Content Of Sentence Research Articles

An unsupervised method for extractive multi-document summarization based on centroid approach and sentence embeddings

Extractive multi-document summarization (MDS) is the process of automatically summarizing a collection of documents by ranking sentences according to their importance and informativeness. Text representation is a fundamental process that affects the effectiveness of many text summarization methods. Word embedding representations have shown to be effective for several Natural Language Processing (NLP) tasks including Automatic Text Summarization (ATS). However, most of these representations do not consider the order and the semantic relationships between words in a sentence. This does not fully allow grasping the sentence semantics and the syntactic relationships between sentences constituents. In this paper, to overcome this problem, we propose an unsupervised method for generic extractive multi-document summarization based on the sentence embedding representations and the centroid approach. The proposed method selects relevant sentences according to the final score obtained by combining three scores: sentence content relevance, sentence novelty, and sentence position scores. The sentence content relevance score is computed as the cosine similarity between the centroid embedding vector of the cluster of documents and the sentence embedding vectors. The sentence novelty metric is explicitly adopted to deal with redundancy. The sentence position metric assumes that the first sentences of a document are more relevant to the summary, and it assigns high scores to these sentences. Moreover, this paper provides a comparative analysis of nine sentence embedding models used to represent sentences as dense vectors in a low dimensional vector space in the context of extractive multi-document summarization. Experiments are performed on the standard DUC’2002–2004 benchmark datasets and the Multi-News dataset. The overall obtained results have shown that our method outperforms several state-of-the-art methods and achieves promising results compared to the best performing methods including supervised deep learning based methods.

Mask Transformer: Unpaired Text Style Transfer Based on Masked Language

Currently, most text style transfer methods encode the text into a style-independent latent representation and decode it into new sentences with the target style. Due to the limitation of the latent representation, previous works can hardly get satisfactory target style sentence especially in terms of semantic remaining of the original sentence. We propose a “Mask and Generation” structure, which can obtain an explicit representation of the content of original sentence and generate the target sentence with a transformer. This explicit representation is a masked text that masks the words with the strong style attribute in the sentence. Therefore, it can preserve most of the semantic meaning of the original sentence. In addition, as it is the input of the generator, it also simplified this process compared to the current work who generate the target sentence from scratch. As the explicit representation is readable and the model has better interpretability, we can clearly know which words changed and why the words changed. We evaluate our model on two review datasets with quantitative, qualitative, and human evaluations. The experimental results show that our model generally outperform other methods in terms of transfer accuracy and content preservation.

Evaluative adjective sentences: A question-based analysis of projection

Two contents of evaluative adjective sentences, like Kim was smart to watch the movie, are the prejacent (that Kim watched the movie) and the generalization (that the degree to which Kim watching the movie is smart was higher than the contextual standard of smart). The prejacent is standardly analyzed as a presupposition (e.g., Norrick 1978; Barker 2002; Oshima 2009; Kertz 2010). This paper argues against such analyses of the prejacent because, among other things, they do not capture an interaction between the prejacent and the generalization that has not yet been observed for projective content: when the prejacent projects, the generalization does not, and when the prejacent does not project, the generalization does. We develop an analysis according to which the prejacent is not a lexically specified presupposition but is projective to the extent that it is not at-issue with respect to the question addressed by the utterance of the evaluative adjective sentence. In addition to capturing the interaction between the prejacent and the generalization, our question-based projection analysis extends previous such analyses (e.g., Beaver & Clark 2008; Beaver et al. 2017; Simons et al. 2017) by incorporating a novel constraint on the question addressed by an utterance: the more the interpreter takes the truth of content c to follow from the common ground a priori, the less likely the question is about c. We provide experimental evidence for the analysis and argue that it improves on that of Karttunen et al. (2014), according to which evaluative adjectives are ambiguous.

Recognition and Naming of the Sentence Content in Deaf Children Trough Verbal and Neverbal Approach

The aim of the research is to determine the effects on the recognition and naming of certain sentence structures through a verbal and non-verbal approach, or through the use of a verbal and non-verbal communication channel. The study was conducted on a sample of 30 deaf pupils at age from 11 to 14 years. As a measurement instrument, a modified image description test was used. The test consisted of six complexes, illustrated sentences, in which the actions and subjects performing certain actions are shown in the picture sequence in a way that deaf children understand. The descriptive analysis method was used for data processing. Measures of central tendencies and variations have been made. Testing the difference between verbal and non-verbal approach was performed by t-test. The correlation between the chronological age and the recognition and naming of the spoken structures were verified through the non-verbal approach. The results of the research have shown that respondents have achieved better results in recognizing and naming spoken content using the non-verbal communication channel, that there is a statistically significant difference in the use of verbal and non-verbal communication approaches in recognizing and naming spoken content, and that there is a high correlation between the chronological age of respondents and recognizing and naming of spoken content through a non-verbal approach

