HHousen/ParaSCI · Datasets at Hugging Face

sentence1 stringlengths 16 446	sentence2 stringlengths 14 436
mead is centroid based multi-document summarizer which generates summaries using cluster centroids produced by topic detection and tracking system .	mead is a centroid based multi document summarizer which generates summaries using cluster centroids produced by topic detection and tracking system .
brown et al present a hierarchical word clustering algorithm that can handle a large number of classes and a large vocabulary .	our first choice is the bottom-up agglomerative word clustering algorithm of brown et al , which derives a hierarchical clustering of words from unlabeled data .
in this paper , we present a greedy non-directional parsing algorithm which doesn ’ t need a fully connected parse and can learn from partial parses .	in this paper , we present a dependency parsing algorithm which can train on partial projected parses and can take rich syntactic information as features for learning .
in this paper , we propose a neural architecture for coherence assessment that can capture long range entity transitions .	following this tradition , in this paper we propose to neuralize the popular entity grid models .
word sense disambiguation ( wsd ) is a particular problem of computational linguistics which consists in determining the correct sense for a given ambiguous word .	word sense disambiguation ( wsd ) is a key enabling technology that automatically chooses the intended sense of a word in context .
the weights of the log-linear interpolation model were optimized via minimum error rate training on the ted development set , using 200 best translations at each tuning iteration .	the weights of the different feature functions were optimised by means of minimum error rate training on the 2013 wmt test set .
kalchbrenner et al show that a cnn for modeling sentences can achieve competitive results in polarity classification .	to capture the relation between words , kalchbrenner et al propose a novel cnn model with a dynamic k-max pooling .
kim et al proposed a convolutional module to process complex inputs for the problem of language modeling .	kim et al apply a simple convolutional neural network model , which uses character level inputs for word representations .
ranking methods based on importance scores are proposed for keyphrase extraction .	therefore , automatic keyphrase extraction is an important research task .
similarly , korhonen et al relied on the information bottleneck and subcategorisation frame types to induce soft verb clusters .	korhonen et al used verb-frame pairs to cluster verbs into levin-style semantic classes .
twitter is the medium where people post real time messages to discuss on the different topics , and express their sentiments .	twitter is a popular microblogging service , which , among other things , is used for knowledge sharing among friends and peers .
the language model is a large interpolated 5-gram lm with modified kneser-ney smoothing .	we use sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .
success , however , depends on a high-coverage dictionary .	this success rests on a high-coverage dictionary .
for the phrase based system , we use moses with its default settings .	in all submitted systems , we use the phrase-based moses decoder .
the rules were extracted using the pos tags generated by the treetagger .	the source and target sentences are tagged respectively using the treetagger and amira toolkits .
to encode the original sentences we used word2vec embeddings pre-trained on google news .	we used the pre-trained word embeddings that were learned using the word2vec toolkit on google news dataset .
for the information-access applications described above .	however , consider the interactive information-access application described above .
we use 100-dimension glove vectors which are pre-trained on a large twitter corpus and fine-tuned during training .	we use glove word embeddings , which are 50-dimension word vectors trained with a crawled large corpus with 840 billion tokens .
currently , recurrent neural network based models are widely used on natural language processing tasks for excellent performance .	recurrent neural network architectures have proven to be well suited for many natural language generation tasks .
and consequently , we propose a weakly supervised fully-bayesian approach to pos tagging , which relaxes the unrealistic assumption by automatically acquiring the lexicon from a small amount of pos-tagged data .	as a result , we investigated a weakly supervised fully-bayesian approach to pos tagging , which relaxes the unrealistic assumption by automatically acquiring the lexicon from a small amount of postagged data .
we use pre-trained 50 dimensional glove vectors 4 for word embeddings initialization .	we use 300d glove vectors trained on 840b tokens as the word embedding input to the lstm .
while automatic evaluation methods like bleu can be useful for estimating translation quality , a higher score is no guarantee of quality improvement .	aggregating evaluation methods like bleu give a useful overview of the quality of a translation , but they do not afford specific information and leave too many details to chance .
for building our ap e b2 system , we set a maximum phrase length of 7 for the translation model , and a 5-gram language model was trained using kenlm .	we calculated the language model probabilities using kenlm , and built a 5-gram language model from the english gigaword fifth edition .
which is a generalization of current perceptron-based reranking methods .	the subtree ranking approach is a generalization of the perceptron-based approach .
neural networks , working on top of conventional n-gram back-off language models , have been introduced in as a potential means to improve discrete language models .	neural networks , working on top of conventional n-gram back-off language models , have been introduced in as a potential means to improve discrete language models .
