Search results for: longer text joke

Authors: Tanzeela Anbreen, Ammara Maqsood

Abstract:

Reading into writing is one of university students' most required academic skills. The current study explored postgraduate university students’ writing quality using a corpus-based approach. Twelve postgraduate students’ literature review essays were chosen for the corpus-based analysis. These essays were chosen because students had to incorporate multiple reading sources in these essays, which was a new writing exercise for them. The students were provided feedback at least two times which comprised of the written comments by the tutor highlighting the areas of improvement and also by using the ‘track changes’ function. This exercise was repeated two times, and students submitted two drafts. This investigation included only the finally submitted work of the students. A corpus-based approach was adopted to analyse the essays because it promotes autonomous discovery and personalised learning. The aim of this analysis was to understand the existing level of students’ writing before the start of their postgraduate thesis. Text Inspector was used to analyse the quality of essays. With the help of the Text Inspector tool, the vocabulary used in the essays was compared to the English Vocabulary Profile (EVP), which describes what learners know and can do at each Common European Framework of Reference (CEFR) level. Writing quality was also measured for the Flesch reading ease score, which is a standard to describe the ease of understanding the writing content. The results reflected that students found writing essays using multiple sources challenging. In most essays, the vocabulary level achieved was between B1-B2 of the CEFL level. The study recommends that students need extensive training in developing academic writing skills, particularly in writing the literature review type assignment, which requires multiple sources citations.

Keywords: literature review essays, postgraduate students, corpus-based analysis, vocabulary proficiency

Procedia PDF Downloads 73

Authors: Ali Algryani

Abstract:

Ellipsis involves the omission of an item or items that can be recovered from the preceding clause. Ellipsis is used as a cohesion marker; it enhances the cohesiveness of a text/discourse as a clause is interpretable only through making reference to an antecedent clause. The present study attempts to investigate the linguistic phenomenon of ellipsis from a translation perspective. It is mainly concerned with how ellipsis is translated from English into Arabic. The study covers different forms of ellipsis, such as noun phrase ellipsis, verb phrase ellipsis, gapping, pseudo-gapping, stripping, and sluicing. The primary aim of the study, apart from discussing the use and function of ellipsis, is to find out how such ellipsis phenomena are dealt with in English-Arabic translation and determine the implications of the translations of elliptical structures into Arabic. The study is based on the analysis of Dhofar University (DU) students' translations of sentences containing different forms of ellipsis. The initial findings of the study indicate that due to differences in syntactic structures and stylistic preferences between English and Arabic, Arabic tends to use lexical repetition in the translation of some elliptical structures, thus achieving a higher level of explicitness. This implies that Arabic tends to prefer lexical repetition to create cohesion more than English does. Furthermore, the study also reveals that the improper translation of ellipsis leads to interpretations different from those understood from the source text. Such mistranslations can be attributed to student translators’ lack of awareness of the use and function of ellipsis as well as the stylistic preferences of both languages. This has pedagogical implications on the teaching and training of translation students at DU. Students' linguistic competence needs to be enhanced through teaching linguistics-related issues with reference to translation and both languages, .i.e. source and target languages and with special emphasis on their use, function and stylistic preferences.

Keywords: cohesion, ellipsis, explicitness, lexical repetition

Procedia PDF Downloads 124

Authors: Judith Hanke

Abstract:

Due to the increase of students with reading difficulties, digital reading support with diagnostics was developed to foster the individual student's reading comprehension. The digital reading support focused on the reading comprehension of elementary school students. The digital reading packages consist of literary texts with aligned reading exercises. The number of students with German as a second language is growing in Germany. Students with multilingualism, language 1, and language 2 learn German together in school. The research's focus is on determining whether and to what extent multilingualism, language 1, and language 2 affect reading comprehension. For the methodology, an ABA design was selected for the intervention study to examine the reading support. The study was expedited from April 2023 until July 2023 and collected quantitative data of individuals, groups, and classes. It comprised a survey group (N = 58) and a control group (N = 53). The quantitative data was collected from 3 classes of 3 teachers and 47 students for all three test times. To show differences between the groups, a standardized reading comprehension test was used for the three test times, pretest, posttest, and follow-up. The standardized test consists of three subtests regarding word comprehension, sentence comprehension, and text comprehension. The main findings include that students who spoke German as their first language had the best test scores. Interestingly, students with a different language had better testing scores than students with German as the first language and (an) other language/s. Also, the students with another language outperformed the native language speakers in one of the subtests of the post-testing. The variables of spoken language at home and German as a second language were also examined and correlated with the test results. One significant correlation was found between spoken language at home and the text comprehension test of the pretesting. Additionally, the variable German as a second language had multiple significant correlations in the pretest, posttest and follow-up. The study's significance is to understand the influence of several languages, language 1, and language 2, on reading comprehension.