Recognition and naming of the sentence content in deaf children through verbal and non-verbal approach

The aim of the research is to determine the effects on the recognition and naming of certain sentence structures through a verbal and non-verbal approach, or through the use of a verbal and non-verbal communication channel. The study was conducted on a sample of 30 deaf pupils at age from 11 to 14 years. As a measurement instrument, a modified image description test was used. The test consisted of six complexes, illustrated sentences, in which the actions and subjects performing certain actions are shown in the picture sequence in a way that deaf children understand. The descriptive analysis method was used for data processing. Measures of central tendencies and variations have been made. Testing the difference between verbal and non-verbal approach was performed by t-test. The correlation between the chronological age and the recognition and naming of the spoken structures were verified through the non-verbal approach. The results of the research have shown that respondents have achieved better results in recognizing and naming spoken content using the non-verbal communication channel, that there is a statistically significant difference in the use of verbal and non-verbal communication approaches in recognizing and naming spoken content, and that there is a high correlation between the chronological age of respondents and recognizing and naming of spoken content through a non-verbal approach.

Improving Semantic Parsing Using Statistical Word Sense Disambiguation (Student Abstract)

A Semantic Parser generates a logical form graph from an utterance where the edges are semantic roles and nodes are word senses in an ontology that supports reasoning. The generated representation attempts to capture the full meaning of the utterance. While the process of parsing works to resolve lexical ambiguity, a number of errors in the logical forms arise from incorrectly assigned word sense determinations. This is especially true in logical and rule-based semantic parsers. Although the performance of statistical word sense disambiguation methods is superior to the word sense output of semantic parser, these systems do not produce the rich role structure or a detailed semantic representation of the sentence content. In this work, we use decisions from a statistical WSD system to inform a logical semantic parser and greatly improve semantic type assignments in the resulting logical forms.

The Evolution of the Crime-Terror Nexus in Europe

Empirical evidence confirms that in Europe there are growing links between criminal organizations (OCs) and terrorist groups (GTs) – known as the “crime-terror nexus”. After the end of the Cold War and, particularly, since global financial operations came under greater control, the financing of illicit and criminal actions has consequently been conditioned. This resulted in greater proximity and cooperation between OCs and WGs not only in fragile states or undemocratic regimes, but in the western democracies themselves. The article also analyzes the content of the judicial sentences on cooperation and collusions between terrorist groups and criminal organizations.

Risk awareness, medication adherence, and driving behavior as determined by the provision of drug information to patients

ObjectiveThe study identifies appropriate risk expressions by healthcare professionals in communicating the risks of driving-impairing medicine to patients, gauging changed patient behavior, preventing traffic accidents due to drugs, and improving drug adherence. MethodsAn online questionnaire survey was conducted on participants’ perception of driving-related risks, and risk awareness, as well as reports of healthcare professionals’ expressions and warning messages regarding driving-impairing drugs. ResultsApproximately 80 % of participants were aware of the effects of pharmaceutical drugs on driving ability, and 50 % responded that they had received an explanation from their respective health professionals. As reported by participants, although healthcare professionals typically used more indirect expressions, direct warning messages were associated with high-risk awareness. ConclusionThe content of the explanatory sentences and debriefing influenced risk perception among participants. Direct expressions were more desirable for appropriate risk perception by participants. Providing information from healthcare professional about the degree of risks and patients’ determining their influence on driving behavior based on risk perception was necessary to clarify the predictors of driving behavior. Practice implicationsHealth professionals should be aware that their warning messages could have a significant impact on patients’ risk perception and driving behavior.

O poder simbólico do discurso judicial – a construção de uma “cultura jurídica” no contexto da judicialização das políticas públicas

O presente artigo problematiza a judicialização das políticas públicas a partir da construção de uma cultura jurídica que utiliza o argumento de autoridade no discurso jurídico. Analisando o conteúdo de sentenças concedendo o Benefício de Prestação Continuada o artigo reflete sobre um habitus que valoriza um direito jurisprudencial em detrimento das definições postas pela política pública.

Naive Russellians and Schiffer\u2019s Puzzle

Neo-Russellians like Salmon and Braun hold that: (A) the semantic contents of sentences are structured propositions whose basic components are objects and properties, (B) names are directly referential terms, and (C) a sentence of the form ‘n believes that S’ is true in a context c iff the referent of the name n in c believes the proposition expressed by S in c. This is sometimes referred to as ‘the Naive Russellian theory’. In this talk, I will discuss the Naive Russellian theory primarily in connection with a problem known as Schiffer’s puzzle. Schiffer first presented the puzzle as an argument against the Naive Russellian theory. Schiffer’s argument proceeds in two steps. In step one, Schiffer argues that the Naive Russellian theory is committed to two principles regarding de re belief; the special-case consequence and Frege’s constraint. Then, in step two, Schiffer argues that the special-case consequence is not consistent with Frege’s constraint. Salmon and Braun reply to Schiffer’s argument that although the Naive Russellian theory is committed to Frege’s constraint, it is not committed to the special-case consequence. However, in this paper, I will argue with a new Schiffer-case that even if the Naive Russellian theory is not committed to the special-case consequence, it is still not consistent with Frege’s constraint. Concluding, I will discuss the possibility to reject Frege’s constraint within the Naive Russellian theory.