model fitting for our model is based on the expectation-maximization algorithm .	we estimate the parameters by maximizingp using the expectation maximization algorithm .
hammarstr枚m and borin give an extensive overview of stateof-the-art unsupervised learning of morphology .	hammarstr枚m and borin presented a literature survey on unsupervised learning of morphology , including methods for learning morphological segmentation .
following the work of koo et al , we used a tagger trained on the training data to provide part-of-speech tags for the development and test sets , and used 10-way jackknifing to generate part-of-speech tags for the training set .	following koo et al , we used the mxpost tagger trained on the full training data to provide part-of-speech tags for the development and the test set , and we used 10-way jackknifing to generate tags for the training set .
word sense disambiguation ( wsd ) is a natural language processing ( nlp ) task in which the correct meaning ( sense ) of a word in a given context is to be determined .	word sense disambiguation ( wsd ) is the task to identify the intended sense of a word in a computational manner based on the context in which it appears ( cite-p-13-3-4 ) .
ahsan and kolachina introduce a hybrid mt system that utilised online mt engines for msmt .	mellebeek et al introduced a hybrid mt system that utilised online mt engines for msmt .
coreference resolution is the next step on the way towards discourse understanding .	coreference resolution is the task of clustering a set of mentions in the text such that all mentions in the same cluster refer to the same entity .
we are the first to tie figurative language to the social context in which it is produced and show its relation to internal and external .	to the best of our knowledge , this is the first time figurative language is tied to the social context in which it appears .
coreference resolution is the process of linking together multiple referring expressions of a given entity in the world .	coreference resolution is the task of partitioning a set of mentions ( i.e . person , organization and location ) into entities .
we show that our method disambiguates a significant proportion of subject-object ambiguities in german .	we show that a simple method disambiguates some subject-object ambiguities in german , while making few errors .
we parsed the corpus with rasp and with the stanford pcfg parser .	we extracted the features from the gigaword corpus , which was first parsed using the rasp parser .
the alignment template approach uses word classes rather than lexical items to model phrase translation .	the alignment template model enhanced phrasal generalizations by using words classes rather than the words themselves .
collection comprises 132 , 229 dialogues containing a total of 764 , 146 turns / utterances that have been extracted from 753 movies .	the collected dataset comprises 132,229 dialogues containing a total of 764,146 turns that have been extracted from 753 movies .
we use the scikit-learn toolkit as our underlying implementation .	we trained the five classifiers using the svm implementation in scikit-learn .
in this paper , we propose a novel task that is crucial and generic from the viewpoint of health surveillance .	in this paper , we propose a more generalized task setting for public surveillance .
wang et al use all amr concepts and relations that appear in the training set as possible parameters if they appear in any sentence containing the same lemma as 蟽 0 and 尾 .	wang et al use all concepts that occur in the training data in the same sentence as the lemma of the node , leading to hundreds or thousands of possible actions from some states .
the language model is trained and applied with the srilm toolkit .	it has been trained with the srilm toolkit on the target side of all the training data .
we propose a method for extracting semantic orientations of phrases ( pairs of an adjective and a noun .	we proposed a method for extracting semantic orientations of phrases ( pairs of an adjective and a noun ) .
semantic role labeling ( srl ) is defined as the task to recognize arguments for a given predicate and assign semantic role labels to them .	semantic role labeling ( srl ) is the process of producing such a markup .
which beyonds the capability of phrase-based mt , we extend the search-aware tuning framework from phrase-based mt to syntax-based mt , in particular the hierarchical phrase-based translation model .	we extend this approach from phrase-based translation to syntax-based translation by generalizing the evaluation metrics for partial translations to handle tree-structured derivations in a way inspired by inside-outside algorithm .
svms have been shown to be robust in classification tasks involving text where the dimensionality is high .	svms are frequently used for text classification and have been applied successfully to nli .
we have adopted a supervised approach , a svm polynomial kernel classifier trained with the data provided by the challenge .	we follow a supervised approach , exploiting a svm polynomial kernel classifier trained with the challenge data .
the model parameters are trained using minimum error-rate training .	minimum error rate training is applied to tune the cn weights .
conditional random fields are undirected graphical models trained to maximize a conditional probability .	conditional random fields are probabilistic models for labelling sequential data .
coreference resolution is the task of clustering a sequence of textual entity mentions into a set of maximal non-overlapping clusters , such that mentions in a cluster refer to the same discourse entity .	coreference resolution is the process of determining whether two expressions in natural language refer to the same entity in the world .
our baseline is a standard phrase-based smt system .	our smt system is a phrase-based system based on the moses smt toolkit .
supervised methods shows that while supervised methods generally outperform the unsupervised ones .	we show that supervised methods outperform the unsupervised ones , while also being more efficient , computed on top of low-dimensional vectors .