Keywords: multilingualism, language 1, language 2, reading comprehension, second language

Procedia PDF Downloads 29

Authors: Iqra Raza

Abstract:

This paper attempts to conceptualize the (de)humanization of the Muslim subject in Karen J. Greenberg and Janet Hamlin’s transmedia Sketching Guantanamo through a close study of the aesthetics and semiotics of the text. The Muslim experience, the paper shall argue, is mediated through a (de)humanization confined and incarcerated within the chains of artistic representation. Hamlin’s reliance on the distortions offered by stereotypes is reminiscent of the late Victorian epistemology on criminality, as evidenced most starkly in the sketch of Khalid Sheikh Mohammad. The position of the white artist thus becomes suspect in the enterprise of neo-Victorian ethnography. The visual stories of movement from within Guantanamo become potent; the paper shall argue, especially in juxtaposition with the images of stillness that came out from the detention centers, which portrayed the enactment of violence on individual bodies with a deliberate erasure of faces. So, while art becomes a way for reclaiming subjectivity or humanizing these identifiable bodies, the medium predicates itself on their objectification. The paper shall explore various questions about what it means for the (criminal?) subjects to be rendered into art rather than being photographed. Does art entail a necessary departure from the assumed objectivity of the photographic images? What makes art the preferred medium for (de)humanization of the violated Muslim bodies? What happens when art is produced without a recognition of the ‘precariousness’ of the life being portrayed? Rendering the detainees into art becomes a slippery task complicated by Hamlin’s privileged position outside the glass walls of the court. The paper shall adjourn analysis at the many dichotomies that exist in the text viz. between the White men and the brown, the Muslims and the Christians, Occident and the Orient problematized by Hamlin’s politics, that of a ‘neutral outsider’ which quickly turns on its head and becomes complicity in her deliberate erasure of the violence that shaped and still shapes Guantanamo.

Keywords: Abu Ghraib, Derrida, Guantanamo, graphic journalism, Muslimness, orient, spectrality

Procedia PDF Downloads 152

Authors: Tooba Sabir, Asima Jaffar, Namra Sabir, Mohammad Amjad Sabir

Abstract:

Ecocriticism is the study of relationship between human and environment which has been represented in literature since the very beginning in pastoral tradition. However, the analysis of such representation is new as compared to the other critical evaluations like Psychoanalysis, Marxism, Post-colonialism, Modernism and many others. Ecocritics seek to find information like anthropocentrism, ecocentrism, ecofeminism, eco-Marxism, representation of environment and environmental concept and several other topics. In the current study the representation of environmental concepts, were ecocritically analyzed in Paulo Coelho’s The Alchemist, one of the most read novels throughout the world, having been translated into many languages. Analysis of the text revealed, the representations of environmental ideas like landscapes and tourism, biodiversity, land-sea displacement, environmental disasters and warfare, desert winds and sand dunes. 'This desert was once a sea' throws light on different theories of land-sea displacement, one being the plate-tectonic theory which proposes Earth’s lithosphere to be divided into different large and small plates, continuously moving toward, away from or parallel to each other, resulting in land-sea displacement. Another theory is the continental drift theory which holds onto the belief that one large landmass—Pangea, broke down into smaller pieces of land that moved relative to each other and formed continents of the present time. The cause of desertification may, however, be natural i.e. climate change or artificial i.e. by human activities. Imagery of the environmental concepts, at some instances in the novel, is detailed and at other instances, is not as striking, but still is capable of arousing readers’ imagination. The study suggests that ecocritical justifications of environmental concepts in the text will increase the interactions between literature and environment which should be encouraged in order to induce environmental awareness among the readers.

