text stringlengths 82 736 | label int64 0 1 |
|---|---|
on the remaining tweets , we trained a 10-gram word length model , and a 5-gram language model , using srilm with kneyser-ney smoothing---we used the target side of the parallel corpus and the srilm toolkit to train a 5-gram language model | 1 |
xu et al and santos et al both used convolutional architectures along with negative sampling to pursue this task---in this task , we used conditional random fields | 0 |
we showed experimentally that we can reduce running time by an order of magnitude , while at the same time improving mean average precision from .432 to .528 and mean reciprocal rank from .850 to .933---results were evaluated with both bleu and nist metrics | 0 |
for en-de , we used lmplz to estimate a 5-gram language model on all wmt german monolingual data and the german side of the parallel common crawl corpus---the language models were trained using srilm toolkit | 0 |
fader et al recently presented a scalable approach to learning an open domain qa system , where ontological mismatches are resolved with learned paraphrases---fader et al presented a qa system that maps questions onto simple queries against open ie extractions , by learning paraphrases from a large monolingual parallel corpus , and performing a single paraphrasing step | 1 |
with the more fine-grained feedback increasingly available on social media platforms ( e.g . laughter , love , anger , tears ) , it may be possible to distinguish different types of popularity as well as levels , e.g . shared sentiment vs. humor---with the more fine-grained feedback increasingly available on social media platforms ( e . g . laughter , love , anger , tears ) , it may be possible to distinguish different types of popularity | 1 |
the language model was a 5-gram language model estimated on the target side of the parallel corpora by using the modified kneser-ney smoothing implemented in the srilm toolkit---we use glove pre-trained word embeddings , a 100 dimension embedding layer that is followed by a bilstm layer of size 32 | 0 |
for word embedding , we used pre-trained glove word vectors with 300 dimensions , and froze them during training---we used glove vectors trained on common crawl 840b 4 with 300 dimensions as fixed word embeddings | 1 |
the weights in the log-linear model are tuned by minimizing bleu loss through mert on the dev set for each language pair---the feature weights for the log-linear combination of the features are tuned using minimum error rate training on the devset in terms of bleu | 1 |
recently , these models have also been successfully applied to morphological reinflection tasks ( cite-p-14-3-0 , cite-p-14-1-5 )---many neural morphological models have been proposed , most of them have focused on inflectional morphology ( e . g . , see cite-p-14-1-5 ) | 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 a nlp task that deals with extraction of opinion from a piece of text on a topic | 1 |
coreference resolution is the task of clustering referring expressions in a text so that each resulting cluster represents an entity---coreference resolution is a set partitioning problem in which each resulting partition refers to an entity | 1 |
the translation systems were evaluated by bleu score---we used a phrase-based smt model as implemented in the moses toolkit | 0 |
we also investigate whether these methods can outperform other automatic methods---in this paper , we investigate an automatic evaluation method that can reduce the errors of traditional automatic methods | 1 |
as a baseline system for our experiments we use the syntax-based component of the moses toolkit---support vector machines are one class of such model | 0 |
we use the skipgram model with negative sampling implemented in the open-source word2vec toolkit to learn word representations---following , we use the word analogical reasoning task to evaluate the quality of word embeddings | 1 |
we currently achieve coverage of 95.26 % , a bleu score of 0.7227 and string accuracy of 0.7476 on the penn-ii wsj section 23 sentences of length ¡ü20---and their best model achieves coverage of 90 . 56 % and a bleu score of 0 . 7723 on penn-ii wsj section 23 sentences of length ¡ü20 | 1 |
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 microblogging site where people express themselves and react to content in real-time | 1 |
furthermore , we train a 5-gram language model using the sri language toolkit---we use kenlm 3 for computing the target language model score | 0 |
the language models used were 7-gram srilm with kneser-ney smoothing and linear interpolation---the language models in this experiment were trigram models with good-turing smoothing built using srilm | 1 |
the target-syntax system is based on english parses from the collins parser---grammar rules were extracted from europarl using the collins parser for syntax on the english side | 1 |
another recent approach to guide clustering for sentiment analysis was introduced by dasgupta and ng , where they incorporate user feedback into a spectral clustering algorithm---more recently , dasgupta and ng proposed an unsupervised sentiment classification algorithm where user feedbacks are provided on the spectral clustering process in an interactive manner to ensure that text are clustered along the sentiment dimension | 1 |
one reason for this is that katakana noun compounds often include out-of-vocabulary words , which are difficult for the existing segmentation systems to deal with---katakana words ( i . e . , transliterated foreign words ) are particularly difficult to split , because katakana words are highly productive and are often out-of-vocabulary | 1 |
a dependency tree is a rooted , directed spanning tree that represents a set of dependencies between words in a sentence---such a forest is called a dependency tree | 1 |
the trigram language model is implemented in the srilm toolkit---the srilm toolkit was used to build this language model | 1 |