we further used a 5-gram language model trained using the srilm toolkit with modified kneser-ney smoothing .	we use sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .
table 3 gives the results for the penn treebank converted with the head-finding rules of yamada and matsumoto and the labeling rules of nivre .	table 4 shows labeled and unlabeled accuracy scores of previous work reported for the penn2malt conversion with the head finding rules of yamada and matsumoto .
we use the sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .	we first trained a trigram bnlm as the baseline with interpolated kneser-ney smoothing , using srilm toolkit .
for the training of the drank and fixrank models , we utilised svm rank .	for the ranker we used svm rank , an efficient implementation for training ranking svms .
semantic role labeling ( srl ) is the process of extracting simple event structures , i.e. , “ who ” did “ what ” to “ whom ” , “ when ” and “ where ” .	semantic role labeling ( srl ) is the task of identifying semantic arguments of predicates in text .
whilst , the parameters for the maximum entropy model are developed based on the minimum error rate training method .	the weights associated to feature functions are optimally combined using the minimum error rate training .
xing et al pre-defined a set of topics from an external corpus to guide the generation of the seq2seq model .	moreover , xing et al incorporated topic words into seq2seq frameworks , where topic words are obtained from a pre-trained l-da model .
and show that our dependency language model provides improvements on five different test sets , with an overall gain of 0 . 92 in ter and 0 . 45 in bleu scores .	our results show that augmenting a state-of-the-art phrase-based system with this dependency language model leads to significant improvements in ter ( 0.92 % ) and bleu ( 0.45 % ) scores on five nist chinese-english evaluation test sets .
suchanek et al regarded the heading of a wikipedia article as a hyponym and obtained category labels attached to the article as its hypernym candidates .	suchanek et al extracted hyponymy relations from the category pages in the wikipedia using wordnet information .
sun and wan proposed a structure-based stacking model , which makes use of structured features such as sub-words for model combination .	sun and wan further extend the guide-feature method and propose a more complex sub-word stacking approach .
decoding paths adopted by other mt systems , this framework achieves better translation quality with much less re-decoding time .	thanks to the refined translation models , this approach produces better translations with a much shorter re-decoding time .
questions show that a discriminatively trained preference rank model is able to outperform alternative approaches designed for the same task .	a discriminative preference ranking model with a preference for appropriate answers is trained and applied to unseen questions .
we use a conditional random field sequence model , which allows for globally optimal training and decoding .	we use an information extraction tool for named entity recognition based on conditional random fields .
furthermore , tang et al proposed a new neural network approach called sswe to train sentimentaware word representation .	tang et al , proposed a method to learn sentiment specific word embeddings from tweets with emoticons as distantsupervised corpora without any manual annotation .
we use the berkeley probabilistic parser to obtain syntactic trees for english and its bonsai adaptation for french .	we use the berkeley probabilistic parser to obtain syntactic trees for english and its adapted version for french .
cover-based method guarantees that all bursty n-grams including irregularly-formed ones must be covered by extracted bursty phrases .	the proposed set cover-based method finds a minimum set of bursty phrases that cover all bursty n-grams including incomplete ones .
in the first part of the paper a novel , sortally-based approach to aspectual composition .	the first part of the paper develops a novel , sortally-based approach to the problem of aspectual composition .
for language model , we used sri language modeling toolkit to train a 4-gram model with modified kneser-ney smoothing .	we train a kn-smoothed 5-gram language model on the target side of the parallel training data with srilm .
we used trigram language models with interpolated kneser-kney discounting trained using the sri language modeling toolkit .	we use the sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .
the neural network for greedy training is based on the neural networks of chen and manning and .	our base model is a transition-based neural parser of chen and manning .
each context consists of approximately a paragraph of surrounding text , where the word to be discriminated ( the target word ) is found approximately in the middle of the context .	each context consists of several sentences that use a single sense of a target word , where at least one sentence contains the word .
corpus offers two improvements over current resources .	our corpus offers two main contributions .
supervised methods shows that while supervised methods generally outperform the unsupervised ones , the former are sensitive to the distribution of training instances , hurting their reliability .	as a consequence , supervised methods are sensitive to the distribution of examples in a particular dataset , making them less reliable for real-world applications .
finite-state head transducers produces implementations that are much more efficient than those for the ibm model .	the resulting finite-state machines are more expressive than standard leftto-right transducers .
the language model used was a 5-gram with modified kneserney smoothing , built with srilm toolkit .	we used a 5-gram language model with modified kneser-ney smoothing , built with the srilm toolkit .
