text stringlengths 82 736 | label int64 0 1 |
|---|---|
we build all the classifiers using the l2-regularized linear logistic regression from the liblinear package---we train a linear support vector machine classifier using the efficient liblinear package | 1 |
the translation quality is evaluated by case-insensitive bleu and ter metric---the translation quality is evaluated by caseinsensitive bleu-4 metric | 1 |
compared to the traditional word-based translation models , the phrase-based translation model is more effective because it captures contextual information in modeling the translation of phrases as a whole , rather than translating single words in isolation---table 1 shows the performance for the test data measured by case sensitive bleu | 0 |
results show that srl is highly effective for orl , which is consistent with previous findings---the results show that srl information is very helpful for orl , which is consistent with previous studies | 1 |
sentiment classification is the task of labeling a review document according to the polarity of its prevailing opinion ( favorable or unfavorable )---sentiment classification is a special task of text categorization that aims to classify documents according to their opinion of , or sentiment toward a given subject ( e.g. , if an opinion is supported or not ) ( cite-p-11-1-2 ) | 1 |
one of the prominent achievements in this area is the co-training paradigm proposed by blum and mitchell---this idea was formalised by blum and mitchell in their presentation of co-training | 1 |
we used stanford corenlp for sentence splitting , part-of-speech tagging , named entity recognition , co-reference resolution and dependency parsing---automatic image description is the task of producing a natural-language utterance ( usually a sentence ) which correctly reflects the visual content of an image | 0 |
combinatory categorial grammars are a linguistically-motivated model for a wide range of language phenomena---they then searched the propbank wall street journal corpus for sentences containing such lexical items and annotated them with respect to metaphoricity | 0 |
our proposed work is identifying attitudes in sentences that appear in online discussions---in this work , we present a method to identify the attitude of participants in an online discussion | 1 |
some cost functions may act as rule backoffs , generating new rhs given unseen lhs , thus producing transducer rules ¡°onthe-fly¡±---at which some cost functions generate right-hand-sides of previously unseen left-hand-sides , thus creating transducer rules ¡° onthe-fly ¡± | 1 |
the detected spelling variants can then be used to mitigate the problems caused by spelling variation that were described above---detected spelling variants could also be used as the basis for an artifical standard that can then be used | 1 |
this paper describes a novel stacked subword model---coherence is a central aspect in natural language processing of multi-sentence texts | 0 |
the approach to discourse modeling is based on the work of grosz and sidner---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-8-1-9 ) | 0 |
och and ney show that for larger corpora , using word classes leads to lower alignment error rate---gra莽a et al show that alignment error rate can be improved with agreement constraints | 1 |
in particular , we use the liblinear svm 1va classifier---in this paper , we train our linear classifiers using liblinear 4 | 1 |
the similarity between two words is computed with the bi-sim measure---the formal similarity between two words is computed with the bi-sim measure | 1 |
for example , indicate the limited coverage of framenet as one of the main problems of this resource---in this paper we present a novel graph-based wsd algorithm which uses the full graph of wordnet efficiently , performing significantly better that previously published approaches in english | 0 |
44 computational linguistics , volume 14 , number 3 , september 1988 quilici , dyer , and flowers recognizing and responding to plan-oriented misconceptio---computational linguistics , volume 14 , number 3 , september 1988 47 quilici , dyer , and flowers recognizing and responding to plan-oriented misconceptions | 1 |
the penn discourse treebank is the largest available corpus of annotations for discourse relations , covering one million words of the wall street journal---the penn discourse treebank , developed by prasad et al , is currently the largest discourse-annotated corpus , consisting of 2159 wall street journal articles | 1 |
sentiment analysis is a collection of methods and algorithms used to infer and measure affection expressed by a writer---sentiment analysis is a natural language processing ( nlp ) task ( cite-p-10-1-14 ) which aims at classifying documents according to the opinion expressed about a given subject ( federici and dragoni , 2016a , b ) | 1 |
to train word embeddings , we adopt the approach proposed by mikolov et al , to derive a continuous , semantic representation of words based on context---furthermore , the concept of word embedding introduced by mikolov et al allows for words to have vector representations , such that syntactic and semantic similarities are embodied in the vector space | 1 |
we used an unsupervised transliteration model to transliterate the oov words---this year we used an em-based method to induce unsupervised transliteration models | 1 |
the word embeddings are word2vec of dimension 300 pre-trained on google news---relation extraction is a well-studied problem ( cite-p-12-1-6 , cite-p-12-3-7 , cite-p-12-1-5 , cite-p-12-1-7 ) | 0 |