the results show that lexicalized surprisal according to both models is a significant predictor of rt , outperforming its unlexicalized counterparts---and indeed , the results show the ability of lexicalized surprisal to explain a significant amount of variance in rt data | 1 |
we use the 300-dimensional glove embeddings for english , and the 100-dimensional embeddings of reimers et al for german---we use the google news embeddings with 300 dimensions by mikolov et al for english and the 100-dimensional news-and wikipedia-based embeddings by reimers et al for german | 1 |
uchimoto et al showed that an evaluation method combining question-based evaluations with conventional automatic evaluations outperforms the conventional automatic evaluation methods---we ran these ml methods by the weka platform using the default parameters | 0 |
semantic parsing is the task of mapping natural language sentences to complete formal meaning representations---semantic parsing is the task of mapping natural language to a formal meaning representation | 1 |
recently , web mining systems have been built to automatically acquire parallel data from the web---web mining systems have been developed to automatically obtain parallel corpora from the web | 1 |
in this paper , we proposed a novel method for cross-lingual text classification---previous work consistently reported that word-based translation models yielded better performance than traditional methods for question retrieval | 0 |
distributed representations for words and sentences have been shown to significantly boost the performance of a nlp system---distributed word representations have been shown to improve the accuracy of ner systems | 1 |
li and yarowsky proposed an unsupervised method extracting the relation between a full-form phrase and its abbreviation from monolingual corpora---li and yarowsky present methods that take advantage of monolingual distributional similarities to identify the full form of abbreviated chinese words | 1 |
in particular , we assume the phrase-based smt framework---in this work , we apply a standard phrase-based translation system | 1 |
moreover , arabic is a morphologically complex language---morphologically , arabic is a non-concatenative language | 1 |
experimental results on real-world datasets show that our model can capture useful information from noisy data and achieve significant improvements on ds-qa as compared to all baselines---searchqa and triviaqa show that our system achieves significant and consistent improvement as compared to all baseline methods | 1 |
glorot et al first employed stacked denoising auto-encoders to extract meaningful representation for domain adaptation---glorot et al proposed to learn robust feature representations with stacked denoising auto-encoders | 1 |
we implement the pbsmt system with the moses toolkit---the automobile , kitchen and software reviews are taken from blitzer et al | 0 |
such methods are hard to be directly applied to dependency structures due to the great discrepancy between constituency and dependency grammars---for dependency grammars , but the special property of dependency grammars makes it hard to directly adopt the conventional structure transformation methods | 1 |
we use the stanford parser to generate the dependency parse tree of each sentence in the thread---we apply the rules to each sentence with its dependency tree structure acquired from the stanford parser | 1 |
entity type classification is the task of assigning type labels ( e.g. , person , location , organization ) to mentions of entities in documents---in our example , one should treat ¡° page ¡± , ¡° plant ¡± and ¡° gibson ¡± also as named-entity mentions | 0 |
bilingual dictionaries are an essential resource in many multilingual natural language processing tasks such as machine translation and cross-language information retrieval---bilingual lexicons play a vital role in many natural language processing applications such as machine translation or crosslanguage information retrieval | 1 |
we presented our study on research proceedings of approximately two decades from the leading nlp conference venues : emnlp and acl---in this paper , we present our study on research proceedings of approximately two decades from two leading nlp conferences , namely acl and emnlp , to complement a previous study on this topic | 1 |
we then lowercase all data and use all unique headlines in the training data to train a language model with the srilm toolkit---we use srilm toolkit to train a trigram language model with modified kneser-ney smoothing on the target side of training corpus | 1 |
to generate these trees , we employ the stanford pos tagger 8 and the stack version of the malt parser---we use the stanford pos tagger to obtain the lemmatized corpora for the sre task | 1 |
the most widely used are word error rate , position independent word error rate , the bleu score and the nist score---there exists a variety of different metrics , eg , word error rate , position-independent word error rate , bleu score , nist score , meteor , gtm | 1 |
we follow puduppully et al and , applying the learning and search framework of zhang and clark---we apply the global training and beam-search decoding framework of zhang and clark | 1 |
in this paper we present l obby b ack , a system to reconstruct the “ dark corpora ” that is comprised of model bills which are copied ( and modified ) by resource constrained state legislatures---in this paper we propose l obby b ack , a system that automatically identifies clusters of documents that exhibit text reuse , and generates “ prototypes ” | 1 |
distributional semantic models are based on the distributional hypothesis of meaning assuming that semantic similarity between words is a function of the overlap of their linguistic contexts---dsms are based on the distributional hypothesis of meaning assuming that semantic similarity between words is a function of the overlap of their linguistic contexts | 1 |
