HHousen/ParaSCI · Datasets at Hugging Face

sentence1 stringlengths 16 446	sentence2 stringlengths 14 436
lda is a probabilistic model that can be used to model and discover underlying topic structures of documents .	lda is a generative model that learns a set of latent topics for a document collection .
we build a state of the art phrase-based smt system using moses .	we use the moses toolkit to train our phrase-based smt models .
we use byte pair encoding with 45k merge operations to split words into subwords .	for nmt , we applied byte pair encoding to split word into subword segments for both source and target languages .
in this paper , we studied how to modify an lstm model for deletion-based sentence compression .	in this paper , we focus on deletion-based sentence compression , which is a spacial case of extractive sentence compression .
the language model used in our paraphraser and the clarke and lapata baseline system is a kneser-ney discounted 5-gram model estimated on the gigaword corpus using the srilm toolkit .	the language model is a trigram-based backoff language model with kneser-ney smoothing , computed using srilm and trained on the same training data as the translation model .
conditional random fields have been successfully applied to several ie tasks in the past .	conditional random fields are popular models for many nlp tasks .
mikolov et al further proposed continuous bagof-words and skip-gram models , which use a simple single-layer architecture based on inner product between two word vectors .	the skip-gram and continuous bag-of-words models of mikolov et al propose a simple single-layer architecture based on the inner product between two word vectors .
our system is based on the phrase-based part of the statistical machine translation system moses .	we adapted the moses phrase-based decoder to translate word lattices .
in order to efficiently train parameters , we apply a reparameterization technique ( cite-p-22-3-6 , cite-p-22-1-10 ) .	in order to efficiently train parameters , we apply a reparameterization technique ( cite-p-22-3-6 , cite-p-22-1-10 ) on the variational lower bound .
the dataset used was a 1 million sentence aligned english-french corpus , taken from the europarl corpus .	the german-to-english baseline phrasebased system was trained on the europarl v7 corpus .
erkan et al defines similarity functions based on cosine similarity and edit distance between dependency paths , and then incorporate them in svm and knn learning for ppi extraction .	erkan et al first define two similarity functions based on cosine similarity and edit distance among dependency paths between two entities , and then incorporate them in semi-supervised learning for ppi extraction using svm and knn classifiers .
blitzer et al investigate domain adaptation for pos tagging using the method of structural correspondence learning .	blitzer et al proposed a structural correspondence learning method for domain adaptation and applied it to part-of-speech tagging .
in a unified framework , our model provides an effective way to capture context information at different levels for better lexical selection in smt .	the translation probabilities derived from our model are integrated into smt to allow collective lexical selection with both local and global informtion .
the model weights were trained using the minimum error rate training algorithm .	the parameter weights are optimized with minimum error rate training .
we use moses to train our phrasebased statistical mt system using the same parallel text as the nmt model , with the addition of common crawl , 10 for phrase extraction .	we used the phrase-based smt model , as implemented in the moses toolkit , to train an smt system translating from english to arabic .
using machine translation tools , we use the bidirectional lstm network to model the documents in both of the source and the target languages .	we use the bilingual bidirectional lstms to model the word sequences in the source and target languages .
liu et al suggested incorporating additional network architectures to further improve the performance of sdp-based methods , which uses a recursive neural network to model the sub-tree .	liu et al proposed a recursive neural network designed to model the subtrees , and cnn to capture the most important features on the shortest dependency path .
for data preparation and processing we use scikit-learn .	for the classifiers we use the scikit-learn machine learning toolkit .
sites show that the two new techniques enable classification algorithms to significantly improve the accuracy of the current state-of-the-art techniques .	they help achieve significantly higher accuracy than the current state-of-the-art techniques and systems .
we used the moses toolkit to build mt systems using various alignments .	we preprocessed the training corpora with scripts included in the moses toolkit .
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 .	since coreference resolution is a pervasive discourse phenomenon causing performance impediments in current ie systems , we considered a corpus of aligned english and romanian texts to identify coreferring expressions .
this resource was a commissioned translation of the basic traveling expression corpus sentences from english and french to the different dialects .	this resource was created as a commissioned translation of the basic traveling expression corpus sentences from english and french to the different dialects .
burkett and klein adopted a transformation-based method to learn a sequence of monolingual tree transformations for translation .	burkett and klein utilized a transformation-based method to learn a sequence of monolingual tree transformations for translation .
the latter proposed by the conll-2008 shared task is also called semantic dependency parsing , which annotates the heads of arguments rather than phrasal arguments .	the conll 2008-2009 shared tasks introduced a variant where semantic dependencies are annotated rather than phrasal arguments .
