text
stringlengths
82
736
label
int64
0
1
semantic role labeling ( srl ) consists of finding the arguments of a predicate and labeling them with semantic roles ( cite-p-9-1-5 , cite-p-9-3-0 )---semantic role labeling ( srl ) is defined as the task to recognize arguments for a given predicate and assign semantic role labels to them
1
in this paper , we make a description about our submission system for the task---however , in their further study , they reported even lower bleu scores after grouping mwes according to part-of-speech on a large corpus
0
we also show that the most successful error generation methods are those that use knowledge about the article distribution and error patterns observed in non-native text---we used data from the conll-x shared task on multilingual dependency parsing
0
article and preposition errors are the two main research topics---of all the errors , determiner and preposition errors are the two main research topics
1
we used the adam optimization function with default parameters---we used adam optimizer with its standard parameters
1
barzilay and mckeown extract paraphrases from a monolingual parallel corpus , containing multiple translations of the same source---barzilay and mckeown utilized multiple english translations of the same source text for paraphrase extraction
1
coreference resolution is the task of determining which mentions in a text are used to refer to the same real-world entity---coreference resolution is the task of grouping all the mentions of entities 1 in a document into equivalence classes so that all the mentions in a given class refer to the same discourse entity
1
we use the selectfrommodel 4 feature selection method as implemented in scikit-learn---we use a random forest classifier , as implemented in scikit-learn
1
word sense disambiguation ( wsd ) is the task of identifying the correct meaning of a word in context---translation-based approaches are based on the statistical translation models , including the ibm model
0
much of the additional work on generative modeling of 1-to-n word alignments is based on the hmm model---the dominant approach to word alignment has been the ibm models together with the hmm model
1
for example , in extended wordnet , the rich glosses in wordnet are enriched by disambiguating the nouns , verbs , adverbs , and adjectives with synsets---the english side of the parallel corpus is trained into a language model using srilm
0
cardie et al took advantage of opinion summarization to support multi-perspective question answering system which aims to extract opinion-oriented information of a question---our results indicate that a word-based approach is superior to syllable-or vowel-based detection
0
we trained a time-series generation policy for 10,000 runs using the tabular temporaldifference learning---we trained a te generation policy using the above user simulation model for 10,000 runs using the sarsa reinforcement learning algorithm
1
we compare the model against the moses phrase-based translation system , applied to phoneme sequences---we then evaluate the effect of word alignment on machine translation quality using the phrase-based translation system moses
1
we trained a 5-gram language model on the english side of each training corpus using the sri language modeling toolkit---zhang et al introduced a synchronous binarization technique that improved decoding efficiency and accuracy by ensuring that rule binarization avoided gaps on both the source and target sides
0
semantic role labeling ( srl ) has been defined as a sentence-level natural-language processing task in which semantic roles are assigned to the syntactic arguments of a predicate ( cite-p-14-1-7 )---relation extraction ( re ) is a task of identifying typed relations between known entity mentions in a sentence
0
we specify a non-stochastic version of the formalism , noting that probabilities may be attached to the rewrite rules exactly as in stochastic cfg---ionescu et al propose a combination of several string kernels and use multiple kernel learning
0
the common inventory incorporates some of the general relation types defined by gildea and jurafsky for their experiments in classifying semantic relations in framenet using a reduced inventory---the common inventory incorporates some of the general relation types defined by gildea and jurafsky for their experiments in classifying semantic relations in framenet using a reduced relation inventory
1
generally , it has been found that tense high vowels have longer vots than lax low vowels---it is now widely accepted that tense high vowels are correlated with longer vots than lax low vowels
1
in the task-6 results ( cite-p-15-1-4 ) , run2 was ranked 72th out of 85 participants with 0.4169 pearson-correlation all competition rank---n can be done using minimum error rate training on a development set of input sentences and their reference translations
0
our sampling method always gives preference to that example which maximizes training utility---for our baseline , we used a small parallel corpus of 30k english-spanish sentences from the europarl corpus
0
faruqui et al introduce a graph-based retrofitting method where they post-process learned vectors with respect to semantic relationships extracted from additional lexical resources---faruqui et al employ semantic relations of ppdb , wordnet , framenet to retrofit word embeddings for various prediction tasks
