Sentences In Order Research Articles

Hierarchical Concept-Driven Language Model

For guiding natural language generation, many semantic-driven methods have been proposed. While clearly improving the performance of the end-to-end training task, these existing semantic-driven methods still have clear limitations: for example, (i) they only utilize shallow semantic signals (e.g., from topic models) with only a single stochastic hidden layer in their data generation process, which suffer easily from noise (especially adapted for short-text etc.) and lack of interpretation; (ii) they ignore the sentence order and document context, as they treat each document as a bag of sentences, and fail to capture the long-distance dependencies and global semantic meaning of a document. To overcome these problems, we propose a novel semantic-driven language modeling framework, which is a method to learn a Hierarchical Language Model and a Recurrent Conceptualization-enhanced Gamma Belief Network, simultaneously. For scalable inference, we develop the auto-encoding Variational Recurrent Inference, allowing efficient end-to-end training and simultaneously capturing global semantics from a text corpus. Especially, this article introduces concept information derived from high-quality lexical knowledge graph Probase, which leverages strong interpretability and anti-nose capability for the proposed model. Moreover, the proposed model captures not only intra-sentence word dependencies, but also temporal transitions between sentences and inter-sentence concept dependence. Experiments conducted on several NLP tasks validate the superiority of the proposed approach, which could effectively infer meaningful hierarchical concept structure of document and hierarchical multi-scale structures of sequences, even compared with latest state-of-the-art Transformer-based models.

Neural Sentence Ordering Based on Constraint Graphs

Sentence ordering aims at arranging a list of sentences in the correct order. Based on the observation that sentence order at different distances may rely on different types of information, we devise a new approach based on multi-granular orders between sentences. These orders form multiple constraint graphs, which are then encoded by Graph Isomorphism Networks and fused into sentence representations. Finally, sentence order is determined using the order-enhanced sentence representations. Our experiments on five benchmark datasets show that our method outperforms all existing baselines significantly, achieving a new state-of-the-art performance. The results demonstrate the advantage of considering multiple types of order information and using graph neural networks to integrate sentence content and order information for the task. Our code is available at https://github.com/DaoD/ConstraintGraph4NSO.

Особенности окуломоторных реакций у студентов при восприятии текстовой информации

The study of oculomotor reactions in the process of performing cognitive tasks is becoming increasingly popular among psychophysiologists. The use of eye tracking techniques extended our knowledge about eye movements when perceiving visual information. Modern research using eye tracking focuses on oculomotor reactions while perceiving images and texts, performing logical visual tasks, etc. It should be noted that eye tracking parameters in visual cognitive activity not only reflect the peculiarities of eye movements, but also serve as markers of brain activity in terms of perception and processing of information. Of particular interest are eye tracking studies involving reading texts of varying complexity and emotional content with normal and inverted word order in sentences. This paper presents the results of an eyetracking research involving students during visual perception of texts. Two types of texts were used: text with correct order of letters and text with changed order of letters in words. We examined 40 students of Northern (Arctic) Federal University named after M.V. Lomonosov using the eye tracking system iView X ™ RED (SensoMotoric Instruments, Germany). It recorded a significantly higher number of fixations and saccades as well as longer duration of fixations while reading the text with changed order of letters in words. The subjects demonstrated a holistic perception of words when working with the text with changed order of letters, which led to shorter reading time. As a result, we identified two ways of visual information processing with high reproducibility, regardless of the text type: focused and superficial reading. For citation: Zvyagina N.V., Taleeva A.I., Kuznetsova D.A. Oculomotor Reactions in Students During Text Perception. Journal of Medical and Biological Research, 2021, vol. 9, no. 2, pp. 145–152. DOI: 10.37482/2687-1491-Z052