for part-of-speech tagging of the sentences , we used stanford pos tagger .	we use the stanford pos tagger for english and french to tag all sentence pairs .
the target language model is built on the target side of the parallel data with kneser-ney smoothing using the irstlm tool .	we have used the srilm with kneser-ney smoothing for training a language model for the first stage of decoding .
the penn discourse tree bank is the largest resource to date that provides a discourse annotated corpus in english .	the penn discourse treebank is the largest corpus richly annotated with explicit and implicit discourse relations and their senses .
in this task , we used conditional random fields .	we selected conditional random fields as the baseline model .
hovy et al utilized hypernyms and synonyms in wordnet to expand queries for increasing recall .	hovy et al utilized wordnet hypernyms and synonyms to expand queries to increase recall .
word sense disambiguation along with the lexical senses from wordnet are used for this task .	word sense disambiguation is performed using babelnet with the wordnet sense inventory .
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 .	relation extraction ( re ) is the process of generating structured relation knowledge from unstructured natural language texts .
we use the webquestions dataset as our main dataset , which contains 5,810 question-answer pairs .	we use the webquestions dataset , which contains 5,810 question-answer pairs .
the parsing uses parsito , which is a transition-based parser using a neural-network classifier .	the final parsing step is performed using parsito , which is a transitionbased parser with a neural-network classifier .
dependency parsing is a very important nlp task and has wide usage in different tasks such as question answering , semantic parsing , information extraction and machine translation .	dependency parsing is a crucial component of many natural language processing systems , for tasks such as text classification ( o ? zgu ? r and gu ? ngo ? r , 2010 ) , statistical machine translation ( cite-p-13-3-0 ) , relation extraction ( cite-p-13-1-1 ) , and question answering ( cite-p-13-1-3 ) .
coreference resolution is the process of finding discourse entities ( markables ) referring to the same real-world entity or concept .	coreference resolution is the task of partitioning a set of mentions ( i.e . person , organization and location ) into entities .
word alignment is equivalent to orthogonal non-negative matrix factorisation .	here , we view word alignment as matrix factorisation .
most recently , bansal and klein improved the berkeley parser by using surface counts from google n-grams .	more recently , bansal and klein proposed features for both dependency and constituency parsing based on web counts from the google n-grams corpus .
we have used the srilm with kneser-ney smoothing for training a language model for the first stage of decoding .	we use a pbsmt model where the language model is a 5-gram lm with modified kneser-ney smoothing .
morante and daelemans present a machine-learning approach to this task , using token-level , lexical information only .	morante and daelemans use the bioscope corpus to approach the problem of identifying cues and scopes via supervised machine learning .
relation extraction is a challenging task in natural language processing .	relation extraction is the problem of populating a target relation ( representing an entity-level relationship or attribute ) with facts extracted from natural-language text .
and applying a set of constraints , we restrict the space of possible tts templates under consideration , while still allowing new and more accurate templates to emerge from the training data .	in this paper , we explore methods for restricting the space of possible tts templates under consideration , while still allowing good templates to emerge directly from the data as much as possible .
the n-gram language models are trained using the srilm toolkit or similar software developed at hut .	uedin has used the srilm toolkit to train the language model and relies on kenlm for language model scoring during decoding .
this type of data has been found to yield the best correlation with eye-tracking data when different styles of presentation were compared for english .	this had the best correlation with eye-tracking data when different styles of presentation were compared for english .
for these experiments we use a maximum entropy classifier using the liblinear toolkit 2 .	for these experiments we use a maximum entropy classifier using the liblinear toolkit 1 .
sentiment analysis is a technique to classify documents based on the polarity of opinion expressed by the author of the document ( cite-p-16-1-13 ) .	sentiment analysis is the task in natural language processing ( nlp ) that deals with classifying opinions according to the polarity of the sentiment they express .
in this paper , we adopt continuous bag-of-word in word2vec as our context-based embedding model .	we revisit skip-gram model , as one of the most popular context-based embedding approaches .
the combiner we use here is implemented using a rule-based classifier , ripper .	in order to extract rules from the annotated data , we use a rule-based classifier , ripper .
attentional state properties modeled by centering can account for these differences .	we describe how the attentional state properties modeled by centering can account for these differences .
freeparser uses a domain-independent architecture to automatically identify sentences relevant to each new database .	using a self-supervised architecture , freeparser automatically labels these sentences , and then trains a semantic parser for all of freebase .
firstly , we explicitly show that concept-drift is pervasive and serious in real bug report streams .	we demonstrate that concept drift is a real , pervasive issue for learning from issue tracker streams .