Keywords: biodiversity, ecocritical analysis, ecocriticism, environmental disasters, landscapes

Procedia PDF Downloads 264

Authors: Shazia Salam

Abstract:

This paper focuses on studying the literary reactions of selected Muslim women writers to the event of Partition of India in the north western region. It aims to explore how Muslim women experienced the Partition and how that experience was articulated through their writing. There is a serious dearth of research on the experience of Muslim women who had to witness the momentous event of the subcontinent. Since scholars have often questioned the silence around the historiography related to the experiences of Muslim women, this paper aims to explore if literature could provide insights that may be less readily available in other modes of narration. Using literature as an archival source, it aims to delve into the arenas of history that have been cloistered and closed. Muslim women have been silent about their experiences of Partition which at the cost of essentializing could be attributed to patriarchal constraints, and taboos, on speaking of intimate matters. These silences have consigned the question of their experience to a realm of anonymity. The lack of ethnographic research has in a way been compensated in the realm of literature, mainly poetry and fiction. Besides reportage, literature remains an important source of social history about Partition and how Muslim women lived through it. Where traditional history fails to record moments of rupture and dislocation, literature serves the crucial purpose. The central premise in this paper is that there is a need to revise the history of partition owing to the gaps in historiography. It looks into if literature can serve as a ground for developing new approaches to history since the question of the representation always confronts us--between what a text represents and how it represents it since imagination of the writer plays a great role in the construction of any text. With this approach as an entry point, this paper aims to unpack the questions of representation, the coalescing of history /literature and the gendered nature of partition history. It concludes that the gaps in the narratives of Partition and the memory of Partition can be addressed by way of suing literary as a source to fill in the cracks and fissures.

Keywords: gender, history, literature, partition

Procedia PDF Downloads 210

Authors: Brenda Russell

Abstract:

Police officers spend a great deal of time responding to domestic violence calls. Recent research has found that men and women in heterosexual and same-sex relationships are equally likely to initiate intimate partner violence IPV) and likewise susceptible to victimization, yet police training tends to focus primarily on male perpetration and female victimization. Criminal justice studies have found that male perpetrators of IPV are blamed more than female perpetrators who commit the same offense. While previous research has examined officer’s response in IPV cases with male and female heterosexual offenders, research has yet to investigate police response in same-sex relationships. This study examined officers’ decisions to arrest, perceptions of blame, perceived danger to others, disrespect, and beliefs in prosecution, guilt and sentencing. Officers in the U.S. (N = 248) were recruited using word of mouth and access to police association websites where a link to an online study was made available. Officers were provided with one of 4 experimentally manipulated scenarios depicting a male or female perpetrator (heterosexual or same-sex) in a clear domestic assault situation. Officer age, experience with IPV and IPV training were examined as possible covariates. Training in IPV was not correlated to any dependent variable of interest. Age was correlated with perpetrator arrest and blame (.14 and .16, respectively) and years of experience was correlated to arrest, offering informal advice, and mediating the incident (.14 to -.17). A 2(perpetrator gender) X 2 (victim gender) factorial design was conducted. Results revealed that officers were more likely to provide informal advice and mediate in gay male relationships, and were less likely to arrest perpetrators in same-sex relationships. When officer age and years of experience with domestic violence were statistically controlled, effects for perpetrator arrest and providing informal advice were no longer significant. Officers perceived heterosexual male perpetrators as more dangerous, blameworthy, disrespectful, and believed they would receive significantly longer sentences than all other conditions. When officer age and experience were included as covariates in the analyses perpetrator blame was no longer statistically significant. Age, experience and training in IPV were not related to perceptions of victims. Police perceived victims as more truthful and believable when the perpetrator was a male. Police also believed victims of female perpetrators were more responsible for their own victimization. Victims were more likely to be perceived as a danger to their family when the perpetrator was female. Female perpetrators in same-sex relationships and heterosexual males were considered to experience more mental illness than heterosexual female or gay male perpetrators. These results replicate previous research suggesting male perpetrators are more blameworthy and responsible for their own victimization, yet expands upon previous research by identifying potential biases in police response to IPV in same-sex relationships. This study brings to the forefront the importance of evidence-based officer training in IPV and provides insight into the need for a gender inclusive approach as well as addressing the necessity of the practical applications for police.