Salience-weighted agreement feature hierarchy modulates language comprehension

The brain establishes relations between elements of an unfolding sentence in order to incrementally build a representation of who is doing what based on various linguistic cues. Many languages systematically mark the verb and/or its arguments to imply the manner in which they are related. A common mechanism to this end is subject-verb agreement, whereby the marking on the verb covaries with one or more of the features such as person, number and gender of the subject argument in a sentence. The cross-linguistic variability of these features would suggest that they may modulate language comprehension differentially based on their relative weightings in a given language. To test this, we investigated the processing of subject-verb agreement in simple intransitive Arabic sentences in a visual event-related brain potential (ERP) study. Specifically, we examined the differences, if any, that ensue in the processing of person, number and gender features during online comprehension, employing sentences in which the verb either showed full agreement with the subject noun (singular or plural) or did not agree in one of the features. ERP responses were measured at the post-nominal verb. Results showed a biphasic negativity−late-positivity effect when the verb did not agree with its subject noun in either of the features, in line with similar findings from other languages. Crucially however, the biphasic effect for agreement violations was systematically graded based on the feature that was violated, which is a novel finding in view of results from other languages. Furthermore, this graded effect was qualitatively different for singular and plural subjects based on the differing salience of the features for each subject-type. These results suggest that agreement features, varying in their salience due to their language-specific weightings, differentially modulate language comprehension. We postulate a Salience-weighted Feature Hierarchy based on our findings and argue that this parsimoniously accounts for the diversity of existing cross-linguistic neurophysiological results on verb agreement processing.

The Effect of Using Song on Students’ Achivement in Simple Past Tense at SMK Antartika 2 Sidoarjo

SongIn English making sentence in the correct language order is difficult for students because English is a forein language. In this research focus on simple past tense. This research was coducted on tenth grade students in SMK Antartika 2 Sidoarjo who have understanding problems in simple past tense. This research aims to determine wether there is a significant influence on students achievement in simple past tense. This study used the pre-experimental design with Wilcoxon Sign Rank test and saturated sample technique. The researcher use the song as media in delivery of simple past tense to make it easier to understand.

Ceiling Effects on Weight in Heavy NP Shift

This squib tests theories of heavy NP shift that link constituent order to parsing. Our results indicate that increasing the weight of an object NP cannot make a heavy NP shift construction more acceptable than a comparison sentence in the canonical order, contra parsing-based theories. In addition, we examine acceptability as it relates to verb disposition. We find no significant differences between different subcategorizations, contrary to the findings of corpus studies. These results reveal a disconnect between production and comprehension; we further conclude that parser sensitivity to constituent structure is unlikely to affect speaker production of heavy NP shift.

Investigating the Impact of Interlanguage on Adult EFL Learners in Indonesia: Strengths and Weaknesses

The study of second language acquisition (SLA) concerns not only the way to learn a second language (L2), but also the unique language system created by language learners (interlanguage). This qualitative study analyzed EFL learners’ interlanguage by focusing on their strengths and weaknesses in speaking English, especially on phonological, grammatical and pragmatic competence in speaking English. The language data were obtained from a recorded conversation between two non-native English speakers from Indonesia. The data were analysed by referring to contrastive analysis, error analysis and the socio-cultural perspective in SLA. The results indicated that although both learners had a strong L1 accent, the learners did not face any difficulty in constructing English sentences in the right sentence order (SVO). In addition, when facing difficulties in speaking English, the learners were seen to ask and provide support to each other, as well as to talk to themselves (private speech). In terms of weaknesses, the study found that the learners often mispronounced the letter “t” in English as the sound of “t” in Indonesian, made grammatical errors in subject and verb agreement, and applied Indonesia’s pragmatic concepts when speaking English. The study suggests that it is necessary to have a positive view on learners’ interlanguage as it can be used as a tool to learn a second language.
 Keywords: Contrastive analysis, error analysis, interlanguage, second language acquisition, socio-cultural perspective in SLA.

Typological Shift in Bilinguals’ L1: Word Order and Case Marking in Two Varieties of Child Quechua

We compare speech production and find morphosyntactic change among children and adolescents speaking two closely related varieties of Quechua in Cuzco, Peru, and Chuquisaca, Bolivia. Quechua languages traditionally employ Object-Verb (OV) word order in main clauses, but robust case marking permits other orders, especially to focalize new information through constituent fronting. In Chuquisaca, but not Cuzco, we find that schoolchildren often omit the accusative suffix -ta from direct objects while retaining a prosodic trace of -ta. In other varieties, loss of accusative marking is associated with a shift towards Verb-Object (VO) word order, as in Spanish. However, we find that Chuquisaqueños use more canonical OV and possessor-possessed order in declarative sentences than do Cuzqueños, who employ a wide range of word orders at the sentence level and deviate from the possessor-possessed norm for Quechua noun phrases. Our finding of more rigid word order in Chuquisaca highlights the complex factors contributing to typological shift in word order and morphology: Omission of case morphology places a greater burden on word order to identify grammatical roles. Further, we find that Chuquisaqueño schoolchildren alone have begun to use huk, “one,” to mark indefiniteness, perhaps to replace determiner-like functions ascribed to -ta and to obsolescent markers such as evidentials.