The Content and Logic of Imperatives

This paper articulates an account of imperatives that sensibly supports the idea of a logic of imperative inferences. We rebuke common objections to the very possibility of such a logic, from a perspective based on recent linguistic work on the morphosyntax of imperatives. Specifically, we develop the notion that the content of an imperative sentence includes both a force operator alongside an imperational content to which the force applies. We further argue that this account of the content of imperatives constitutes a plausible and flexible framework to develop a logic of imperative by examining a number of reconstructions of this idea using semantical analogs of widespread modal semantics. After studying the performance of those approaches, we conclude that progress in imperative logic has been hindered by the failure to adopt conflict-tolerant and resource-sensitive semantics, but suggest that such considerations can be incorporated in this flexible framework. Finally, we also propose a simple account of the difference between operators applying to the content of imperatives and imperative operators, in a way that sheds light on some of the issues underlying the usual antinomies.

Computational Analysis of Arabic Cursive Steganography using Complex Edge Detection Techniques

Arabic language contains a multiple set of features which complete the process of embedding and extracting text from the compressed image. In specific, the Arabic language covers numerous textual styles and shapes of the Arabic letter. This paper investigated Arabic cursive steganography using complex edge detection techniques via compressed image, which comprises several characteristics (short, medium and Long sentence) as per the research interest. Sample of images from the Berkeley Segmentation Database (BSD) was utilized and compressed with a diverse number of bits per pixel through Least Significant Bit (LSB) technology. The method presented in this paper was evaluated based on five complex edge detectors (Roberts, Prewitt, Sobel, LoG, and Canny) via MATLAB. Canny edge detector has been demonstrated to be the most excellent solution when it is vital to perform superior edge discovery over-compressed image with little several facts, but Sobel appears to be better in term of the execution time for Long sentence contents.

A Joint Sentence Scoring and Selection Framework for Neural Extractive Document Summarization

Extractive document summarization methods aim to extract important sentences to form a summary. Previous works perform this task by first scoring all sentences in the document then selecting most informative ones; while we propose to jointly learn the two steps with a novel end-to-end neural network framework. Specifically, the sentences in the input document are represented as real-valued vectors through a neural document encoder. Then the method builds the output summary by extracting important sentences one by one. Different from previous works, the proposed joint sentence scoring and selection framework directly predicts the relative sentence importance score according to both sentence content and previously selected sentences. We evaluate the proposed framework with two realizations: a hierarchical recurrent neural network based model; and a pre-training based model that uses BERT as the document encoder. Experiments on two datasets show that the proposed joint framework outperforms the state-of-the-art extractive summarization models which treat sentence scoring and selection as two subtasks.

Machine Translation of Concise State Sentences for Quantifier Phrases

With the development of deep learning methods, the machine translation system based on deep neural network has reached a very high accuracy, but for some daily Chinese phenomenon machine translation system is still not able to translate correctly. In this paper, we study a sentence that often appears in Chinese spoken language, that is, a simple state sentence composed of quantitative phrases, and improve the existing machine translation system. The external helper program constructed in this paper is compatible with the current mainstream network translation systems, greatly improving the translation effect of these translation systems on the concise state sentences composed of quantitative phrases.

Text Analysis on Punishment Clauses in Korean Law: Changes in the Number, Level, and Balance of Punishment Clauses Over 30 Years

Abstract For this study, I completed a comprehensive review of punishment clauses in the Korean Legal Code from 1985 to 2016. Using a web crawler and text analysis, I gathered data on the laws and then identified the content of the penal sentence in each clause. By investigating the data, I was able to quantify and assess changes over time in: (1) the number of punishment clauses; (2) the severity of sentences; and (3) the balance between imprisonment and fines. In order to examine the causes of these changes, I separated the data into different sentence levels and sectors. I found that low-level punishment clauses had grown quickly, and some of the sectors responsible for the change included civil engineering and sex offenses. This comprehensive review of the penal sentences revealed issues of concern related to overcriminalization, overpenalization, and an imbalance of punishment level in the Korean Legal Code.

