Datasets:

deepachalapathi

/

parasci_data

like 0

in sections 3 and 4 , we present results of experiments that investigate how humans use colour terms for reference in production and comprehension---convolutional neural networks have recently achieved remarkably strong performance also on the practically important task of sentence classification

a letter-trigram language model with sri lm toolkit was then built using the target side of ne pairs tagged with the above position information---a 5-gram language model was created with the sri language modeling toolkit and trained using the gigaword corpus and english sentences from the parallel data

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

we parse the senseval test data using the stanford parser generating the output in dependency relation format---we convert both data sets to stanford dependencies with the stanford dependency converter

we use the wordsim353 dataset , divided into similarity and relatedness categories---we assess intrinsic embedding quality by considering correlation with human judgment on the wordsim353 test set

this grammar consists of a lexicon which pairs words or phrases with regular expression functions---each grammar consists of a set of rules evaluated in a leftto-right fashion over the input annotations , with multiple grammars cascaded together and evaluated bottom-up

generally phrase-based smt models outperform word-based ones---phrase-based smt systems have been shown to outperform word-based approaches

relation extraction ( re ) is the task of extracting instances of semantic relations between entities in unstructured data such as natural language text---relation extraction is the task of tagging semantic relations between pairs of entities from free text

sentiwordnet is another popular lexical resource for opinion mining---in the intrinsic evaluation , we use bleu , which was proposed as an automatic evaluation measure for smt , and human judgments

we use liblinear logistic regression module to classify document-level embeddings---we use liblinear 9 to solve the lr and svm classification problems

semantic parsing is the task of converting a sentence into a representation of its meaning , usually in a logical form grounded in the symbols of some fixed ontology or relational database ( cite-p-21-3-3 , cite-p-21-3-4 , cite-p-21-1-11 )---semantic parsing is the task of translating natural language utterances into a machine-interpretable meaning representation ( mr )

callison-burch et al used pivot languages for paraphrase extraction to handle the unseen phrases for phrase-based smt---in mt , callison-burch et al utilized paraphrases of unseen source phrases to alleviate data sparseness

distributed representations of words have been widely used in many natural language processing tasks---word embeddings have recently led to improvements in a wide range of tasks in natural language processing

coreference resolution is a challenging task , that involves identification and clustering of noun phrases mentions that refer to the same real-world entity---coreference resolution is a set partitioning problem in which each resulting partition refers to an entity

mauser et al integrated a logistic regression model predicting target words from all the source words in a pbsmt---mauser et al presented discriminative lexicon models to predict target words

as a byproduct , this approach further provides a new , effective perspective on handling those missing relations---as a byproduct , this provides a new perspective on handling missing relations

semantic parsing is the problem of translating human language into computer language , and therefore is at the heart of natural language understanding---semantic parsing is the task of mapping a natural language ( nl ) sentence into a completely formal meaning representation ( mr ) or logical form

our model can make full use of all informative sentences and alleviate the wrong labelling problem for distant supervised relation extraction---show that , our model can make full use of all informative sentences and effectively reduce the influence of wrong labelled instances

the words of input sentences are first converted to vector representations learned from word2vec tool---the character embeddings are computed using a method similar to word2vec

for 23 nouns 8 shown in as examples of nouns used as both mass and count nouns , accuracy was calculated using the bnc and ten-fold cross validation---for 25 nouns shown in as examples of nouns used as both mass and count nouns , accuracy on the bnc was calculated using ten-fold cross validation

in the dr subtask , the system achieved the median score in phase 1 and obtained a lower r in phase 2 , but in both cases it performs better than baseline---the srilm toolkit was used to build this language model

for optimization , we used adam with default parameters---we used adam optimizer with its standard parameters

we use 4-gram language models in both tasks , and conduct minimumerror-rate training to optimize feature weights on the dev set---we adapt the minimum error rate training algorithm to estimate parameters for each member model in co-decoding

in the most likely scenario ¨c porting a parser to a novel domain for which there is little or no annotated data ¨c the improvements can be quite large---and thus both of the case dependencies and specific sense restriction selected by the proposed method have much contribution to improving the performance in subcategorization preference test