Keywords: domestic violence, heterosexual, intimate partner violence, officer response, police officer, same-sex

Procedia PDF Downloads 347

Authors: Alejandro Molina-Villegas, Silvia Fernández-Sabido, Eduardo Mendoza-Vargas, Fátima Miranda-Pestaña

Abstract:

The automatic detection of indigenous languages ​​in digital texts is essential to promote their inclusion in digital media. Underrepresented languages, such as Maya, are often excluded from language detection tools like Google’s language-detection library, LANGDETECT. This study addresses these limitations by developing a hybrid language detection solution that accurately distinguishes Maya (YUA) from Spanish (ES). Two strategies are employed: the first focuses on creating a profile for the Maya language within the LANGDETECT library, while the second involves training a Naive Bayes classification model with two categories, YUA and ES. The process includes comprehensive data preprocessing steps, such as cleaning, normalization, tokenization, and n-gram counting, applied to text samples collected from various sources, including articles from La Jornada Maya, a major newspaper in Mexico and the only media outlet that includes a Maya section. After the training phase, a portion of the data is used to create the YUA profile within LANGDETECT, which achieves an accuracy rate above 95% in identifying the Maya language during testing. Additionally, the Naive Bayes classifier, trained and tested on the same database, achieves an accuracy close to 98% in distinguishing between Maya and Spanish, with further validation through F1 score, recall, and logarithmic scoring, without signs of overfitting. This strategy, which combines the LANGDETECT profile with a Naive Bayes model, highlights an adaptable framework that can be extended to other underrepresented languages in future research. This fills a gap in Natural Language Processing and supports the preservation and revitalization of these languages.

Keywords: indigenous languages, language detection, Maya language, Naive Bayes classifier, natural language processing, low-resource languages

Procedia PDF Downloads 16

Authors: Rona Lee Suelo-Calanao, Mahmoud Loubani

Abstract:

Objective: Sternal wound infection (SWI) following cardiac operation has a significant impact on patient morbidity and mortality. It also contributes to longer hospital stays and increased treatment costs. SWI management is mainly focused on treatment rather than prevention. This study looks at the effect of closed incision negative pressure therapy (ciNPT) dressing to help reduce the incidence of superficial SWI in high-risk patients after cardiac surgery. The ciNPT dressing was evaluated at 3 cardiac hospitals in the United Kingdom". Methods: All patients who had cardiac surgery from 2013 to 2021 were included in the study. The patients were classed as high risk if they have two or more of the recognised risk factors: obesity, age above 80 years old, diabetes, and chronic obstructive pulmonary disease. Patients receiving standard dressing (SD) and patients using ciNPT were propensity matched, and the Fisher’s exact test (two-tailed) and unpaired T-test were used to analyse categorical and continuous data, respectively. Results: There were 766 matched cases in each group. Total SWI incidences are lower in the ciNPT group compared to the SD group (43 (5.6%) vs 119 (15.5%), P=0.0001). There are fewer deep sternal wound infections (14(1.8%) vs. 31(4.04%), p=0.0149) and fewer superficial infections (29(3.7%) vs. 88 (11.4%), p=0.0001) in the ciNPT group compared to the SD group. However, the ciNPT group showed a longer average length of stay (11.23 ± 13 days versus 9.66 ± 10 days; p=0.0083) and higher mean logistic EuroSCORE (11.143 ± 13 versus 8.094 ± 11; p=0.0001). Conclusion: Utilization of ciNPT as an approach to help reduce the incidence of superficial and deep SWI may be effective in high-risk patients requiring cardiac surgery.