for example , turian et al used word embeddings as input features for several nlp systems , including a traditional chunking system based on conditional random fields .	turian et al , for example , used embeddings from existing language models as unsupervised lexical features to improve named entity recognition and chunking .
information extraction ( ie ) is a task of identifying 憽甪acts挕 ? ( entities , relations and events ) within unstructured documents , and converting them into structured representations ( e.g. , databases ) .	information extraction ( ie ) is the task of extracting information from natural language texts to fill a database record following a structure called a template .
for the classifiers we use the scikit-learn machine learning toolkit .	we use the selectfrommodel 4 feature selection method as implemented in scikit-learn .
twitter is a huge microblogging service with more than 500 million tweets per day from different locations of the world and in different languages ( cite-p-10-1-6 ) .	twitter is a communication platform which combines sms , instant messages and social networks .
distributional similarity is used in many proposals to find semantically related words .	a common approach to the automatic extraction of semantically related words is to use distributional similarity .
a context-free grammar ( cfg ) is a 4-tuple math-w-4-1-0-9 , where math-w-4-1-0-18 is the set of nonterminals , σ the set of terminals , math-w-4-1-0-31 the set of production rules and math-w-4-1-0-38 a set of starting nonterminals ( i.e . multiple starting nonterminals are possible ) .	a context-free grammar ( cfg ) is a tuple math-w-3-1-1-9 , where math-w-3-1-1-22 is a finite set of nonterminal symbols , math-w-3-1-1-31 is a finite set of terminal symbols disjoint from n , math-w-3-1-1-44 is the start symbol and math-w-3-1-1-52 is a finite set of rules .
in this work , we detailed the multiple choice questions in subject history of gaokao , present two different approaches to address them .	in this work , we detailed the gaokao history multiple choice questions ( gkhmc ) and proposed two different approaches to address them using various resources .
deep learning has been considered as a generic solution to domain adaptation , and transfer learning problems .	deep learning with knowledge transfer has been previously applied to sentiment analysis in the context of domain adaptation and cross-lingual applications .
the log-linear feature weights are tuned with minimum error rate training on bleu .	the feature weights 位 m are tuned with minimum error rate training .
we trained a 5-gram sri language model using the corpus supplied for this purpose by the shared task organizers .	we trained a specific language model using srilm from each of these corpora in order to estimate n-gram log-probabilities .
word sense disambiguation ( wsd ) is the problem of assigning a sense to an ambiguous word , using its context .	word sense disambiguation ( wsd ) is a key task in computational lexical semantics , inasmuch as it addresses the lexical ambiguity of text by making explicit the meaning of words occurring in a given context ( cite-p-18-3-10 ) .
our method returns an “ explanation ” consisting of sets of input and output tokens that are causally related .	our method returns an “ explanation ” consisting of groups of input-output tokens that are causally related .
we used srilm for training the 5-gram language model with interpolated modified kneser-ney discounting , .	we also use a 4-gram language model trained using srilm with kneser-ney smoothing .
for instance swales develops his notion of genre in academic and research settings , bathia in professional settings , and so on .	for instance swales develops his notion of genre in academic and research settings , bathia and trosborg in professional settings , yates and orlikowsky within organizational communication .
work , we developed an approach based on distributional semantics to check whether a word in an answer is similar enough to a word in the question to count as given .	in place of surface-based givenness checks , as a first step in this direction we developed an approach integrating distributional semantics to check whether a word in a sentence is similar enough to a word in the context to count as given .
we use 5-grams for all language models implemented using the srilm toolkit .	we use srilm with its default parameters for this purpose .
trigram language models were estimated using the sri language modeling toolkit with modified kneser-ney smoothing .	the language model was constructed using the srilm toolkit with interpolated kneser-ney discounting .
a 4-gram language model was trained on the target side of the parallel data using the srilm toolkit .	a 4-gram language model is trained on the xinhua portion of the gigaword corpus with the srilm toolkit .
riloff and wiebe performed pattern learning through bootstrapping while extracting subjective expressions .	riloff and wiebe extracted subjective expressions from sentences using a bootstrapping pattern learning process .
shang et al and serban et al apply the rnn-based general encoder-decoder framework to the open-domain dialogue response generation task .	serban et al propose a hierarchical recurrent encoder-decoder neural network to the open domain dialogue .
bidirectional lstm is an extension of traditional lst-m to train two lstms on the input sequence .	the charner model uses bidirectional stacked lstms to map character sequences to tag sequences .
this tree kernel was slightly generalized by culotta and sorensen to compute similarity between two dependency trees .	culotta used this kernel on dependency trees to train a svm classifier for relation extraction .
relation extraction ( re ) is the task of recognizing relationships between entities mentioned in text .	relation extraction is the task of predicting attributes and relations for entities in a sentence ( zelenko et al. , 2003 ; bunescu and mooney , 2005 ; guodong et al. , 2005 ) .
we used the phrase-based smt in moses 5 for the translation experiments .	we used moses with the default configuration for phrase-based translation .

