among the words in two different languages---haghighi et al presented a generative model based on canonical correlation analysis , in which monolingual features such as the context and orthographic substrings of words were taken into account | 1 |
we used the stanford parser to generate the grammatical structure of sentences---we use the stanford parser to generate the grammar structure of review sentences for extracting syntactic d-features | 1 |
conditional random fields are discriminative structured classification models for sequential tagging and segmentation---conditional random fields are global discriminative learning algorithms for problems with structured output spaces , such as dependency parsing | 1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.