1
the composite kernel consists of an entity kernel and a convolution parse tree kernel---biadsy et al present a system that identifies dialectal words in speech and their dialect of origin through the acoustic signals
0
this model is also ¡®row-less¡¯ and does not directly model entities or entity pairs---that is ¡® row-less ¡¯ having no explicit parameters for entity pairs and entities
1
xing et al incorporated the topic information from an external corpus into the seq2seq framework to guide the generation---as expected , this analysis suggests that including context in the model helps more
0
mem2seq is the first neural generative model that combines the multi-hop attention over memories with the idea of pointer network---v茅ronis proposed a graph based model named hyperlex based on the small-world properties of co-occurrence graphs
0
finally in section 6 we give a conclusion and outlook to future work---in section 6 we give a conclusion and outlook to future work
1
twitter is a microblogging service that has 313 million monthly active users 1---recently , the focus has also moved to mining from user-generated content , such as online debates , discussions on regulations , and product reviews
0
experiments on chinese-english translation show that our approach outperforms two state-of-the-art baselines significantly---experiments on chinese-english translation show that joint training with generalized agreement achieves significant improvements over two baselines for ( hierarchical )
1
in particular , we use the neural-network based models from , also referred as word embeddings---we use srilm toolkit to build a 5-gram language model with modified kneser-ney smoothing
0
the baseline further contains a hierarchical reordering model and a 7-gram word class language model---using our new objective , we train large multi-layer lstms
0
in comparison to co-training , self-training achieves substantially superior performance and is less sensitive to its input parameters---coreference data sets indicate that self-training outperforms co-training under various parameter settings and is comparatively less sensitive to parameter changes
1
we used the google news pretrained word2vec word embeddings for our model---we used the pre-trained word embeddings that are learned using the word2vec toolkit on google news dataset
1
for all classifiers , we used the scikit-learn implementation---we used standard classifiers available in scikit-learn package
1
thus , we observe a marginal improvement by using similarity-based metrics for wordnet---we also examine the possibility of using similarity metrics defined on wordnet
1
the detection model is implemented as a conditional random field , with features over the morphology and context---srilm toolkit is used to build these language models
0
ngram features have been generated with the srilm toolkit---srilm toolkit is used to build these language models
1
the model in this work is trained using transcribed child-directed speech from the babysrl portions of childes---as that study was aimed at modeling facts of child language acquisition , it uses child-directed speech from the thomas corpus , part of the childes database
1
for ptb pos tags , we tagged the text with the stanford parser---we used the stanford parser to extract dependency features for each quote and response
1
we show in this paper that following this intuition leads to suboptimal results---hara et al derived turn level ratings from overall ratings of the dialogue which were applied by the users afterwards on a five point scale
0
we use the open-source toolkit groundhog , which implements the attention-based encoder-decoder framework---we use opennmt , which is an implementation of the popular nmt approach that uses an attentional encoder-decoder network
1
textual entailment has been proposed as a generic framework for modeling language variability---textual entailment has been recently defined as a common solution for modelling language variability in different nlp tasks
1
this paper presented real-life use cases that require understanding audience segments in social media---this paper presents an approach that detects various audience attributes , including author
1
language models were built with srilm , modified kneser-ney smoothing , default pruning , and order 5---semantic parsing is the task of mapping a natural language ( nl ) sentence into a completely formal meaning representation ( mr ) or logical form
0
table 4 shows end-to-end translation bleu score results---base nps provides an accurate and fast bracketing method , running in time linear in the length of the tagged text
0
klein and manning show that much of the gain in statistical parsing using lexicalized models comes from the use of a small set of function words---we measured the overall translation quality with the help of 4-gram bleu , which was computed on tokenized and lowercased data for both systems
0
we first generate potential positive interpretations manipulating syntactic dependencies---while we rely heavily on syntactic dependencies , positive interpretations are generated in plain text
1
coreference resolution is the process of linking together multiple expressions of a given entity---although coreference resolution is a subproblem of natural language understanding , coreference resolution evaluation metrics have predominately been discussed in terms of abstract entities and hypothetical system errors