Fictional Content

Abstract It is usually taken for granted that a necessary condition for knowing that P is the truth of P. It may therefore be claimed that if we assume that we gain some kind of knowledge through fiction (let us call it fictional knowledge) of P*, then P* should be true—in at least a certain sense. My hypothesis is that this assumption grounds the different ways adopted by philosophers for attributing truth-conditions to fictional sentences. My claim in this work is that fictional sentences do not have truth-values and truth-conditions, but I want to maintain that we gain some kind of knowledge through fiction: to this aim, I will characterize the objective content of fictional sentences not in terms of truth-conditions (which are usually described by appealing to rules of the language or rules of interpretation of language independent of the actual users), but in dispositional terms and I will define a necessary condition for fictional knowledge accordingly.

The nature of entailment: an informational approach

In this paper we elaborate a conception of entailment based on what we call the Ackermann principle, which explicates valid entailment through a logical connection between (sets of) sentences depending on their informational content. We reconstruct Dunn’s informational semantics for entailment on the basis of Restall’s approach, with assertion and denial as two independent (primary) speech acts, by introducing the notion of a ‘position description’. We show how the machinery of position descriptions can effectively be used to define the positive and the negative information carried by sentences of a given language and to present a formalized version of the Ackermann principle as an inclusion relationship between the informational contents of the conclusions and the premises of a valid entailment. Being so interpreted, the entailment relation exhibits certain properties, including the property of transitivity (and, more generally, admissibility of the cut rule). Whereas properties such as Anderson and Belnap’s variable sharing property or Parry’s proscriptive principle are normally presented as imposing a relevance requirement on valid entailment, the suggested formalization of the Ackermann principle supports all of Gentzen’s structural rules, including weakening, a rule that is normally given up in sequent-style proof systems for relevance logics. In this way we propose an Ackermann-inspired explication of the nature of entailment as a relation between the informational contents of sentences.

Absolutism About Taste and Faultless Disagreement

It is usually claimed that taste utterances have judge-dependent semantic content. Jeremy Wyatt recently proposed a semantic theory that rejects this claim. According to him, the semantic content of taste sentences is judge-independent, but the content of our assertions made by uttering taste sentences is judge-dependent. He showed that this account explains faultless disagreements about tastes. My paper aims to raise some challenges to his proposal. First, a judge-independent taste proposition semantically expressed by a taste sentence seems unrelated to a judge-dependent taste proposition asserted by the speaker. It means that the latter proposition is not systematically obtained from the former. Second, the theory assumes that there are judge-independent taste properties (like being tasty, full stop). The existence of such properties can be questioned because it is problematic to state instantiation conditions for them. The paper ends with a sketch of an alternative explanation; it shares Wyatt’s account of faultless disagreements but does not suffer from its drawbacks.

Deep contextualized embeddings for quantifying the informative content in biomedical text summarization

Background and ObjectiveCapturing the context of text is a challenging task in biomedical text summarization. The objective of this research is to show how contextualized embeddings produced by a deep bidirectional language model can be utilized to quantify the informative content of sentences in biomedical text summarization. MethodsWe propose a novel summarization method that utilizes contextualized embeddings generated by the Bidirectional Encoder Representations from Transformers (BERT) model, a deep learning model that recently demonstrated state-of-the-art results in several natural language processing tasks. We combine different versions of BERT with a clustering method to identify the most relevant and informative sentences of input documents. Using the ROUGE toolkit, we evaluate the summarizer against several methods previously described in literature. ResultsThe summarizer obtains state-of-the-art results and significantly improves the performance of biomedical text summarization in comparison to a set of domain-specific and domain-independent methods. The largest language model not specifically pretrained on biomedical text outperformed other models. However, among language models of the same size, the one further pretrained on biomedical text obtained best results. ConclusionsWe demonstrate that a hybrid system combining a deep bidirectional language model and a clustering method yields state-of-the-art results without requiring labor-intensive creation of annotated features or knowledge bases or computationally demanding domain-specific pretraining. This study provides a starting point towards investigating deep contextualized language models for biomedical text summarization.

Abstractive Sentence Compression with Event Attention

Sentence compression aims at generating a shorter sentence from a long and complex source sentence while preserving the important content of the source sentence. Since it provides enhanced comprehensibility and readability to readers, sentence compression is required for summarizing news articles in which event words play a key role in delivering the meaning of the source sentence. Therefore, this paper proposes an abstractive sentence compression with event attention. In compressing a sentence of news articles, event words should be preserved as important information for sentence compression. For this, event attention is proposed which focuses on the event words of the source sentence in generating a compressed sentence. The global information in the source sentence is as significant as event words, since it captures the information of a whole source sentence. As a result, the proposed model generates a compressed sentence by combining both attentions. According to experimental results, the proposed model outperforms both the normal sequence-to-sequence model and the pointer generator on three datasets, namely the MSR dataset, Filippova dataset, and Korean sentence compression dataset. In particular, it shows 122% higher BLEU score than the sequence-to-sequence model. Therefore, the proposed model is effective in sentence compression.