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 popular microblogging service , which , among other things , is used for knowledge sharing among friends and peers .
however , due to their heterogeneous characteristics , mwes present a tough challenge for both linguistic and computational work .	because of their frequency and their peculiar behaviour , mwes pose a great challenge to the creation of natural language processing systems .
all system component weights were tuned using minimum error-rate training , with three tuning runs for each condition .	feature weights were set with minimum error rate training on a tuning set using bleu as the objective function .
in the project , we focus on content-related criteria .	in this work , we chose to start with criteria related to content choice .
we built a 5-gram language model on the english side of europarl and used the kneser-ney smoothing method and srilm as the language model toolkit .	we used the sri language modeling toolkit to train a fivegram model with modified kneser-ney smoothing .
in section 6 , the proposed word embeddings show evident improvements on sentiment classification , as compared to the base model .	in section 6 , the proposed word embeddings show evident improvements on sentiment classification , as compared to the base model word2vec and other baselines using the same lexical resource .
we base our extrinsic evaluation on the seminal work of collobert et al on the use of neural methods for nlp .	we compare the proposed model to our implementation of the iobes-based model described in collobert et al , applied to mwe tagging .
we use the 300-dimensional skip-gram word embeddings built on the google-news corpus .	for feature building , we use word2vec pre-trained word embeddings .
we used the srilm toolkit to create 5-gram language models with interpolated modified kneser-ney discounting .	we apply sri language modeling toolkit to train a 4-gram language model with kneser-ney smoothing .
empirical experiments on a labeled set of words show that the proposed method outperforms the state of the art methods .	the method is experimentally tested using a manually labeled set of positive and negative words .
finin et al use amazons mechanical turk service 2 and crowdflower 3 to annotate named entities in tweets and train a crf model to evaluate the effectiveness of human labeling .	finin et al use amazons mechanical turk service 3 and crowdflower 4 to annotate named entities in tweets and train a crf model to evaluate the effectiveness of human labeling .
we applied liblinear via its scikitlearn python interface to train the logistic regression model with l2 regularization .	we use the wrapper of the scikit learn python library over the liblinear logistic regression implementation .
we further used a 5-gram language model trained using the srilm toolkit with modified kneser-ney smoothing .	we trained a standard 5-gram language model with modified kneser-ney smoothing using the kenlm toolkit on 4 billion running words .
r the brown corpus is in some sense fundamentally more difficult .	r the brown corpus is in some sense fundamentally more difficult for this problem .
apart from the original space of features , we have used the so called svd features , obtained from the projection of the feature vectors into the reduced space .	apart from the original space of features , we have the so called svd features , obtained from the projection of the feature vectors into the reduced space .
udpipe 1 . 2 participated in the shared task , placing as the 8th best system , while achieving low running times .	udpipe 1.1 provided a strong baseline for the task , placing as the 13 th ( out of 33 ) best system in the official ranking .
to train the link embeddings , we use the speedy , skip-gram neural language model of mikolov et al via their toolkit word2vec .	we use the perplexity computation method of mikolov et al suitable for skip-gram models .
we use srilm toolkits to train two 4-gram language models on the filtered english blog authorship corpus and the xinhua portion of gigaword corpus , respectively .	we use the sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus .
agate is an acronym for general architecture for text engineering .	ttea is an acronym for text engineering architecture .
in this context , we present a method based on character p-grams that we designed for the arabic dialect identification shared task of the dsl 2016 challenge .	we have presented a method based on character p-grams for the arabic dialect identification shared task of the dsl 2016 challenge .
ucca ¡¯ s representation is guided by conceptual notions and has its roots in the cognitive linguistics tradition .	ucca is supported by extensive typological cross-linguistic evidence and accords with the leading cognitive linguistics theories .
and , based on theoretical and empirical desiderata , we outline a more comprehensive framework to model the acquisition of allophonic rules .	therefore , the main extension towards a comprehensive model of the acquisition of allophonic rules would be to include acoustic indicators .
we compute statistical significance using the approximate randomization test .	we test this hypothesis with an approximate randomization approach .
automated essay scoring utilizes natural language processing and machine learning techniques to automatically rate essays written for a target prompt .	automated essay scoring utilizes the nlp techniques to automatically rate essays written for given prompts , namely , essay topics , in an educational setting .
krishnakumaran and zhu use the isa relation in wordnet for metaphor recognition .	krishnakumaran and zhu use hyponymy relation in wordnet to detect semantic violations .