A Judge’s Attempt at Sentencing Consistency After Booker

This article examines how Judge Jack B. Weinstein, who served as a federal district judge in the Eastern District of New York for 53 years, approached the issue of consistency in sentencing after the Federal Sentencing Guidelines were rendered advisory in United States v. Booker. It was Judge Weinstein’s practice to publish a statement of reasons for each sentence he imposed, and the article relies on these statements of reason – as well as articles and speeches published by the Judge – to demonstrate how he attempted to ensure that each defendant before him received an individualized sentencing determination, and that comparable cases and sentencing factors were treated consistently. It examines how Judge Weinstein developed a specific sentencing procedure, took into account the personal history and characteristics of each defendant, and considered the need for the sentence in order to avoid unwanted sentencing disparities. It is a condensed version of an article that was originally published in the Cardozo Law Review in 2019.

Judge Weinstein's Contributions to Sentencing Law

This article examines how Judge Jack B. Weinstein, who served as a federal district judge in the Eastern District of New York for 53 years, approached the issue of consistency in sentencing after the Federal Sentencing Guidelines were rendered advisory in United States v. Booker. It was Judge Weinstein’s practice to publish a statement of reasons for each sentence he imposed, and the article relies on these statements of reason – as well as articles and speeches published by the Judge – to demonstrate how he attempted to ensure that each defendant before him received an individualized sentencing determination, and that comparable cases and sentencing factors were treated consistently. It examines how Judge Weinstein developed a specific sentencing procedure, took into account the personal history and characteristics of each defendant, and considered the need for the sentence in order to avoid unwanted sentencing disparities. It is a condensed version of an article that was originally published in the Cardozo Law Review in 2019.

The impact of formative quizzes using CorrectWriting question type on learning word order in an ESL course

Studying word order in English sentences is a problem for ESL students whose native languages allow flexible word orders. Developing skills in formulating English sentences often requires trial-and-error process that takes time and requires supervision. Quiz software that is able to analyse students’ answers, and give meaningful feedback on mistakes in word order and hints on fixing them allows training without teacher’s supervision that greatly enhances the amount of attempts a student can do. In this study, we used CorrectWriting question type for Moodle LMS to create formative and summative quizzes for Russian students learning English as a second language. The experiments show that the students who made more than one attempt of formative quizzes had significantly better learning gains than the students who didn’t use formative quizzes or attempted them only once. It also showed vulnerabilities in question design that need to be fixed to utilise software features better in learning process.

An External Knowledge Enhanced Graph-based Neural Network for Sentence Ordering

As an important text coherence modeling task, sentence ordering aims to coherently organize a given set of unordered sentences. To achieve this goal, the most important step is to effectively capture and exploit global dependencies among these sentences. In this paper, we propose a novel and flexible external knowledge enhanced graph-based neural network for sentence ordering. Specifically, we first represent the input sentences as a graph, where various kinds of relations (i.e., entity-entity, sentence-sentence and entity-sentence) are exploited to make the graph representation more expressive and less noisy. Then, we introduce graph recurrent network to learn semantic representations of the sentences. To demonstrate the effectiveness of our model, we conduct experiments on several benchmark datasets. The experimental results and in-depth analysis show our model significantly outperforms the existing state-of-the-art models.

Graph-based cumulative score using statistical features for multilingual automatic text summarisation

Multilingual summarisation began to receive more attention these late years. Many approaches can be used to achieving this, among them: statistical and graph-based approaches. Our idea is to combine these two approaches into a new extractive text summarisation method. Surface statistical features are used to calculate a primary score for each sentence. The graph is used to selecting some candidate sentences and calculating a final score for each sentence based on its primary score and those of its neighbours in the graph. We propose four variants to calculating the cumulative score of a sentence. Also, the order of sentences is an important aspect of summary readability. We propose some other algorithms to generating the summary not just based on final scores but on sentences connections in the graph. The method is tested using MultiLing'15 workshop's MSS corpus and ROUGE metric. It is evaluated against some well known methods and it gives promising results.