Keywords: closed incision negative pressure therapy, surgical wound infection, cardiac surgery complication, high risk cardiac patients

Procedia PDF Downloads 96

Authors: Irene Yi

Abstract:

Machine learning, natural language processing, and neural network models of language are becoming more and more prevalent in the fields of technology and linguistics today. Training data for machines are at best, large corpora of human literature and at worst, a reflection of the ugliness in society. Machines have been trained on millions of human books, only to find that in the course of human history, derogatory and sexist adjectives are used significantly more frequently when describing females in history and literature than when describing males. This is extremely problematic, both as training data, and as the outcome of natural language processing. As machines start to handle more responsibilities, it is crucial to ensure that they do not take with them historical sexist and misogynistic notions. This paper gathers data and algorithms from neural network models of language having to deal with syntax, semantics, sociolinguistics, and text classification. Results are significant in showing the existing intentional and unintentional misogynistic notions used to train machines, as well as in developing better technologies that take into account the semantics and syntax of text to be more mindful and reflect gender equality. Further, this paper deals with the idea of non-binary gender pronouns and how machines can process these pronouns correctly, given its semantic and syntactic context. This paper also delves into the implications of gendered grammar and its effect, cross-linguistically, on natural language processing. Languages such as French or Spanish not only have rigid gendered grammar rules, but also historically patriarchal societies. The progression of society comes hand in hand with not only its language, but how machines process those natural languages. These ideas are all extremely vital to the development of natural language models in technology, and they must be taken into account immediately.

Keywords: gendered grammar, misogynistic language, natural language processing, neural networks

Procedia PDF Downloads 120

Authors: Lucilla Crosta, Anthony Edwards

Abstract:

Companies and the market place nowadays struggle to find employees with adequate skills in relation to anticipated growth of their businesses. At least half of workers will need to undertake some form of up-skilling process in the next five years in order to remain aligned with the requests of the market . In order to meet these challenges, there is a clear need to explore the potential uses of AI (artificial Intelligence) based tools in assessing transversal skills (critical thinking, communication and soft skills of different types in general) of workers and adult students while empowering them to develop those same skills in a reliable trustworthy way. Companies seek workers with key transversal skills that can make a difference between workers now and in the future. However, critical thinking seems to be the one of the most imprtant skill, bringing unexplored ideas and company growth in business contexts. What employers have been reporting since years now, is that this skill is lacking in the majority of workers and adult students, and this is particularly visible trough their writing. This paper investigates how critical thinking and communication skills are currently developed in Higher Education environments through use of AI tools at postgraduate levels. It analyses the use of a branch of AI namely Machine Learning and Big Data and of Neural Network Analysis. It also examines the potential effect the acquisition of these skills through AI tools and what kind of effects this has on employability This paper will draw information from researchers and studies both at national (Italy & UK) and international level in Higher Education. The issues associated with the development and use of one specific AI tool Edulai, will be examined in details. Finally comparisons will be also made between these tools and the more recent phenomenon of Chat GPT and forthcomings and drawbacks will be analysed.

Keywords: critical thinking, artificial intelligence, higher education, soft skills, chat GPT

Procedia PDF Downloads 110

Authors: Svetlana Cojocaru, Mircea Petic, Inga Titchiev

Abstract:

Humanity faces more and more often with different social disasters, which in turn can generate new accidents and catastrophes. To mitigate their consequences, it is important to obtain early possible signals about the events which are or can occur and to prepare the corresponding scenarios that could be applied. Our research is focused on solving two problems in this domain: identifying signals related that an accident occurred or may occur and mitigation of some consequences of disasters. To solve the first problem, methods of selecting and processing texts from global network Internet are developed. Information in Romanian is of special interest for us. In order to obtain the mentioned tools, we should follow several steps, divided into preparatory stage and processing stage. Throughout the first stage, we manually collected over 724 news articles and classified them into 10 categories of social disasters. It constitutes more than 150 thousand words. Using this information, a controlled vocabulary of more than 300 keywords was elaborated, that will help in the process of classification and identification of the texts related to the field of social disasters. To solve the second problem, the formalism of Petri net has been used. We deal with the problem of inhabitants’ evacuation in useful time. The analysis methods such as reachability or coverability tree and invariants technique to determine dynamic properties of the modeled systems will be used. To perform a case study of properties of extended evacuation system by adding time, the analysis modules of PIPE such as Generalized Stochastic Petri Nets (GSPN) Analysis, Simulation, State Space Analysis, and Invariant Analysis have been used. These modules helped us to obtain the average number of persons situated in the rooms and the other quantitative properties and characteristics related to its dynamics.