borrowing is a major type of word formation in japanese , and numerous foreign words ( proper names or neologisms etc . ) are continuously being imported from other languages ( cite-p-26-3-22 ) .	borrowing is the pervasive linguistic phenomenon of transferring and adapting linguistic constructions ( lexical , phonological , morphological , and syntactic ) from a “ donor ” language into a “ recipient ” language ( cite-p-10-3-16 ) .
many existing active learning methods are to select the most uncertain examples using various measures .	many existing active learning methods are based on selecting the most uncertain examples using various measures .
coreference resolution is a key task in natural language processing ( cite-p-13-1-8 ) aiming to detect the referential expressions ( mentions ) in a text that point to the same entity .	coreference resolution is a challenging task , that involves identification and clustering of noun phrases mentions that refer to the same real-world entity .
contrary to both intuition and past conclusions , results show no significant evidence of reference bias .	results showed no significant evidence of reference bias , contrary to prior reports and intuition .
named entity linking is the task of mapping mentions of named entities to their canonical reference in a knowledge base .	entity resolution is the task of mapping mentions of entities in text to corresponding records in a knowledge base .
vector representations of words and phrases have been successfully applied in many natural language processing tasks .	distributed representations of words have been widely used in many natural language processing tasks .
the lingo grammar matrix is situated theoretically within head-driven phrase structure grammar , a lexicalist , constraint-based framework .	this model is inspired by formalisms based on structural features like head-driven phrase structure grammar .
this means in practice that the language model was trained using the srilm toolkit .	uedin has used the srilm toolkit to train the language model and relies on kenlm for language model scoring during decoding .
for all models , we use fixed pre-trained glove vectors and character embeddings .	for word-level embedding e w , we utilize pre-trained , 300-dimensional embedding vectors from glove 6b .
we used the scikit-learn python machine learning library to implement the feature extraction pipeline and the support vector machine classifier .	we used scikit-learn 4 for more details , a machine learning library for python , to build a question classifier based on the svm algorithm and linear kernel function .
for the automatic evaluation , we used the bleu metric from ibm .	we report the mt performance using the original bleu metric .
by the base form of the head verb , we achieve a better statistical word alignment performance , and are able to better estimate the translation model and generalize to unseen verb forms during translation .	this leads to an improved statistical word alignment performance , and has the advantages of improving the translation model and generalizing to unseen verb forms , during translation .
we substitute our language model and use mert to optimize the bleu score .	we use case-sensitive bleu-4 to measure the quality of translation result .
since its introduction , topic modeling has been tailored to perform better on short texts such as microblogs .	topic models , such as plsa and lda , have shown great success in discovering latent topics in text collections .
we use the stanford nlp pos tagger to generate the tagged text .	we use the stanford part of speech tagger to annotate each word with its pos tag .
for language model , we use a trigram language model trained with the srilm toolkit on the english side of the training corpus .	we use srilm for training a trigram language model on the english side of the training corpus .
sentiment analysis is the task of identifying positive and negative opinions , sentiments , emotions and attitudes expressed in text .	sentiment analysis is a growing research field , especially on web social networks .
recently , attentive neural networks have shown success in several nlp tasks such as machine translation , image captioning , speech recognition and document classification .	recently , rnns with attention mechanisms have demonstrated success in various nlp tasks , such as machine translation , parsing , image captioning , and textual entailment .
relation extraction ( re ) is the task of identifying instances of relations , such as nationality ( person , country ) or place of birth ( person , location ) , in passages of natural text .	relation extraction ( re ) has been defined as the task of identifying a given set of semantic binary relations in text .
that word ( containing at least one character ) is the appropriate unit for chinese language processing .	word is usually adopted as the smallest unit in most tasks of chinese language processing .
hasegawa et al , 2004 ) use ner to identify frequently co-occurring entities as likely relation phrases .	hasegawa et al 2004 , used large corpora and an extended named entity tagger to find novel relations and their participants .
we also use a 4-gram language model trained using srilm with kneser-ney smoothing .	we use 5-gram models with modified kneser-ney smoothing and interpolated back-off .
since our dataset is not so large , we make use of pre-trained word embeddings , which are trained on a much larger corpus with word2vec toolkit .	in all of our experiments , the word embeddings are trained using word2vec on the wikipedia corpus .
we have used the srilm with kneser-ney smoothing for training a language model for the first stage of decoding .	we choose modified kneser ney as the smoothing algorithm when learning the ngram model .
long short-term memory is a special type of rnn that leverages multiple gate vectors and a memory cell vector to solve the vanishing and exploding gradient problems of training rnns .	long short term memory is a variant of recurrent neural network , which enables to address the gradient vanishing and exploding problems in rnn via introducing gate mechanism and memory cell .