mead is centroid based multi-document summarizer which generates summaries using cluster centroids produced by topic detection and tracking system .

mead is a centroid based multi document summarizer which generates summaries using cluster centroids produced by topic detection and tracking system .

brown et al present a hierarchical word clustering algorithm that can handle a large number of classes and a large vocabulary .

our first choice is the bottom-up agglomerative word clustering algorithm of brown et al , which derives a hierarchical clustering of words from unlabeled data .

in this paper , we present a greedy non-directional parsing algorithm which doesn ’ t need a fully connected parse and can learn from partial parses .

in this paper , we present a dependency parsing algorithm which can train on partial projected parses and can take rich syntactic information as features for learning .

in this paper , we propose a neural architecture for coherence assessment that can capture long range entity transitions .

following this tradition , in this paper we propose to neuralize the popular entity grid models .

word sense disambiguation ( wsd ) is a particular problem of computational linguistics which consists in determining the correct sense for a given ambiguous word .

word sense disambiguation ( wsd ) is a key enabling technology that automatically chooses the intended sense of a word in context .

the weights of the log-linear interpolation model were optimized via minimum error rate training on the ted development set , using 200 best translations at each tuning iteration .

the weights of the different feature functions were optimised by means of minimum error rate training on the 2013 wmt test set .

kalchbrenner et al show that a cnn for modeling sentences can achieve competitive results in polarity classification .

to capture the relation between words , kalchbrenner et al propose a novel cnn model with a dynamic k-max pooling .

kim et al proposed a convolutional module to process complex inputs for the problem of language modeling .

kim et al apply a simple convolutional neural network model , which uses character level inputs for word representations .

ranking methods based on importance scores are proposed for keyphrase extraction .

therefore , automatic keyphrase extraction is an important research task .

similarly , korhonen et al relied on the information bottleneck and subcategorisation frame types to induce soft verb clusters .

korhonen et al used verb-frame pairs to cluster verbs into levin-style semantic classes .

twitter is the medium where people post real time messages to discuss on the different topics , and express their sentiments .

twitter is a popular microblogging service , which , among other things , is used for knowledge sharing among friends and peers .

the language model is a large interpolated 5-gram lm with modified kneser-ney smoothing .

we use sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .

success , however , depends on a high-coverage dictionary .

this success rests on a high-coverage dictionary .

for the phrase based system , we use moses with its default settings .

in all submitted systems , we use the phrase-based moses decoder .

the rules were extracted using the pos tags generated by the treetagger .

the source and target sentences are tagged respectively using the treetagger and amira toolkits .

to encode the original sentences we used word2vec embeddings pre-trained on google news .

we used the pre-trained word embeddings that were learned using the word2vec toolkit on google news dataset .

for the information-access applications described above .

however , consider the interactive information-access application described above .

we use 100-dimension glove vectors which are pre-trained on a large twitter corpus and fine-tuned during training .

we use glove word embeddings , which are 50-dimension word vectors trained with a crawled large corpus with 840 billion tokens .

currently , recurrent neural network based models are widely used on natural language processing tasks for excellent performance .

recurrent neural network architectures have proven to be well suited for many natural language generation tasks .

and consequently , we propose a weakly supervised fully-bayesian approach to pos tagging , which relaxes the unrealistic assumption by automatically acquiring the lexicon from a small amount of pos-tagged data .

as a result , we investigated a weakly supervised fully-bayesian approach to pos tagging , which relaxes the unrealistic assumption by automatically acquiring the lexicon from a small amount of postagged data .

we use pre-trained 50 dimensional glove vectors 4 for word embeddings initialization .

we use 300d glove vectors trained on 840b tokens as the word embedding input to the lstm .

while automatic evaluation methods like bleu can be useful for estimating translation quality , a higher score is no guarantee of quality improvement .

aggregating evaluation methods like bleu give a useful overview of the quality of a translation , but they do not afford specific information and leave too many details to chance .

for building our ap e b2 system , we set a maximum phrase length of 7 for the translation model , and a 5-gram language model was trained using kenlm .

we calculated the language model probabilities using kenlm , and built a 5-gram language model from the english gigaword fifth edition .

which is a generalization of current perceptron-based reranking methods .

the subtree ranking approach is a generalization of the perceptron-based approach .

neural networks , working on top of conventional n-gram back-off language models , have been introduced in as a potential means to improve discrete language models .

model fitting for our model is based on the expectation-maximization algorithm .

we estimate the parameters by maximizingp using the expectation maximization algorithm .

hammarstr枚m and borin give an extensive overview of stateof-the-art unsupervised learning of morphology .