GAN-GLS: Generative Lyric Steganography Based on Generative Adversarial Networks

Steganography based on generative adversarial networks (GANs) has become a hot topic among researchers. Due to GANs being unsuitable for text fields with discrete characteristics, researchers have proposed GAN-based steganography methods that are less dependent on text. In this paper, we propose a new method of generative lyrics steganography based on GANs, called GAN-GLS. The proposed method uses the GAN model and the large-scale lyrics corpus to construct and train a lyrics generator. In this method, the GAN uses a previously generated line of a lyric as the input sentence in order to generate the next line of the lyric. Using a strategy based on the penalty mechanism in training, the GAN model generates non-repetitive and diverse lyrics. The secret information is then processed according to the data characteristics of the generated lyrics in order to hide information. Unlike other text generation-based linguistic steganographic methods, our method changes the way that multiple generated candidate items are selected as the candidate groups in order to encode the conditional probability distribution. The experimental results demonstrate that our method can generate high-quality lyrics as stego-texts. Moreover, compared with other similar methods, the proposed method achieves good performance in terms of imperceptibility, embedding rate, effectiveness, extraction success rate and security.

Multi-Sentence Auxiliary Adversarial Networks for Fine-Grained Text-to-Image Synthesis

Due to the development of Generative Adversarial Networks (GANs), significant progress has been achieved in text-to-image synthesis task. However, most previous works have only focus on learning the semantic consistency between paired images and sentences, without exploring the semantic correlation between different yet related sentences that describe the same image, which leads to significant visual variation among the synthesized images. Accordingly, in this article, we propose a new method for text-to-image synthesis, dubbed Multi-sentence Auxiliary Generative Adversarial Networks (MA-GAN); this approach not only improves the generation quality but also guarantees the generation similarity of related sentences by exploring the semantic correlation between different sentences describing the same image. More specifically, we propose a Single-sentence Generation and Multi-sentence Discrimination (SGMD) module that explores the semantic correlation between multiple related sentences in order to reduce the variation between their generated images and enhance the reliability of the generated results. Moreover, a Progressive Negative Sample Selection mechanism (PNSS) is designed to mine more suitable negative samples for training, which can effectively promote detailed discrimination ability in the generative model and facilitate the generation of more fine-grained results. Extensive experiments on Oxford-102 and CUB datasets reveal that our MA-GAN significantly outperforms the state-of-the-art methods.

Синтаксис простого речення у поезії Євгена Маланюка

The article considers the peculiarities of the syntax in the simple sentence on the basis of small forms poetry which was written by Yevhen Malanyuk. It is analyzed and established that the stylistic resources of a simple sentence include special cases of using homogeneity, varieties of simple sentence models, various complicating components. Incomplete sentences are the most expressive syntactic constructions in Yevhen Malanyuk's poetic texts. It was found out, that the poet often uses the inversion, so in the sentence he changes the parts of the sentence in order to elicit different emotions (e.g. positive or negative feelings) of the reader, to focus his attention on a particular parts of the sentence. Also the author uses productive sentence models, but it is established that simple sentences with complex components are dominated in this way. The study clarifies the features of syntactic and stylistic organization of the language of poetic works, such as the specifics of subject and predicate coordination, cases of pictorial use of homogeneous parts of the sentence, the expressive possibilities of a simple sentence in poetic texts. The result of this work highlighted the features of syntactic and stylistic organization of the poetic text of the author. Thus, the study make it possible to recognize the contribution of the author in the development of syntactic system resources of Ukrainian artistic speech. Also, it helps to determine the patterns and trends in the development of literary language. The research of the syntactic linguistic aspect of Yevhen Malanyuk highlights the individual uniqueness of the poet. The perspective of the study is to research the syntax of a complex sentence in the works of Ye. Malaniuk. Besides to understand the features of the language of the poet's language as a unique phenomenon in the history of the Ukrainian literary language.

Specifies of construction industry terminology and methods of analysis of the natural language sentences similarity by using isomorphism of their structures

This article discusses methods of automatic translation by analogy. Each dictionary entry in the online dictionary contains a certain number of examples highlighting the use of this lexeme in different contexts. The purpose of this study is to find an example from an online dictionary that most fully conveys the meaning of a word in a particular context. Based on the fact that natural language sentences can be represented as graphs, the author considers several methods based on set theory, graph theory and linear algebra to compare sentences in order to find the closest example in meaning. In addition to purely formal comparison methods based on syntax, the author presents methods for analyzing sentences based on the semantics of their components. The balance between syntactic and semantic structure when comparing sentences gives the best result for finding the correct meaning of the predicate, which ensures an adequate transfer of the idea of the sentence during translation.