lda is a probabilistic model that can be used to model and discover underlying topic structures of documents .

lda is a generative model that learns a set of latent topics for a document collection .

we build a state of the art phrase-based smt system using moses .

we use the moses toolkit to train our phrase-based smt models .

we use byte pair encoding with 45k merge operations to split words into subwords .

for nmt , we applied byte pair encoding to split word into subword segments for both source and target languages .

in this paper , we studied how to modify an lstm model for deletion-based sentence compression .

in this paper , we focus on deletion-based sentence compression , which is a spacial case of extractive sentence compression .

the language model used in our paraphraser and the clarke and lapata baseline system is a kneser-ney discounted 5-gram model estimated on the gigaword corpus using the srilm toolkit .

the language model is a trigram-based backoff language model with kneser-ney smoothing , computed using srilm and trained on the same training data as the translation model .

conditional random fields have been successfully applied to several ie tasks in the past .

conditional random fields are popular models for many nlp tasks .

mikolov et al further proposed continuous bagof-words and skip-gram models , which use a simple single-layer architecture based on inner product between two word vectors .

the skip-gram and continuous bag-of-words models of mikolov et al propose a simple single-layer architecture based on the inner product between two word vectors .

our system is based on the phrase-based part of the statistical machine translation system moses .

we adapted the moses phrase-based decoder to translate word lattices .

in order to efficiently train parameters , we apply a reparameterization technique ( cite-p-22-3-6 , cite-p-22-1-10 ) .

in order to efficiently train parameters , we apply a reparameterization technique ( cite-p-22-3-6 , cite-p-22-1-10 ) on the variational lower bound .

the dataset used was a 1 million sentence aligned english-french corpus , taken from the europarl corpus .

the german-to-english baseline phrasebased system was trained on the europarl v7 corpus .

erkan et al defines similarity functions based on cosine similarity and edit distance between dependency paths , and then incorporate them in svm and knn learning for ppi extraction .

erkan et al first define two similarity functions based on cosine similarity and edit distance among dependency paths between two entities , and then incorporate them in semi-supervised learning for ppi extraction using svm and knn classifiers .

blitzer et al investigate domain adaptation for pos tagging using the method of structural correspondence learning .

blitzer et al proposed a structural correspondence learning method for domain adaptation and applied it to part-of-speech tagging .

in a unified framework , our model provides an effective way to capture context information at different levels for better lexical selection in smt .

the translation probabilities derived from our model are integrated into smt to allow collective lexical selection with both local and global informtion .

the model weights were trained using the minimum error rate training algorithm .

the parameter weights are optimized with minimum error rate training .

we use moses to train our phrasebased statistical mt system using the same parallel text as the nmt model , with the addition of common crawl , 10 for phrase extraction .

we used the phrase-based smt model , as implemented in the moses toolkit , to train an smt system translating from english to arabic .

using machine translation tools , we use the bidirectional lstm network to model the documents in both of the source and the target languages .

we use the bilingual bidirectional lstms to model the word sequences in the source and target languages .

liu et al suggested incorporating additional network architectures to further improve the performance of sdp-based methods , which uses a recursive neural network to model the sub-tree .

liu et al proposed a recursive neural network designed to model the subtrees , and cnn to capture the most important features on the shortest dependency path .

for data preparation and processing we use scikit-learn .

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

sites show that the two new techniques enable classification algorithms to significantly improve the accuracy of the current state-of-the-art techniques .

they help achieve significantly higher accuracy than the current state-of-the-art techniques and systems .

we used the moses toolkit to build mt systems using various alignments .

we preprocessed the training corpora with scripts included in the moses toolkit .