we derive 100-dimensional word vectors using word2vec skip-gram model trained over the domain corpus---with english gigaword corpus , we use the skip-gram model as implemented in word2vec 3 to induce embeddings | 1 |
this is a problem of estimation of classifier effectiveness---in principle , this is a problem of estimation of classifier effectiveness | 1 |
parallel lda , which is lda with mpi , was used for training and inference for lda---parallel lda , which is lda with mpi , was used for training 100 mixture topic models and inference | 1 |
importantly , word embeddings have been effectively used for several nlp tasks , such as named entity recognition , machine translation and part-of-speech tagging---such features have been useful in a variety of english nlp models , including chunking , named entity recognition , and spoken language understanding | 1 |
the model was built using the srilm toolkit with backoff and good-turing smoothing---the target-side language models were estimated using the srilm toolkit | 1 |
to explicitly treat translation pairs from multiple domains , our system extends a domain adaptation method for neural networks , and apply it to the baseline anmt model---we used the mit java wordnet interface version 1 . 1 . 1 | 0 |
our framework was built with the cleartk toolkit with its wrapper for svmlight---we built a linear svm classifier using svm light package | 1 |
previous work has focused on congressional debates , company-internal discussions , and debates in online forums---in addition , we use early stopping based on the performance achieved on the development sets | 0 |
for probabilistic parsing , we can cite lfg , head-driven phrase structure grammar and probabilistic context-free grammars---we can cite lexical-functional grammar , head-driven phrase structure grammar and probabilistic context-free grammars | 1 |
cross-lingual textual entailment has been proposed as an extension of textual entailment---cross-lingual textual entailment is an extension of textual entailment | 1 |
in this paper , we propose a scalable approach to term extraction , which is based on string b-trees---in this paper , we propose a scalable approach and is capable of handling huge numbers of text documents | 1 |
we show that lvegs can subsume latent variable grammars and compositional vector grammars as special cases---such as latent variable grammars and compositional vector grammars can be interpreted as special cases of lvegs | 1 |
we use the well-known word embedding model that is a robust framework to incorporate word representation features---we use a popular word2vec neural language model to learn the word embeddings on an unsupervised tweet corpus | 1 |
the knowledge representation system kl-one , was the first dl---the knowledge representation system kl-one was the first dl | 1 |
one of the most frequently used methods for removing redundancy is maximal marginal relevance---one of the most well-known solutions of extractive text summarization is to use maximal marginal relevance | 1 |
experimental results illustrate that our method outperforms several baseline systems---the language model is a large interpolated 5-gram lm with modified kneser-ney smoothing | 0 |
we used a phrase-based smt model as implemented in the moses toolkit---we used the moses toolkit with its default settings to build three phrase-based translation systems | 1 |
we apply our model to the english portion of the conll 2012 shared task data , which is derived from the ontonotes corpus---we perform all our experiments on the english section of the conll-2012 corpus , which is based on ontonotes | 1 |
information extraction ( ie ) is a fundamental technology for nlp---information extraction ( ie ) is the nlp field of research that is concerned with obtaining structured information from unstructured text | 1 |
in this paper , we only focus on knowledge-based word sense disambiguation---in this paper , we present a unified model for both word sense representation and disambiguation | 1 |
the basic model of the our system is a log-linear model---following li et al , we define our model in the well-known log-linear framework | 1 |
we used srilm for training the 5-gram language model with interpolated modified kneser-ney discounting ,---in this paper , we propose a novel structure named augmented dependency path ( adp ) which attaches dependency subtrees to words on a shortest dependency path | 0 |
the word-based approach assumes one-to-one aligned source and target sentences---the language model is a 3-gram language model trained using the srilm toolkit on the english side of the training data | 0 |
in this work , we solve the inconsistency problem above by adapting the inter-sentence model of cite-p-12-3-3 to ccg parsing---we use the moses toolkit to train our phrase-based smt models | 0 |
a hybrid model of the word-based and the character-based model has also been proposed by luong and manning---optimization with regard to the bleu score is done using minimum error rate training as described by venugopal et al | 0 |
for sentence segmentation , we used the stanford corenlp library , which includes a probabilistic parser---we used the stanford factored parser to parse sentences into constituency grammar tree representations | 1 |
in this paper we present a new application of aligned multilinguai texts---in this paper we present a new , multilingual data-driven method for coreference resolution | 1 |