1
for our first hypothesis , we induce pos distribution information from a corpus , and approximate the probability of occurrence of pos blocks as per two statistical estimators separately---for our first hypothesis , we induce pos distribution information from a corpus , and approximate the probability of occurrence of pos blocks
1
initially , a seed lexicon of source and target language words was needed to learn a mapping between the two spaces---different from most work relying on a large number of handcrafted features , collobert and weston proposed a convolutional neural network for srl
0
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---in nlp , mikolov et al show that a linear mapping between vector spaces of different languages can be learned to infer missing dictionary entries by relying on a small amount of bilingual information
0
to do this , we relied on a neural network with a long short-term memory layer , which is fed from the word embeddings---we used the single layer long short-term memory networks to extract the features of each text
1
we train a 4-gram language model on the xinhua portion of the english gigaword corpus using the srilm toolkits with modified kneser-ney smoothing---for language model , we used sri language modeling toolkit to train a 4-gram model with modified kneser-ney smoothing
1
twitter is a huge microbloging service with more than 500 million tweets per day 1 from different locations in the world and in different languages---twitter is a widely used social networking service
1
the penn discourse treebank is the largest corpus richly annotated with explicit and implicit discourse relations and their senses---the penn discourse treebank is the largest available annotated corpora of discourse relations over 2,312 wall street journal articles
1
the output was evaluated against reference translations using bleu score which ranges from 0 to 1---the accuracy was measured using the bleu score and the string edit distance by comparing the generated sentences with the original sentences
1
a 4-gram language model is trained on the monolingual data by srilm toolkit---an lm is trained on 462 million words in english using the srilm toolkit
1
we tune phrase-based smt models using minimum error rate training and the development data for each language pair---for language model , we train a 5-gram modified kneser-ney language model and use minimum error rate training to tune the smt
1
we also use a 4-gram language model trained using srilm with kneser-ney smoothing---for language models , we use the srilm linear interpolation feature
1
we used the moses toolkit to train the phrase tables and lexicalized reordering models---we trained the statistical phrase-based systems using the moses toolkit with mert tuning
1
it has been widely recognized that verb meaning plays an important role in the syntactic realization of arguments and their interpretation---much research has focused on explaining the varied expression of verb arguments within syntactic positions
1
wordnet is a general english thesaurus which additionally covers biological terms---we used 5-gram models , estimated using the sri language modeling toolkit with modified kneser-ney smoothing
0
we train the model using the adam optimizer with the default hyper parameters---we considered one layer and used the adam optimizer for parameter optimization
1
phan et al firstly learned hidden topics from substantial external resources to enrich the features in short text---phan et al presented a general framework to expand the short and sparse text by appending topic names discovered using lda
1
we use the collapsed tree formalism of the stanford dependency parser---we apply the rules to each sentence with its dependency tree structure acquired from the stanford parser
1
in this paper , we have proposed the task of lexical normalisation for short text messages , as found in twitter and sms data---in this paper is this task of lexical normalisation of noisy english text , with a particular focus on twitter and sms messages
1
we implemented linear models with the scikit learn package---we use scikit learn python machine learning library for implementing these models
1
a pun is a form of wordplay in which a word suggests two or more meanings by exploiting polysemy , homonymy , or phonological similarity to another word , for an intended humorous or rhetorical effect---pun is a figure of speech that consists of a deliberate confusion of similar words or phrases for rhetorical effect , whether humorous or serious
1
we use the moses toolkit to train various statistical machine translation systems---we use the moses statistical mt toolkit to perform the translation
1
for the task of event trigger prediction , we train a multi-class logistic regression classifier using liblinear---we use logistic regression as the per-class binary classifier , implemented using liblinear
1
we compare our system with all participants of qald-6 as well as ganswer , nff and aqqu---our empirical results further confirm the strength of the model
0
we used standard classifiers available in scikit-learn package---we used scikit-learn library for all the machine learning models
1
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 )---in the task , our neural network approach is competitive with the systems