Keywords: lexicon of disasters, modelling, Petri nets, text annotation, social disasters

Procedia PDF Downloads 197

Authors: Natalia Banasik

Abstract:

Comprehension of verbal irony is an example of pragmatic competence in understanding figurative language. The knowledge of how it develops may shed new light on the understanding of social and communicative competence that is crucial for one's effective functioning in the society. Researchers agree it is a competence that develops late in a child’s development. One of the abilities that seems crucial for irony comprehension is theory of mind (ToM), that is the ability to understand that others may have beliefs, desires and intentions different from one’s own. Although both theory of mind and irony comprehension require the ability to understand the figurative use of the false description of the reality, the exact relationship between them is still unknown. Also, even though irony comprehension in children has been studied for over thirty years, the results of the studies are inconsistent as to the age when this competence are acquired. The presented study aimed to answer questions about the developmental trajectories of irony comprehension and ascribing function to ironic utterances by preschool children. Specifically, we were interested in how it is related to the development of ToM and how comprehension of the function of irony changes with age. Data was collected from over 150 monolingual, Polish-speaking children and (so far) thirty bilingual children speaking Polish and English who live in the US. Four-, five- and six-year-olds were presented with a story comprehension task in the form of audio and visual stimuli programmed in the E-prime software (pre-recorded narrated stories, some of which included ironic utterances, and pictures accompanying the stories displayed on a touch screen). Following the presentation, the children were then asked to answer a series of questions. The questions checked the children’s understanding of the intended utterance meaning, evaluation of the degree to which it was funny and evaluation of how nice the speaker was. The children responded by touching the screen, which made it possible to measure reaction times. Additionally, the children were asked to explain why the speaker had uttered the ironic statement. Both quantitive and qualitative analyses were applied. The results of our study indicate that for irony recognition there is a significant difference among the three age groups, but what is new is that children as young as four do understand the real meaning behind the ironic statement as long as the utterance is not grammtically or lexically complex also, there is a clear correlation of ToM and irony comprehension. Although four-year olds and six-year olds understand the real meaning of the ironic utterance, it is not earlier than at the age of six when children start to explain the reason of using this marked form of expression. They talk about the speaker's intention to tell a joke, be funny, or to protect the listener's emotions. There are also some metalinguistic references, such as "mommy sometimes says things that don't make sense and this is called a metaphor".

Keywords: child's pragmatics, figurative speech, irony comprehension in children, theory of mind and irony

Procedia PDF Downloads 312

Authors: Yulan Wu

Abstract:

With the rapid development of social media, the issue of fake news has gained considerable prominence, drawing the attention of both the public and governments. The widespread dissemination of false information poses a tangible threat across multiple domains of society, including politics, economy, and health. However, much research has concentrated on supervised training models within specific domains, their effectiveness diminishes when applied to identify fake news across multiple domains. To solve this problem, some approaches based on domain labels have been proposed. By segmenting news to their specific area in advance, judges in the corresponding field may be more accurate on fake news. However, these approaches disregard the fact that news records can pertain to multiple domains, resulting in a significant loss of valuable information. In addition, the datasets used for training must all be domain-labeled, which creates unnecessary complexity. To solve these problems, an unsupervised domain knowledge discovery framework for fake news detection is proposed. Firstly, to effectively retain the multidomain knowledge of the text, a low-dimensional vector for each news text to capture domain embeddings is generated. Subsequently, a feature extraction module utilizing the unsupervisedly discovered domain embeddings is used to extract the comprehensive features of news. Finally, a classifier is employed to determine the authenticity of the news. To verify the proposed framework, a test is conducted on the existing widely used datasets, and the experimental results demonstrate that this method is able to improve the detection performance for fake news across multiple domains. Moreover, even in datasets that lack domain labels, this method can still effectively transfer domain knowledge, which can educe the time consumed by tagging without sacrificing the detection accuracy.