we apply sri language modeling toolkit to train a 4-gram language model with kneser-ney smoothing---we are releasing a black box for generating sentential paraphrases : machine translation language packs | 0 |
the parameters of the log-linear model are tuned by optimizing bleu on the development data using mert---in tuning the systems , mert iterative parameter estimation under ibm bleu 8 is performed on the development set | 1 |
burstein et al , used it for an educational purpose , and used it to predict the readability of essays---burstein et al employ this idea for evaluating coherence in student essays | 1 |
the parameters of the log-linear model are tuned by optimizing bleu on the development data using mert---the log-linear parameter weights are tuned with mert on the development set | 0 |
we propose two approaches to construct perturbed data to adversarially train the encoder and stabilize the decoder---to this end , we propose two approaches to constructing noisy inputs with small perturbations to make nmt | 1 |
chen et al used lstm to capture long distance dependencies---chen et al used long short-term memoryto capture long term dependency | 1 |
the charniak-lease phrase structure parses are transformed into the collapsed stanford dependency scheme using the stanford tools---this paper presents a parser-based word reordering model that employs a shift-reduce parser for inversion transduction grammars | 0 |
we used cdec as our hierarchical phrase-based decoder , and tuned the parameters of the system to optimize bleu on the nist mt06 corpus---we used cdec as our decoder , and tuned the parameters of the system to optimize bleu on the nist mt06 tuning corpus using the margin infused relaxed algorithm | 1 |
we use a pbsmt model where the language model is a 5-gram lm with modified kneser-ney smoothing---we have used the srilm with kneser-ney smoothing for training a language model for the first stage of decoding | 1 |
riloff et al . ( cite-p-23-1-8 ) applied bootstrapping to recognise subjective noun keywords and classify sentences as subjective or objective in newswire texts---riloff et al . ( cite-p-23-1-8 ) applied bootstrapping to recognise subjective noun keywords and classify sentences as subjective or objective | 1 |
here , we propose an alternative method based on a simple rule generator and decision tree learning---we used word2vec , a powerful continuous bag-of-words model to train word similarity | 0 |
document summarization can be treated as a special kind of translation process : translating from a bunch of related source documents to a short target summary---li and li have shown that word translation and bilingual bootstrapping is a good combination for disambiguation | 0 |
in this paper , we analyze the reasons that cause errors in chinese pinyin input method---in this paper , we study user input behaviors in chinese pinyin input method | 1 |
word alignment is a central problem in statistical machine translation ( smt )---word alignment is a natural language processing task that aims to specify the correspondence between words in two languages ( cite-p-19-1-0 ) | 1 |
however , we find no improvement as compared to using an equivalent amount of random-string autoencoder examples---we have presented a technique for creating a ∗ estimates for inference | 0 |
in this paper , we propose a new model for representing documents while automatically learning richer structural dependencies---in this paper , we focus on learning structure-aware document representations from data | 1 |
unlike previous research , we focus on the pairwise relationship between morphologically related wordforms , which we treat as potential paraphrases , and which we handle using paraphrasing techniques at various levels : word , phrase , and sentence level---unlike previous research , which has targeted word inflections and concatenations , we focus on the pairwise relationship between morphologically related words , which we treat as potential paraphrases and handle using paraphrasing techniques | 1 |
we trained two 5-gram language models on the entire target side of the parallel data , with srilm---as soft constraint , we propose a novel unsupervised model in the framework of posterior regularization | 0 |
the resulting kernel function is the cosine similarity between tweet vector pairs , in line with---finally , the resulting kernel function is the cosine similarity between vector pairs , in line with | 1 |
we show that these alignment models trained directly from discourse structures imposed on free text improve performance considerably over an information retrieval baseline and a neural network language model trained on the same data---we address this issue , and investigate whether alignment models for qa can be trained from artificial question-answer pairs generated from discourse structures imposed on free text | 1 |
statistical measures of word similarity have application in many areas of natural language processing , such as language modeling and information retrieval---statistical tests have been proposed to measure the strength of word similarity or word association in natural language texts | 1 |
this study explored the role of linguistic context in predicting quantifiers---in this study , we explore the role of linguistic context in predicting generalized quantifiers | 1 |
we used a 4-gram language model which was trained on the xinhua section of the english gigaword corpus using the srilm 4 toolkit with modified kneser-ney smoothing---for language model , we used sri language modeling toolkit to train a 4-gram model with modified kneser-ney smoothing | 1 |
transe ( cite-p-20-3-0 ) is a typical method in the neural-based approach , which learns vectors ( i.e. , embeddings ) for both entities and relations---transe ( cite-p-13-1-3 ) is a typical model considering relation vector as translating operations between head and tail vector , i.e. , math-w-2-3-0-13 when math-w-2-3-0-21 holds | 1 |