0
we trained a 5-grams language model by the srilm toolkit---our 5-gram language model is trained by the sri language modeling toolkit
1
relation extraction is a core task in information extraction and natural language understanding---relation extraction is the problem of populating a target relation ( representing an entity-level relationship or attribute ) with facts extracted from natural-language text
1
for phrase-based smt translation , we used the moses decoder and its support training scripts---for our baseline we use the moses software to train a phrase based machine translation model
1
we also use a 4-gram language model trained using srilm with kneser-ney smoothing---this type of features are based on a trigram model with kneser-ney smoothing
1
in this paper , we present a neural extractive document summarization ( n eu s um ) framework which jointly learns to score and select sentences---in this paper , we present a novel neural network framework for extractive document summarization by jointly learning to score and select sentences
1
we use the word2vec cbow model with a window size of 5 and a minimum frequency of 5 to generate 200-dimensional vectors---the word embeddings are initialized with 100-dimensions vectors pre-trained by the cbow model
1
moreover , for regularization , we place dropout after each lstm layer as suggested in---to prevent overfitting , we apply dropout operators to non-recurrent connections between lstm layers
1
we use word2vec from as the pretrained word embeddings---the annotation scheme leans on the universal stanford dependencies complemented with the google universal pos tagset and the interset interlingua for morphological tagsets
0
we show how this approach can be combined with additional features , in particular , the discourse features presented by cite-p-13-1-7---we also show how this approach can be combined with discourse features previously shown to be beneficial for the task of answer
1
below we divide related works into five broad categories based on which of these subtasks they addressed---in our experiment , svms and hm-svm training are carried out with svm struct packages
0
the feature weights were tuned on the wmt newstest2008 development set using mert---bannard and callison-burch , for instance , used a bilingual parallel corpus and obtained english paraphrases by pivoting through foreign language phrases
0
its complexity is linear in the sentence length---complexity of this algorithm is linear in the sentence length
1
we used the svd implementation provided in the scikit-learn toolkit---we used svm classifier that implements linearsvc from the scikit-learn library
1
in recent years , recurrent neural networks have risen in popularity among different nlp tasks---soricut and marcu use a standard bottomup chart parsing algorithm to determine the discourse structure of sentences
0
we use srilm to train a 5-gram language model on the xinhua portion of the english gigaword corpus 5th edition with modified kneser-ney discounting---coreference resolution is the task of grouping mentions to entities
0
we used a 5-gram language model with modified kneser-ney smoothing , built with the srilm toolkit---we use srilm to build 5-gram language models with modified kneser-ney smoothing
1
in this work , we present a sentence similarity using esa and syntactic similarities---in this paper , we describe our system submitted for the semantic textual similarity ( sts ) task
1
this paper presents the first attempt at a fully automatic extraction of news values from headline text---this paper presents the first attempt at a fully automatic and topic-independent extraction of news values which is applied and validated on headlines
1
stance detection is the task of automatically determining from text whether the author of the text is in favor of , against , or neutral towards a proposition or target---stance detection is a difficult task since it often requires reasoning in order to determine whether an utterance is in favor of or against a specific issue
1
in addition , we utilize the pre-trained word embeddings with 300 dimensions from for initialization---text simplification is the process of reducing the complexity of a text while preserving the original meaning
0
the data collection methods used to compile the dataset used in the shared task is described in---the way the dataset used in the trac 2018 shared task was built is described in
1
we use stanford ner for named entity recognition---for emd we used the stanford named entity recognizer
1
we find that , in our sample of languages , lexical semantic spaces largely coincide with genealogical relations---we use the adam optimizer for the gradient-based optimization
0
we used the weka implementation of na茂ve bayes for this baseline nb system---we create a manually-labeled dataset of dialogue from tv series ¡® friends ¡¯
0
our translation model is implemented as an n-gram model of operations using the srilm toolkit with 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
1
collobert et al use a convolutional neural network over the sequence of word embeddings---collobert et al set the neural network architecture for many current approaches
1
our neural generator follows the standard encoder-decoder paradigm---the translation quality is evaluated by case-insensitive bleu and ter metrics using multeval
0