word alignment models were first introduced in statistical machine translation---we use the pre-trained word2vec embeddings provided by mikolov et al as model input

we define a conditional random field for this task---as a classifier , we choose a first-order conditional random field model

coreference resolution is a task aimed at identifying phrases ( mentions ) referring to the same entity---coreference resolution is the task of automatically grouping references to the same real-world entity in a document into a set

coreference resolution is the task of determining when two textual mentions name the same individual---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

we use glove vectors with 100 dimensions trained on wikipedia and gigaword as word embeddings---we represent terms using pre-trained glove wikipedia 6b word embeddings

these tools are , however , not directly applicable to the task of multi-document summarization---tools are not well-suited for this task , as they do not support cross-document annotations , the modeling of complex tasks

the feature weights are tuned with minimum error-rate training to optimise the character error rate of the output---the feature weights of the translation system are tuned with the standard minimum-error-ratetraining to maximize the systems bleu score on the development set

thus , event extraction is a difficult task and requires substantial training data---event extraction is a challenging task , which aims to discover event triggers in a sentence and classify them by type

sentence compression is a paraphrasing task where the goal is to generate sentences shorter than given while preserving the essential content---sentence compression is the task of shortening a sentence while preserving its important information and grammaticality

translation quality is measured by case-insensitive bleu on newstest13 using one reference translation---the translation quality is evaluated by case-insensitive bleu and ter metrics using multeval

coreference resolution is a multi-faceted task : humans resolve references by exploiting contextual and grammatical clues , as well as semantic information and world knowledge , so capturing each of these will be necessary for an automatic system to fully solve the problem---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

sentiment analysis is the task of identifying the polarity ( positive , negative or neutral ) of review---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 )

since discourse is a natural form of communication , it favors the observation of the patient ’ s functionality in everyday life---discourse is a structurally organized set of coherent text segments

we use the cdec decoder 5 and induce scfg grammars from two sets of symmetrized alignments using the method described by chiang---for direct translation , we use the scfg decoder cdec 4 and build grammars using its implementation of the suffix array extraction method described in lopez

we present a weakly-supervised induction method to assign semantic information to food items---we present a semi-supervised graph-based approach to induce these food

in this work , we look at methods for bootstrapping the production of these statistical models without having an annotated treebank---in this work , we show that , by taking advantage of the constrained nature of these hpsg grammars , we can learn a discriminative parse selection model from raw text

for this , we used the combination of the entire swedish-english europarl corpus and the smultron data---for our baseline , we used a small parallel corpus of 30k english-spanish sentences from the europarl corpus

for example , turian et al have improved the performance of chunking and named entity recognition by using word embedding also as one of the features in their crf model---similarly , turian et al find that using brown clusters , cw embeddings and hlbl embeddings for name entity recognition and chunking tasks together gives better performance than using these representations individually

among others , there are studies using phrase-based statistical machine translation , which does not limit the types of grammatical errors made by a learner---there are several studies about grammatical error correction using phrase-based statistical machine translation

we used the svm implementation provided within scikit-learn---we trained the five classifiers using the svm implementation in scikit-learn

in this work , we propose a role identification model , which iteratively optimizes a team member role assignment that can predict the teamwork quality to the utmost extent---we utilise liblinear-java 3 with the l2-regularised l2-loss linear svm setting for the svm implementation , and snowball 4 for the stemmer

for the word-embedding based classifier , we use the glove pre-trained word embeddings---we use 300d glove vectors trained on 840b tokens as the word embedding input to the lstm

for language modeling , we use the english gigaword corpus with 5-gram lm implemented with the kenlm toolkit---for example , citation structure or rebuttal links , was used as extra information to model agreements or disagreements in debate posts and to infer their labels

djuric et al , 2015 ) highlighted the effectiveness of comment embeddings in detection of hate speech , by joint modelling comments and words using continuous-bag of words to generate a low dimensional embedding---the translation quality is evaluated by bleu and ribes

furthermore , we put a particular focus on recording the interactions between the users and the annotation tool---as a key property of our tool , we store all intermediate annotation results and record the user – system interaction

knowledge bases are usually highly incomplete---because the knowledge base is incomplete