metaphor is a natural consequence of our ability to reason by analogy ( cite-p-16-1-12 )---we trained the l1-regularized logistic regression classifier implemented in liblinear | 0 |
in this paper , we proposed a novel method for cross-lingual text classification---in this paper , we propose a new approach to cltc , which trains a classification model in the source language | 1 |
this approach was pioneered by sch眉tze using second order co-occurrences to construct the context representation---sch眉tze proposed the word vectors as one such contextualized feature vector | 1 |
gale et al state and quantify the observation that words strongly tend to exhibit only one sense in a given discourse or document---one sense per discoursethe observation that words strongly tend to exhibit only one sense in a given discourse or document was stated and quantified in gale , church and yarowsky | 1 |
in addition , we use an english corpus of roughly 227 million words to build a target-side 5-gram language model with srilm in combination with kenlm---for the language model we use the corpus of 60,000 simple english wikipedia articles 3 and build a 3-gram language model with kneser-ney smoothing trained with srilm | 1 |
we used the logistic regression implemented in the scikit-learn library with the default settings---we used the scikit-learn implementation of svrs and the skll toolkit | 1 |
the evaluations were performed with scikit-learn using the skll toolkit 6 that makes it easy to run batch scikit-learn experiments---experiments were run with a variety of machine learning algorithms using the scikit-learn toolkit | 1 |
additionally , coreference resolution is a pervasive problem in nlp and many nlp applications could benefit from an effective coreference resolver that can be easily configured and customized---case-insensitive bleu4 was used as the evaluation metric | 0 |
we trained a 5-gram language model on the xinhua portion of gigaword corpus using the srilm toolkit---we used the srilm toolkit to train a 4-gram language model on the xinhua portion of the gigaword corpus , which contains 238m english words | 1 |
riezler et al demonstrate the advantages of translation-based approach to answer retrieval by utilizing a more complex translation model also trained from a large amount of data extracted from faqs on the web---in order to improve the word-based translation model with some contextual information , riezler et al and proposed a phrase-based translation model for question and answer retrieval | 1 |
the approach was further extended by ionescu , popescu , and cahill to combine several string kernels via multiple kernel learning---ionescu et al propose a combination of several string kernels and use multiple kernel learning | 1 |
neural machine translation has recently become the dominant approach to machine translation---neural machine translation is currently the state-of-the art paradigm for machine translation | 1 |
each system is optimized using mert with bleu as an evaluation measure---we trained a 4-gram language model on this data with kneser-ney discounting using srilm | 0 |
since text categorization is a task based on predefined categories , we know the categories for classifying documents---the tagger is based on the implementation of conditional random fields in the mallet toolkit | 0 |
in this paper , we addressed the task of deception detection within- and across-cultures---in this paper , we explore within-and across-culture deception detection | 1 |
a pun is a means of expression , the essence of which is in the given context the word or phrase can be understood in two meanings simultaneously ( cite-p-22-3-7 )---pun is a way of using the characteristics of the language to cause a word , a sentence or a discourse to involve two or more different meanings | 1 |
for this labeling , we estimate translation quality by the translation edit rate ter metric---the baseline is the psmt system used for the 2006 naacl smt workshop with phrase length 3 and a trigram language model | 0 |
to train the link embeddings , we use the speedy , skip-gram neural language model of mikolov et al via their toolkit word2vec---we first obtain word representations using the popular skip-gram model with negative sampling introduced by mikolov et al and implemented in the gensim package | 1 |
conditional random fields is a popular and efficient ml technique for supervised sequence labeling---lda is a simple model for topic modeling where topic probabilities are assigned words in documents | 0 |
furthermore , we train a 5-gram language model using the sri language toolkit---our 5-gram language model was trained by srilm toolkit | 1 |
co-training and its boostrapped adaptation require disjoint views of the features of the data---co-training methods make crucial usage of redundant models of the data | 1 |
this paper proposes a single document summarization method based on the trimming of a discourse tree---in this paper , we propose a single document summarization method based on the trimming of a discourse tree | 1 |
the system is demonstrable on a conventional pc laptop computer---functionality of the system is demonstrable on a laptop computer | 1 |
we applied dropout to each layer in our approach to mitigate overfitting---we applied dropout strategy to each layer of our model to mitigate overfitting | 1 |
an area that might benefit from a semi-supervised ne tagger is machine translation---a semi-supervised ne tagger can be successfully developed | 1 |