bahdanau et al incorporated the attention model into the sequence to sequence learning framework---bahdanau et al introduced soft alignments as part of the network architecture | 1 |
narayanan et al discuss a pos-based approach for identifying conditional types for the task of sentiment analysis---narayanan et al proposed a method for sentiment classification targeting conditional sentences | 1 |
the word embeddings are word2vec of dimension 300 pre-trained on google news---bengio et al propose a feedforward neural network to train a word-level language model with a limited n-gram history | 0 |
knowtator provides a very flexible mechanism for defining annotation schemas---knowtator facilitates the manual creation of annotated corpora that can be used for evaluating or training | 1 |
in order to decrease training times , we follow and eliminate unlikely dependencies using a form of coarse-to-fine pruning---following koo and collins , we eliminate unlikely dependencies using a form of coarse-to-fine pruning | 1 |
topic models , which identify latent semantic themes from text corpora , have previously been successfully used to discover aspects for sentiment analysis---topic models have recently been applied to information retrieval , text classification , and dialogue segmentation | 1 |
we use bleu , rouge , and meteor scores as automatic evaluation metrics---during evaluation , we employ rouge as our evaluation metric | 1 |
we proposed a novel framework that incorporates synonyms from monolingual linguistic resources in a word alignment generative model---we present a novel framework for word alignment that incorporates synonym knowledge collected from monolingual linguistic resources | 1 |
conquest uses a ravenclaw-based dialog manager---we use srilm toolkit to build a 5-gram language model with modified kneser-ney smoothing | 0 |
moreover , the results show the robustness of the proposed model---in addition , it has been shown that argument detection and argument classification need different sets of features | 0 |
callison-burch et al extract phrase-level paraphrases by mapping input phrases into a phrase table and then mapping back to the source language---callison-burch et al tackle the problem of unseen phrases in smt by adding source language paraphrases to the phrase table with appropriate probabilities | 1 |
we obtained these scores by training a word2vec model on the wiki corpus---we then used word2vec to train word embeddings with 512 dimensions on each of the prepared corpora | 1 |
sentiment analysis is the study of the subjectivity and polarity ( positive vs. negative ) of a text ( cite-p-7-1-10 )---sentiment analysis is a multi-faceted problem | 1 |
the target-side language models were estimated using the srilm toolkit---trigram language models are implemented using the srilm toolkit | 1 |
we evaluated the translation quality of the system using the bleu metric---we measure machine translation performance using the bleu metric | 1 |
a 5-gram language model built using kenlm was used for decoding---the 5-gram target language model was trained using kenlm | 1 |
in this paper , we proposed the hrde model and ltc module---in this paper , we propose a novel endto-end neural architecture | 1 |
we use the arc-based features of turboparser , which descend from several other feature models from the literature on syntactic dependency parsing---in aggregate , these constraints can greatly improve the consistency over the overall document-level predictions | 0 |
negation is a grammatical category that comprises devices used to reverse the truth value of propositions---negation is a linguistic phenomenon that can alter the meaning of a textual segment | 1 |
semantic role labeling ( srl ) is defined as the task to recognize arguments for a given predicate and assign semantic role labels to them---semantic role labeling ( srl ) is the process of extracting simple event structures , i.e. , “ who ” did “ what ” to “ whom ” , “ when ” and “ where ” | 1 |
the first step of sa research consists in building up lexical resources of affect , such as wordnet affect , sentiwordnet , or micro-wnop , cerini et---thus , the first step involved resided in building lexical resources of affect , such as wordnet affect , sentiwordnet , micro-wnop , cerini et | 1 |
dependency parsing is a fundamental task for language processing which has been investigated for decades---some of the well-known readability formulas include the smog formula , the fk formula , and the dalechall formula | 0 |
we use minimal error rate training to maximize bleu on the complete development data---we used minimum error rate training to tune the feature weights for maximum bleu on the development set | 1 |
semantic parsing is the task of automatically translating natural language text to formal meaning representations ( e.g. , statements in a formal logic )---semantic parsing is the task of converting natural language utterances into their complete formal meaning representations which are executable for some application | 1 |
we found that using a maximum phrase length of 10 for the translation model and a 6-gram language model produces the best results in terms of bleu scores for our sape model---we found that using a maximum phrase length of 7 for the translation model and a 5-gram language model produces the best results in terms of bleu scores for our sape model | 1 |
we used a 5-gram language model with modified kneser-ney smoothing implemented using the srilm toolkit---we trained kneser-ney discounted 5-gram language models on each available corpus using the srilm toolkit | 1 |
magatti et al introduced an approach for labelling topics that relied on two hierarchical knowledge resources labelled by humans , the google directory and the openoffice english thesaurus---we measure the overall translation quality using 4-gram bleu , which is computed on tokenized and lowercased data for all systems | 0 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.