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 .

since coreference resolution is a pervasive discourse phenomenon causing performance impediments in current ie systems , we considered a corpus of aligned english and romanian texts to identify coreferring expressions .

this resource was a commissioned translation of the basic traveling expression corpus sentences from english and french to the different dialects .

this resource was created as a commissioned translation of the basic traveling expression corpus sentences from english and french to the different dialects .

burkett and klein adopted a transformation-based method to learn a sequence of monolingual tree transformations for translation .

burkett and klein utilized a transformation-based method to learn a sequence of monolingual tree transformations for translation .

the latter proposed by the conll-2008 shared task is also called semantic dependency parsing , which annotates the heads of arguments rather than phrasal arguments .

the conll 2008-2009 shared tasks introduced a variant where semantic dependencies are annotated rather than phrasal arguments .

for part-of-speech tagging of the sentences , we used stanford pos tagger .

we use the stanford pos tagger for english and french to tag all sentence pairs .

the target language model is built on the target side of the parallel data with kneser-ney smoothing using the irstlm tool .

we have used the srilm with kneser-ney smoothing for training a language model for the first stage of decoding .

the penn discourse tree bank is the largest resource to date that provides a discourse annotated corpus in english .

the penn discourse treebank is the largest corpus richly annotated with explicit and implicit discourse relations and their senses .

in this task , we used conditional random fields .

we selected conditional random fields as the baseline model .

hovy et al utilized hypernyms and synonyms in wordnet to expand queries for increasing recall .

hovy et al utilized wordnet hypernyms and synonyms to expand queries to increase recall .

word sense disambiguation along with the lexical senses from wordnet are used for this task .

word sense disambiguation is performed using babelnet with the wordnet sense inventory .

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 .

relation extraction ( re ) is the process of generating structured relation knowledge from unstructured natural language texts .

we use the webquestions dataset as our main dataset , which contains 5,810 question-answer pairs .

we use the webquestions dataset , which contains 5,810 question-answer pairs .

the parsing uses parsito , which is a transition-based parser using a neural-network classifier .

the final parsing step is performed using parsito , which is a transitionbased parser with a neural-network classifier .

dependency parsing is a very important nlp task and has wide usage in different tasks such as question answering , semantic parsing , information extraction and machine translation .

dependency parsing is a crucial component of many natural language processing systems , for tasks such as text classification ( o ? zgu ? r and gu ? ngo ? r , 2010 ) , statistical machine translation ( cite-p-13-3-0 ) , relation extraction ( cite-p-13-1-1 ) , and question answering ( cite-p-13-1-3 ) .

coreference resolution is the process of finding discourse entities ( markables ) referring to the same real-world entity or concept .

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

word alignment is equivalent to orthogonal non-negative matrix factorisation .

here , we view word alignment as matrix factorisation .

most recently , bansal and klein improved the berkeley parser by using surface counts from google n-grams .

more recently , bansal and klein proposed features for both dependency and constituency parsing based on web counts from the google n-grams corpus .

we have used the srilm with kneser-ney smoothing for training a language model for the first stage of decoding .

we use a pbsmt model where the language model is a 5-gram lm with modified kneser-ney smoothing .

morante and daelemans present a machine-learning approach to this task , using token-level , lexical information only .

morante and daelemans use the bioscope corpus to approach the problem of identifying cues and scopes via supervised machine learning .

relation extraction is a challenging task in natural language processing .

relation extraction is the problem of populating a target relation ( representing an entity-level relationship or attribute ) with facts extracted from natural-language text .

and applying a set of constraints , we restrict the space of possible tts templates under consideration , while still allowing new and more accurate templates to emerge from the training data .

in this paper , we explore methods for restricting the space of possible tts templates under consideration , while still allowing good templates to emerge directly from the data as much as possible .

the n-gram language models are trained using the srilm toolkit or similar software developed at hut .

uedin has used the srilm toolkit to train the language model and relies on kenlm for language model scoring during decoding .

this type of data has been found to yield the best correlation with eye-tracking data when different styles of presentation were compared for english .

this had the best correlation with eye-tracking data when different styles of presentation were compared for english .

for these experiments we use a maximum entropy classifier using the liblinear toolkit 2 .

for these experiments we use a maximum entropy classifier using the liblinear toolkit 1 .

sentiment analysis is a technique to classify documents based on the polarity of opinion expressed by the author of the document ( cite-p-16-1-13 ) .

sentiment analysis is the task in natural language processing ( nlp ) that deals with classifying opinions according to the polarity of the sentiment they express .

in this paper , we adopt continuous bag-of-word in word2vec as our context-based embedding model .

we revisit skip-gram model , as one of the most popular context-based embedding approaches .

the combiner we use here is implemented using a rule-based classifier , ripper .

in order to extract rules from the annotated data , we use a rule-based classifier , ripper .

attentional state properties modeled by centering can account for these differences .

we describe how the attentional state properties modeled by centering can account for these differences .

freeparser uses a domain-independent architecture to automatically identify sentences relevant to each new database .

using a self-supervised architecture , freeparser automatically labels these sentences , and then trains a semantic parser for all of freebase .

firstly , we explicitly show that concept-drift is pervasive and serious in real bug report streams .

we demonstrate that concept drift is a real , pervasive issue for learning from issue tracker streams .