we propose a framework for generating an abstractive summary from a semantic model of a multimodal document---into the semantic model , we can produce unified summaries of multimodal documents , resulting in an abstract

heilman and smith used tree kernels to search for the alignment that yields the lowest tree edit distance---heilman and smith presented a classification-based approach with tree-edit features extracted from a tree kernel

all features were log-linearly combined and their weights were optimized by performing minimum error rate training---the weights 位 m in the log-linear model were trained using minimum error rate training with the news 2009 development set

furthermore , the emotions in different dataset can be varied---then , the type of the emotions can be interpreted by observing the top

these models are built on recently translated sentences---models are able to substantially improve translation

we present hyp , an open-source toolkit that provides data structures and algorithms to process weighted directed hypergraphs---we have presented hyp , an open-source toolkit for representing and manipulating weighted directed hypergraphs , including functionality for learning arc

kiela and bottou showed that transferring representations from deep convolutional neural networks yield much better performance than bag-of-visual-words in multi-modal semantics---such cnnderived image representations have been found to be of higher quality than traditional bag of visual words models that were previously used in multi-modal semantics

we use a set of 318 english function words from the scikit-learn package---misra et al use a latent dirichlet allocation topic model to find coherent segment boundaries

we use sri language model toolkit to train a 5-gram model with modified kneser-ney smoothing on the target-side training corpus---we trained a standard 5-gram language model with modified kneser-ney smoothing using the kenlm toolkit on 4 billion running words

we estimate the parameters by maximizingp using the expectation maximization algorithm---in this work , we use the expectation-maximization algorithm

many words have multiple meanings , and the process of identifying the correct meaning , or sense of a word in context , is known as word sense disambiguation ( wsd )---the process of identifying the correct meaning , or sense of a word in context , is known as word sense disambiguation ( wsd )

our algorithm filters incorrect inference rules and identifies the directionality of the correct ones---for our baseline we use the moses software to train a phrase based machine translation model

semantic role labeling ( srl ) is a task of automatically identifying semantic relations between predicate and its related arguments in the sentence---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 )

in our implementation , we train a tri-gram language model on each phone set using the srilm toolkit---we use srilm to train a 5-gram language model on the target side of our training corpus with modified kneser-ney discounting

the basic idea of the neural network lm is to project the word indices onto a continuous space and to use a probability estimator operating on this space---the basic idea of this approach is to project the word indices onto a continuous space and to use a probability estimator operating on this space

our discriminative model is a linear model trained with the margin-infused relaxed algorithm---in this work , we use the margin infused relaxed algorithm with a hamming-loss margin

the weights associated to feature functions are optimally combined using the minimum error rate training---the parameter for each feature function in log-linear model is optimized by mert training

we trained a 4-gram language model with kneser-ney smoothing and unigram caching using the sri-lm toolkit---we use srilm to train a 5-gram language model on the target side of our training corpus with modified kneser-ney discounting

a back-off 2-gram model with good-turing discounting and no lexical classes was also created from the training set , using the srilm toolkit ,---a 5-gram language model was created with the sri language modeling toolkit and trained using the gigaword corpus and english sentences from the parallel data

this observation is evidence that the neural network can find good representations for pos tagging---but the neural networks leverage these features for improving tagging

for this target word the synonym ¡®record¡¯ was picked , which matches ¡®disc¡¯ in its musical sense---in contrast , cnn is able to extract local and position-invariant features well

in addition , we investigate the utility of incorporating additional specialized features tailored to peer review---in particular , we consider conditional random fields and a variation of autoslog

we use pre-trained word embeddings of size 300 provided by---we evaluated the translation quality using the case-insensitive bleu-4 metric

we use 300 dimension word2vec word embeddings for the experiments---then , we trained word embeddings using word2vec

we describe the development of xhpsg , a large-scale english grammar in the hpsg formalism translated from the xtag grammar---we constructed a type signature for the xtag english grammar , an existing broad-coverage grammar of english

we use online learning to train model parameters , updating the parameters using the adagrad algorithm---we apply online training , where model parameters are optimized by using adagrad