hammarstr枚m and borin presented a literature survey on unsupervised learning of morphology , including methods for learning morphological segmentation .

following the work of koo et al , we used a tagger trained on the training data to provide part-of-speech tags for the development and test sets , and used 10-way jackknifing to generate part-of-speech tags for the training set .

following koo et al , we used the mxpost tagger trained on the full training data to provide part-of-speech tags for the development and the test set , and we used 10-way jackknifing to generate tags for the training set .

word sense disambiguation ( wsd ) is a natural language processing ( nlp ) task in which the correct meaning ( sense ) of a word in a given context is to be determined .

word sense disambiguation ( wsd ) is the task to identify the intended sense of a word in a computational manner based on the context in which it appears ( cite-p-13-3-4 ) .

ahsan and kolachina introduce a hybrid mt system that utilised online mt engines for msmt .

mellebeek et al introduced a hybrid mt system that utilised online mt engines for msmt .

coreference resolution is the next step on the way towards discourse understanding .

coreference resolution is the task of clustering a set of mentions in the text such that all mentions in the same cluster refer to the same entity .

we are the first to tie figurative language to the social context in which it is produced and show its relation to internal and external .

to the best of our knowledge , this is the first time figurative language is tied to the social context in which it appears .

coreference resolution is the process of linking together multiple referring expressions of a given entity in the world .

coreference resolution is the task of partitioning a set of mentions ( i.e . person , organization and location ) into entities .

we show that our method disambiguates a significant proportion of subject-object ambiguities in german .

we show that a simple method disambiguates some subject-object ambiguities in german , while making few errors .

we parsed the corpus with rasp and with the stanford pcfg parser .

we extracted the features from the gigaword corpus , which was first parsed using the rasp parser .

the alignment template approach uses word classes rather than lexical items to model phrase translation .

the alignment template model enhanced phrasal generalizations by using words classes rather than the words themselves .

collection comprises 132 , 229 dialogues containing a total of 764 , 146 turns / utterances that have been extracted from 753 movies .

the collected dataset comprises 132,229 dialogues containing a total of 764,146 turns that have been extracted from 753 movies .

we use the scikit-learn toolkit as our underlying implementation .

we trained the five classifiers using the svm implementation in scikit-learn .

in this paper , we propose a novel task that is crucial and generic from the viewpoint of health surveillance .

in this paper , we propose a more generalized task setting for public surveillance .

wang et al use all amr concepts and relations that appear in the training set as possible parameters if they appear in any sentence containing the same lemma as 蟽 0 and 尾 .

wang et al use all concepts that occur in the training data in the same sentence as the lemma of the node , leading to hundreds or thousands of possible actions from some states .

the language model is trained and applied with the srilm toolkit .

it has been trained with the srilm toolkit on the target side of all the training data .

we propose a method for extracting semantic orientations of phrases ( pairs of an adjective and a noun .

we proposed a method for extracting semantic orientations of phrases ( pairs of an adjective and a noun ) .

semantic role labeling ( srl ) is defined as the task to recognize arguments for a given predicate and assign semantic role labels to them .

semantic role labeling ( srl ) is the process of producing such a markup .

which beyonds the capability of phrase-based mt , we extend the search-aware tuning framework from phrase-based mt to syntax-based mt , in particular the hierarchical phrase-based translation model .

we extend this approach from phrase-based translation to syntax-based translation by generalizing the evaluation metrics for partial translations to handle tree-structured derivations in a way inspired by inside-outside algorithm .

svms have been shown to be robust in classification tasks involving text where the dimensionality is high .

svms are frequently used for text classification and have been applied successfully to nli .

we have adopted a supervised approach , a svm polynomial kernel classifier trained with the data provided by the challenge .

we follow a supervised approach , exploiting a svm polynomial kernel classifier trained with the challenge data .

the model parameters are trained using minimum error-rate training .

minimum error rate training is applied to tune the cn weights .

conditional random fields are undirected graphical models trained to maximize a conditional probability .

conditional random fields are probabilistic models for labelling sequential data .

coreference resolution is the task of clustering a sequence of textual entity mentions into a set of maximal non-overlapping clusters , such that mentions in a cluster refer to the same discourse entity .

coreference resolution is the process of determining whether two expressions in natural language refer to the same entity in the world .

our baseline is a standard phrase-based smt system .

our smt system is a phrase-based system based on the moses smt toolkit .

supervised methods shows that while supervised methods generally outperform the unsupervised ones .

we show that supervised methods outperform the unsupervised ones , while also being more efficient , computed on top of low-dimensional vectors .

we further used a 5-gram language model trained using the srilm toolkit with modified kneser-ney smoothing .

we use sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .

table 3 gives the results for the penn treebank converted with the head-finding rules of yamada and matsumoto and the labeling rules of nivre .

table 4 shows labeled and unlabeled accuracy scores of previous work reported for the penn2malt conversion with the head finding rules of yamada and matsumoto .

we use the sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .

we first trained a trigram bnlm as the baseline with interpolated kneser-ney smoothing , using srilm toolkit .

for the training of the drank and fixrank models , we utilised svm rank .

for the ranker we used svm rank , an efficient implementation for training ranking svms .

semantic role labeling ( srl ) is the process of extracting simple event structures , i.e. , “ who ” did “ what ” to “ whom ” , “ when ” and “ where ” .

semantic role labeling ( srl ) is the task of identifying semantic arguments of predicates in text .

whilst , the parameters for the maximum entropy model are developed based on the minimum error rate training method .

the weights associated to feature functions are optimally combined using the minimum error rate training .

xing et al pre-defined a set of topics from an external corpus to guide the generation of the seq2seq model .

moreover , xing et al incorporated topic words into seq2seq frameworks , where topic words are obtained from a pre-trained l-da model .

and show that our dependency language model provides improvements on five different test sets , with an overall gain of 0 . 92 in ter and 0 . 45 in bleu scores .

our results show that augmenting a state-of-the-art phrase-based system with this dependency language model leads to significant improvements in ter ( 0.92 % ) and bleu ( 0.45 % ) scores on five nist chinese-english evaluation test sets .

suchanek et al regarded the heading of a wikipedia article as a hyponym and obtained category labels attached to the article as its hypernym candidates .

suchanek et al extracted hyponymy relations from the category pages in the wikipedia using wordnet information .

sun and wan proposed a structure-based stacking model , which makes use of structured features such as sub-words for model combination .

sun and wan further extend the guide-feature method and propose a more complex sub-word stacking approach .

decoding paths adopted by other mt systems , this framework achieves better translation quality with much less re-decoding time .

thanks to the refined translation models , this approach produces better translations with a much shorter re-decoding time .

questions show that a discriminatively trained preference rank model is able to outperform alternative approaches designed for the same task .

a discriminative preference ranking model with a preference for appropriate answers is trained and applied to unseen questions .

we use a conditional random field sequence model , which allows for globally optimal training and decoding .

we use an information extraction tool for named entity recognition based on conditional random fields .

furthermore , tang et al proposed a new neural network approach called sswe to train sentimentaware word representation .

tang et al , proposed a method to learn sentiment specific word embeddings from tweets with emoticons as distantsupervised corpora without any manual annotation .

we use the berkeley probabilistic parser to obtain syntactic trees for english and its bonsai adaptation for french .

we use the berkeley probabilistic parser to obtain syntactic trees for english and its adapted version for french .

cover-based method guarantees that all bursty n-grams including irregularly-formed ones must be covered by extracted bursty phrases .

the proposed set cover-based method finds a minimum set of bursty phrases that cover all bursty n-grams including incomplete ones .

in the first part of the paper a novel , sortally-based approach to aspectual composition .

the first part of the paper develops a novel , sortally-based approach to the problem of aspectual composition .

for language model , we used sri language modeling toolkit to train a 4-gram model with modified kneser-ney smoothing .

we train a kn-smoothed 5-gram language model on the target side of the parallel training data with srilm .

we used trigram language models with interpolated kneser-kney discounting trained using the sri language modeling toolkit .

we use the sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .

the neural network for greedy training is based on the neural networks of chen and manning and .

our base model is a transition-based neural parser of chen and manning .

each context consists of approximately a paragraph of surrounding text , where the word to be discriminated ( the target word ) is found approximately in the middle of the context .

each context consists of several sentences that use a single sense of a target word , where at least one sentence contains the word .

corpus offers two improvements over current resources .

our corpus offers two main contributions .

supervised methods shows that while supervised methods generally outperform the unsupervised ones , the former are sensitive to the distribution of training instances , hurting their reliability .

as a consequence , supervised methods are sensitive to the distribution of examples in a particular dataset , making them less reliable for real-world applications .

finite-state head transducers produces implementations that are much more efficient than those for the ibm model .

the resulting finite-state machines are more expressive than standard leftto-right transducers .

the language model used was a 5-gram with modified kneserney smoothing , built with srilm toolkit .

we used a 5-gram language model with modified kneser-ney smoothing , built with the srilm toolkit .

for example , turian et al used word embeddings as input features for several nlp systems , including a traditional chunking system based on conditional random fields .

turian et al , for example , used embeddings from existing language models as unsupervised lexical features to improve named entity recognition and chunking .