Sign Order in Slovenian Sign Language Transitive and Ditransitive Sentences

The basic sign order in Slovenian Sign Language (SZJ) is Subject-Verb-Object (SVO). This is shown by analysing non-topicalised or focalised transitive and ditransitive sentences that were elicited from first language SZJ informants using Picture Description Task. The data further reveal that the visual-gestural modality, through which SZJ is transmitted, plays a role in linearization since visually influenced classifier predicates trigger the non-basic SOV sign order in this language.

Implementing Teacher Races Game (TRG) to Improve Students' Grammar Skill Simple Past Tense

Vocabulary and grammar are the most essential components of all use of language as they have to be taught to the language learner to support to the development of the four skills, especially for writing skills Richard (1987:45). Unfortunately, Grammar has been believed as the most terrific problem for most of adolescence learners. They were terrific to construct wrong order of sentence (grammar) and this problem will effect on creating ideas in writing. This case indicates that grammar should be mastered by English learners especially students who want to create ideas in writing. Teacher races game provides a series of fun activity, and gives all the members’ opportunity to work cooperatively (Smith, 2008). This game can fulfill teacher’s need that is link and integrate four basics English language skills—listening, speaking, reading and writing skills into learning and teaching activity. This study was an experimental quantitative study which needs two subject groups – experimental and control group. The population chosen is one state junior high school in Surabaya. The simple random samples used; and VIII C as an experimental class and VIII A as a control class. The pre-test and post-test were given to both of the groups. Tests consist of tryout, pre-test and post-test that were used to collect the data. According to the collected data, the researcher made statistical calculation by using t-test and analysis to answer the research question. Ttest results were examined to know the comparison of experimental and control groups’ mean. The ttest of pre-test and post-test scores of the experimental and control group presented tvalue10.93. Meanwhile, the tvalue for rejecting the null hypothesis with df 50 at t.05 was 2.009. It indicated that tvalue > than tvalue (10.93>2.009) so that it can be concluded that null hypothesis (Ho) is rejected and the alternative hypothesis (Ha) is accepted.

Valence partikulí na základě dat z Pražského mluveného korpusu

Artykuł dotyczy opisu walencji partykuł w języku czeskim. Oparty jest na materiale Praskiego korpusu języka mówionego (PMK), pierwszego czeskiego korpusu języka mówionego. Po krótkim wprowadzeniu wstępnego materiału korpusowego i głównych wyników badań, artykuł skupia się na opisie walencji partykuł, czyli ich zdolności wiązania się z innymi słowami w zdaniu lub z całym zdaniem, co jest ściśle związane z miejscem ich występowania. Koncentruje się również na kumulacji partykuł w mowie, która z punktu widzenia syntagmatyki jest głównie tekstowa i wychwytuje tylko przypadkowo powtarzające się tekstowe kombinacje dwóch lub więcej partykuł obok siebie, w przeciwieństwie do systemowej, wynikiem które są partykuły wielowyrazowe. Jednak oba procesy wskazują na kombinatoryczne możliwości partykuł. Na koniec przedstawiono bezpośrednie otoczenie wszystkich partykuł, które wystąpiły w PMK.
 Abstrakt Článek se zabývá popisem valence partikulí v mluvené češtině. Vychází z materiálu Pražského mluveného korpusu (PMK), prvního českého mluveného korpusu. Po stručném představení výchozího korpusového materiálu a hlavních výsledků šetření se soustředí na popis valence partikulí, tj. jejich schopnost vázat se na jiná slova ve větě, nebo na celou větu, což úzce souvisí s místem jejich výskytu. Dále se zaměřuje na kumulacipartikulí v promluvě, která je z pohledu syntagmatiky převážně textová a zachycuje pouze náhodně se opakující textové kombinace dvou nebo více partikulí vedle sebe, na rozdíl od systémové, jejímž výsledkem jsou víceslovné partikule. Oba procesy však ukazují na kombinatorické možnosti partikulí. Nakonec je představeno bezprostřední okolí všech partikulí, které se vyskytly v PMK.