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

however , due to their heterogeneous characteristics , mwes present a tough challenge for both linguistic and computational work .

because of their frequency and their peculiar behaviour , mwes pose a great challenge to the creation of natural language processing systems .

all system component weights were tuned using minimum error-rate training , with three tuning runs for each condition .

feature weights were set with minimum error rate training on a tuning set using bleu as the objective function .

in the project , we focus on content-related criteria .

in this work , we chose to start with criteria related to content choice .

we built a 5-gram language model on the english side of europarl and used the kneser-ney smoothing method and srilm as the language model toolkit .

we used the sri language modeling toolkit to train a fivegram model with modified kneser-ney smoothing .

in section 6 , the proposed word embeddings show evident improvements on sentiment classification , as compared to the base model .

in section 6 , the proposed word embeddings show evident improvements on sentiment classification , as compared to the base model word2vec and other baselines using the same lexical resource .

we base our extrinsic evaluation on the seminal work of collobert et al on the use of neural methods for nlp .

we compare the proposed model to our implementation of the iobes-based model described in collobert et al , applied to mwe tagging .

we use the 300-dimensional skip-gram word embeddings built on the google-news corpus .

for feature building , we use word2vec pre-trained word embeddings .

we used the srilm toolkit to create 5-gram language models with interpolated modified kneser-ney discounting .

we apply sri language modeling toolkit to train a 4-gram language model with kneser-ney smoothing .

empirical experiments on a labeled set of words show that the proposed method outperforms the state of the art methods .

the method is experimentally tested using a manually labeled set of positive and negative words .

finin et al use amazons mechanical turk service 2 and crowdflower 3 to annotate named entities in tweets and train a crf model to evaluate the effectiveness of human labeling .

finin et al use amazons mechanical turk service 3 and crowdflower 4 to annotate named entities in tweets and train a crf model to evaluate the effectiveness of human labeling .

we applied liblinear via its scikitlearn python interface to train the logistic regression model with l2 regularization .

we use the wrapper of the scikit learn python library over the liblinear logistic regression implementation .

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

we trained a standard 5-gram language model with modified kneser-ney smoothing using the kenlm toolkit on 4 billion running words .

r the brown corpus is in some sense fundamentally more difficult .

r the brown corpus is in some sense fundamentally more difficult for this problem .

apart from the original space of features , we have used the so called svd features , obtained from the projection of the feature vectors into the reduced space .

apart from the original space of features , we have the so called svd features , obtained from the projection of the feature vectors into the reduced space .

udpipe 1 . 2 participated in the shared task , placing as the 8th best system , while achieving low running times .

udpipe 1.1 provided a strong baseline for the task , placing as the 13 th ( out of 33 ) best system in the official ranking .

to train the link embeddings , we use the speedy , skip-gram neural language model of mikolov et al via their toolkit word2vec .

we use the perplexity computation method of mikolov et al suitable for skip-gram models .

we use srilm toolkits to train two 4-gram language models on the filtered english blog authorship corpus and the xinhua portion of gigaword corpus , respectively .

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

agate is an acronym for general architecture for text engineering .

ttea is an acronym for text engineering architecture .

in this context , we present a method based on character p-grams that we designed for the arabic dialect identification shared task of the dsl 2016 challenge .

we have presented a method based on character p-grams for the arabic dialect identification shared task of the dsl 2016 challenge .

ucca ¡¯ s representation is guided by conceptual notions and has its roots in the cognitive linguistics tradition .

ucca is supported by extensive typological cross-linguistic evidence and accords with the leading cognitive linguistics theories .

and , based on theoretical and empirical desiderata , we outline a more comprehensive framework to model the acquisition of allophonic rules .

therefore , the main extension towards a comprehensive model of the acquisition of allophonic rules would be to include acoustic indicators .

we compute statistical significance using the approximate randomization test .

we test this hypothesis with an approximate randomization approach .

automated essay scoring utilizes natural language processing and machine learning techniques to automatically rate essays written for a target prompt .

automated essay scoring utilizes the nlp techniques to automatically rate essays written for given prompts , namely , essay topics , in an educational setting .

krishnakumaran and zhu use the isa relation in wordnet for metaphor recognition .

krishnakumaran and zhu use hyponymy relation in wordnet to detect semantic violations .

borrowing is a major type of word formation in japanese , and numerous foreign words ( proper names or neologisms etc . ) are continuously being imported from other languages ( cite-p-26-3-22 ) .