we measure the translation quality with automatic metrics including bleu and ter---we use case-sensitive bleu to assess translation quality | 1 |
two recent measures incorporate the notion of statistical significance in basic pmi formulation---two new measures have been proposed that incorporate the notion of statistical significance in basic pmi formulation | 1 |
quickly , he crawled under the car and unscrewed the drain bolt---we present a reinforcement learning ( cite-p-24-3-11 ) framework to learn user-adaptive referring expression generation policies | 0 |
we initialize the word embedding matrix with pre-trained glove embeddings---we used the pre-trained word embeddings that are learned using the word2vec toolkit on google news dataset | 0 |
with respect to the model optimization , we adopt the contrastive objective function used in previous works---for pcfg parsing , we select the berkeley parser | 0 |
in this paper we present a method for using lsa analysis to initialize a plsa model---from a set of models , in this paper we focus on finding a good way to initialize the plsa model | 1 |
semantic parsing is the task of mapping natural language to a formal meaning representation---semantic parsing is the task of mapping a natural language ( nl ) sentence into a complete , formal meaning representation ( mr ) which a computer program can execute to perform some task , like answering database queries or controlling a robot | 1 |
the two baseline methods were implemented using scikit-learn in python---both the transfer and transducer systems were trained and evaluated on english-to-mandarin chinese translation of transcribed utterances from the atis corpus | 0 |
the probabilistic language model is constructed on google web 1t 5-gram corpus by using the srilm toolkit---the target language model is trained by the sri language modeling toolkit on the news monolingual corpus | 1 |
this work provides the essential foundations for modular construction of signatures in typed unification grammars---for the experiment reported in section 5 , we use one of the largest , multi-lingual , freely available aligned corpus , europarl | 0 |
following the setup of duan et al , zhang and clark and huang and sagae , we split ctb5 into training , development , and test sets---we follow the setup of duan et al and split ctb5 into training , development , and test sets | 1 |
we used the stanford parser to generate dependency trees of sentences---kalchbrenner et al introduced a convolutional neural network for sentence modeling that uses dynamic k-max pooling to better model inputs of varying sizes | 0 |
two use wordnet and one uses the entries from a distributional thesaurus as classes for representation---two use wordnet and the other uses the entries in a thesaurus of distributionally similar words acquired automatically | 1 |
these grammars were trained with the thrax grammar extractor using its default settings , and translated using joshua---the translation models were trained with thrax , a grammar extractor for machine translation | 1 |
zens and ney show that itg constraints allow a higher flexibility in word ordering for longer sentences than the conventional ibm model---we adapt the minimum error rate training algorithm to estimate parameters for each member model in co-decoding | 0 |
the fundamental work for the pattern-based approaches is that of hearst---the most representative study is the group of patterns proposed by hearst | 1 |
for language model , we used sri language modeling toolkit to train a 4-gram model with modified kneser-ney smoothing---for the fluency and grammaticality features , we train 4-gram lms using the development dataset with the sri toolkit | 1 |
word embeddings learned from a large amount of unlabeled data have been shown to be able to capture the meaningful semantic regularities of words---we used the open source moses phrase-based mt system to test the impact of the preprocessing technique on translation results | 0 |
we apply sri language modeling toolkit to train a 4-gram language model with kneser-ney smoothing---for language model scoring , we use the srilm toolkit training a 5-gram language model for english | 1 |
furthermore , we propose top-rank enhanced loss functions , which are more sensitive to ranking errors at higher positions---for ranking , we propose top-rank enhanced loss functions , which incorporate a position-dependent cost that penalizes errors occurring at the top of the list | 1 |
nivre and mcdonald presented an integrating method to provide additional information for graph-based and transition-based parsers---nivre and mcdonald uses an ensemble model between transition-based and graph-based parsing approaches | 1 |
each document may be marked with multiple keyphrases that express unseen semantic properties---text and the selection of keyphrases are governed by the underlying hidden properties of the document | 1 |
a lexical analogy is a pair of word-pairs that share a similar semantic relation---lexical analogy is a pair of word-pairs that share a similar semantic relation | 1 |
we use the skip-gram model with negative sampling to learn word embeddings from a corpus of 400 million tweets also used in---conditional random fields are a class of undirected graphical models with exponent distribution | 0 |
for the actioneffect embedding model , we use pre-trained glove word embeddings as input to the lstm---we also use glove vectors to initialize the word embedding matrix in the caption embedding module | 1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.