information extraction ( ie ) is a task of identifying 憽甪acts挕 ? ( entities , relations and events ) within unstructured documents , and converting them into structured representations ( e.g. , databases ) .

information extraction ( ie ) is the task of extracting information from natural language texts to fill a database record following a structure called a template .

for the classifiers we use the scikit-learn machine learning toolkit .

we use the selectfrommodel 4 feature selection method as implemented in scikit-learn .

twitter is a huge microblogging service with more than 500 million tweets per day from different locations of the world and in different languages ( cite-p-10-1-6 ) .

twitter is a communication platform which combines sms , instant messages and social networks .

distributional similarity is used in many proposals to find semantically related words .

a common approach to the automatic extraction of semantically related words is to use distributional similarity .

a context-free grammar ( cfg ) is a 4-tuple math-w-4-1-0-9 , where math-w-4-1-0-18 is the set of nonterminals , σ the set of terminals , math-w-4-1-0-31 the set of production rules and math-w-4-1-0-38 a set of starting nonterminals ( i.e . multiple starting nonterminals are possible ) .

a context-free grammar ( cfg ) is a tuple math-w-3-1-1-9 , where math-w-3-1-1-22 is a finite set of nonterminal symbols , math-w-3-1-1-31 is a finite set of terminal symbols disjoint from n , math-w-3-1-1-44 is the start symbol and math-w-3-1-1-52 is a finite set of rules .

in this work , we detailed the multiple choice questions in subject history of gaokao , present two different approaches to address them .

in this work , we detailed the gaokao history multiple choice questions ( gkhmc ) and proposed two different approaches to address them using various resources .

deep learning has been considered as a generic solution to domain adaptation , and transfer learning problems .

deep learning with knowledge transfer has been previously applied to sentiment analysis in the context of domain adaptation and cross-lingual applications .

the log-linear feature weights are tuned with minimum error rate training on bleu .

the feature weights 位 m are tuned with minimum error rate training .

we trained a 5-gram sri language model using the corpus supplied for this purpose by the shared task organizers .

we trained a specific language model using srilm from each of these corpora in order to estimate n-gram log-probabilities .

word sense disambiguation ( wsd ) is the problem of assigning a sense to an ambiguous word , using its context .

word sense disambiguation ( wsd ) is a key task in computational lexical semantics , inasmuch as it addresses the lexical ambiguity of text by making explicit the meaning of words occurring in a given context ( cite-p-18-3-10 ) .

our method returns an “ explanation ” consisting of sets of input and output tokens that are causally related .

our method returns an “ explanation ” consisting of groups of input-output tokens that are causally related .

we used srilm for training the 5-gram language model with interpolated modified kneser-ney discounting , .

we also use a 4-gram language model trained using srilm with kneser-ney smoothing .

for instance swales develops his notion of genre in academic and research settings , bathia in professional settings , and so on .

for instance swales develops his notion of genre in academic and research settings , bathia and trosborg in professional settings , yates and orlikowsky within organizational communication .

work , we developed an approach based on distributional semantics to check whether a word in an answer is similar enough to a word in the question to count as given .

in place of surface-based givenness checks , as a first step in this direction we developed an approach integrating distributional semantics to check whether a word in a sentence is similar enough to a word in the context to count as given .

we use 5-grams for all language models implemented using the srilm toolkit .

we use srilm with its default parameters for this purpose .

trigram language models were estimated using the sri language modeling toolkit with modified kneser-ney smoothing .

the language model was constructed using the srilm toolkit with interpolated kneser-ney discounting .

a 4-gram language model was trained on the target side of the parallel data using the srilm toolkit .

a 4-gram language model is trained on the xinhua portion of the gigaword corpus with the srilm toolkit .

riloff and wiebe performed pattern learning through bootstrapping while extracting subjective expressions .

riloff and wiebe extracted subjective expressions from sentences using a bootstrapping pattern learning process .

shang et al and serban et al apply the rnn-based general encoder-decoder framework to the open-domain dialogue response generation task .

serban et al propose a hierarchical recurrent encoder-decoder neural network to the open domain dialogue .

bidirectional lstm is an extension of traditional lst-m to train two lstms on the input sequence .

the charner model uses bidirectional stacked lstms to map character sequences to tag sequences .

this tree kernel was slightly generalized by culotta and sorensen to compute similarity between two dependency trees .

culotta used this kernel on dependency trees to train a svm classifier for relation extraction .

relation extraction ( re ) is the task of recognizing relationships between entities mentioned in text .

relation extraction is the task of predicting attributes and relations for entities in a sentence ( zelenko et al. , 2003 ; bunescu and mooney , 2005 ; guodong et al. , 2005 ) .

we used the phrase-based smt in moses 5 for the translation experiments .

we used moses with the default configuration for phrase-based translation .