we investigate active learning techniques to reduce the size of these datasets and thus annotation effort---we investigate active learning ( al ) techniques to reduce the size of the dataset

the phrase-level and sentence-level precision of the generated paraphrases exceed 60 % and 55 % , respectively---lexical simplification is the task to find and substitute a complex word or phrase in a sentence with its simpler synonymous expression

we use a pbsmt model where the language model is a 5-gram lm with modified kneser-ney smoothing---for example , bengio et al introduced a model that learns word vector representations as part of a simple neural network architecture for language modeling

while negation is a grammatical category which comprises various kinds of devices to reverse the truth value of a proposition , speculation is a grammatical category which expresses the attitude of a speaker towards a statement in terms of degree of certainty , * corresponding author reliability , subjectivity , sources of information , and perspective ( cite-p-20-1-12 )---we also report the results using bleu and ter metrics

1 hindi is a verb final language with free word order and a rich case marking system---hindi is a verb final , flexible word order language and therefore , has frequent occurrences of non-projectivity in its dependency structures

since katakana words are basically transliterations from english , back-transliterating katakana noun compounds is also useful for splitting---memory-based learning , also known as instance-based , example-based , or lazy learning , is a supervised inductive learning algorithm for learning classification tasks

this intuition has been exploited in some systems to produce summaries---systems that use the discourse structure to produce summaries are also based on this intuition

finally , we apply our model to the recently established semeval-2015 diachronic text evaluation subtasks---we quantitatively evaluate our model on the semeval-2015 benchmark datasets released as part of the diachronic text evaluation exercise

we also use a 4-gram language model trained using srilm with kneser-ney smoothing---we implemented this model using the srilm toolkit with the modified kneser-ney discounting and interpolation options

the frequent frame you it , for example , largely identifies verbs , as shown in , taken from child-directed speech in the childes database---this setting is the same as that used in other studies

commonly used word vectors are word2vec , glove and fasttext---all word vectors are trained on the skipgram architecture

the maximum entropy approach presents a powerful framework for the combination of several knowledge sources---under the maximum entropy framework , evidence from different features can be combined with no assumptions of feature independence

semantic role labeling ( srl ) is defined as the task to recognize arguments for a given predicate and assign semantic role labels to them---semantic role labeling ( srl ) is the process of producing such a markup

in all submitted systems , we use the phrase-based moses decoder---for phrase-based smt translation , we used the moses decoder and its support training scripts

k枚nig et al looked also at mci and ad subjects and examined vocal features using support vector machine---mead is centroid based multi-document summarizer which generates summaries using cluster centroids produced by topic detection and tracking system

the database of typological features we used is the online edition 8 of the world atlas of language structures---the dataset we used in the present study is the online edition 2 of the world atlas of language structures

information extraction ( ie ) is the process of finding relevant entities and their relationships within textual documents---information extraction ( ie ) is a task of identifying 憽甪acts挕 ? ( entities , relations and events ) within unstructured documents , and converting them into structured representations ( e.g. , databases )

we investigate linguistic features that correlate with the readability of texts for adults with intellectual disabilities ( id )---we present a corpus of texts with readability judgments from adults with id ; ( 2 ) we propose a set of cognitively-motivated features which operate at the discourse level

thus , we use an attention mechanism to focus on the important words---we run skip-gram model on training dataset , and use the obtained word vector to initialize the word embedding part of model input

kawahara and uchimoto used a separately trained binary classifier to select reliable sentences as additional training data---kawahara and uchimoto used a separately trained binary classifier to select sentences as additional training data

many words have multiple meanings , and the process of identifying the correct meaning , or sense of a word in context , is known as word sense disambiguation ( wsd )---sentiment analysis is a much-researched area that deals with identification of positive , negative and neutral opinions in text

word similarity is typically low for synonyms that have many word senses since information about different senses are mashed together---we apply online training , where model parameters are optimized by using adagrad

in particular , svms achieve high generalization even with training data of a very high dimension---svms are known to achieve high generalization performance even with input data of high dimensional

for the feature-based system we used logistic regression classifier from the scikit-learn library---we use the logistic regression classifier in the skll package , which is based on scikit-learn , optimizing for f 1 score