borrowing is the pervasive linguistic phenomenon of transferring and adapting linguistic constructions ( lexical , phonological , morphological , and syntactic ) from a “ donor ” language into a “ recipient ” language ( cite-p-10-3-16 ) .

many existing active learning methods are to select the most uncertain examples using various measures .

many existing active learning methods are based on selecting the most uncertain examples using various measures .

coreference resolution is a key task in natural language processing ( cite-p-13-1-8 ) aiming to detect the referential expressions ( mentions ) in a text that point to the same entity .

coreference resolution is a challenging task , that involves identification and clustering of noun phrases mentions that refer to the same real-world entity .

contrary to both intuition and past conclusions , results show no significant evidence of reference bias .

results showed no significant evidence of reference bias , contrary to prior reports and intuition .

named entity linking is the task of mapping mentions of named entities to their canonical reference in a knowledge base .

entity resolution is the task of mapping mentions of entities in text to corresponding records in a knowledge base .

vector representations of words and phrases have been successfully applied in many natural language processing tasks .

distributed representations of words have been widely used in many natural language processing tasks .

the lingo grammar matrix is situated theoretically within head-driven phrase structure grammar , a lexicalist , constraint-based framework .

this model is inspired by formalisms based on structural features like head-driven phrase structure grammar .

this means in practice that the language model was trained using the srilm toolkit .

uedin has used the srilm toolkit to train the language model and relies on kenlm for language model scoring during decoding .

for all models , we use fixed pre-trained glove vectors and character embeddings .

for word-level embedding e w , we utilize pre-trained , 300-dimensional embedding vectors from glove 6b .

we used the scikit-learn python machine learning library to implement the feature extraction pipeline and the support vector machine classifier .

we used scikit-learn 4 for more details , a machine learning library for python , to build a question classifier based on the svm algorithm and linear kernel function .

for the automatic evaluation , we used the bleu metric from ibm .

we report the mt performance using the original bleu metric .

by the base form of the head verb , we achieve a better statistical word alignment performance , and are able to better estimate the translation model and generalize to unseen verb forms during translation .

this leads to an improved statistical word alignment performance , and has the advantages of improving the translation model and generalizing to unseen verb forms , during translation .

we substitute our language model and use mert to optimize the bleu score .

we use case-sensitive bleu-4 to measure the quality of translation result .

since its introduction , topic modeling has been tailored to perform better on short texts such as microblogs .

topic models , such as plsa and lda , have shown great success in discovering latent topics in text collections .

we use the stanford nlp pos tagger to generate the tagged text .

we use the stanford part of speech tagger to annotate each word with its pos tag .

for language model , we use a trigram language model trained with the srilm toolkit on the english side of the training corpus .

we use srilm for training a trigram language model on the english side of the training corpus .

sentiment analysis is the task of identifying positive and negative opinions , sentiments , emotions and attitudes expressed in text .

sentiment analysis is a growing research field , especially on web social networks .

recently , attentive neural networks have shown success in several nlp tasks such as machine translation , image captioning , speech recognition and document classification .

recently , rnns with attention mechanisms have demonstrated success in various nlp tasks , such as machine translation , parsing , image captioning , and textual entailment .

relation extraction ( re ) is the task of identifying instances of relations , such as nationality ( person , country ) or place of birth ( person , location ) , in passages of natural text .

relation extraction ( re ) has been defined as the task of identifying a given set of semantic binary relations in text .

that word ( containing at least one character ) is the appropriate unit for chinese language processing .

word is usually adopted as the smallest unit in most tasks of chinese language processing .

hasegawa et al , 2004 ) use ner to identify frequently co-occurring entities as likely relation phrases .

hasegawa et al 2004 , used large corpora and an extended named entity tagger to find novel relations and their participants .

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

we use 5-gram models with modified kneser-ney smoothing and interpolated back-off .

since our dataset is not so large , we make use of pre-trained word embeddings , which are trained on a much larger corpus with word2vec toolkit .

in all of our experiments , the word embeddings are trained using word2vec on the wikipedia corpus .

we have used the srilm with kneser-ney smoothing for training a language model for the first stage of decoding .

we choose modified kneser ney as the smoothing algorithm when learning the ngram model .

long short-term memory is a special type of rnn that leverages multiple gate vectors and a memory cell vector to solve the vanishing and exploding gradient problems of training rnns .

long short term memory is a variant of recurrent neural network , which enables to address the gradient vanishing and exploding problems in rnn via introducing gate mechanism and memory cell .