Keywords: fake news, deep learning, natural language processing, multiple domains

Procedia PDF Downloads 96

Authors: Dean J. Hill

Abstract:

This paper undertakes the task of conducting an in-depth textual analysis of the cli-fi genre. It examines how writing in the genre contributes to expressing and facilitating the articulation of environmental consciousness through the form of narrative. The paper begins by situating cli-fi within the literary continuum of ecological narratives and identifying the unique textual characteristics and thematic preoccupations of this area. The paper unfolds how cli-fi transforms the esoteric nature of climate science into credible narrative forms by drawing on language use, metaphorical constructs, and narrative framing. It also involves how descriptive and figurative language in the description of nature and disaster makes climate change so vivid and emotionally resonant. The work also points out the dialogic nature of cli-fi, whereby the characters and the narrators experience inner disputes in the novel regarding the ethical dilemma of environmental destruction, thus demanding the readers challenge and re-evaluate their standpoints on sustainability and ecological responsibilities. The paper proceeds with analysing the feature of narrative voice and its role in eliciting empathy, as well as reader involvement with the ecological material. In looking at how different narratorial perspectives contribute to the emotional and cognitive reaction of the reader to text, this study demonstrates the profound power of perspective in developing intimacy with the dominating concerns. Finally, the emotional arc of cli-fi narratives, running its course over themes of loss, hope, and resilience, is analysed in relation to how these elements function to marshal public feeling and discourse into action around climate change. Therefore, we can say that the complexity of the text in the cli-fi not only shows the hard edge of the reality of climate change but also influences public perception and behaviour toward a more sustainable future.

Keywords: cli-fi genre, ecological narratives, emotional arc, narrative voice, public perception

Procedia PDF Downloads 31

Authors: Zayd Khayi

Abstract:

This paper deals with the construction and analysis of the Tamazight text corpus. The grammatical structure of the Tamazight remains poorly understood, and a lack of comparative grammar leads to linguistic issues. In order to fill this gap, even though it is small, by constructed the diachronic corpus of the Tamazight language, and elaborated the program tool. In addition, this work is devoted to constructing that tool to analyze the different aspects of the Tamazight, with its different dialects used in the north of Africa, specifically in Morocco. It also focused on three Moroccan dialects: Tamazight, Tarifiyt, and Tachlhit. The Latin version was good choice because of the many sources it has. The corpus is based on the grammatical parameters and features of that language. The text collection contains more than 500 texts that cover a long historical period. It is free, and it will be useful for further investigations. The texts were transformed into an XML-format standardization goal. The corpus counts more than 200,000 words. Based on the linguistic rules and statistical methods, the original user interface and software prototype were developed by combining the technologies of web design and Python. The corpus presents more details and features about how this corpus provides users with the ability to distinguish easily between feminine/masculine nouns and verbs. The interface used has three languages: TMZ, FR, and EN. Selected texts were not initially categorized. This work was done in a manual way. Within corpus linguistics, there is currently no commonly accepted approach to the classification of texts. Texts are distinguished into ten categories. To describe and represent the texts in the corpus, we elaborated the XML structure according to the TEI recommendations. Using the search function may provide us with the types of words we would search for, like feminine/masculine nouns and verbs. Nouns are divided into two parts. The gender in the corpus has two forms. The neutral form of the word corresponds to masculine, while feminine is indicated by a double t-t affix (the prefix t- and the suffix -t), ex: Tarbat (girl), Tamtut (woman), Taxamt (tent), and Tislit (bride). However, there are some words whose feminine form contains only the prefix t- and the suffix –a, ex: Tasa (liver), tawja (family), and tarwa (progenitors). Generally, Tamazight masculine words have prefixes that distinguish them from other words. For instance, 'a', 'u', 'i', ex: Asklu (tree), udi (cheese), ighef (head). Verbs in the corpus are for the first person singular and plural that have suffixes 'agh','ex', 'egh', ex: 'ghrex' (I study), 'fegh' (I go out), 'nadagh' (I call). The program tool permits the following characteristics of this corpus: list of all tokens; list of unique words; lexical diversity; realize different grammatical requests. To conclude, this corpus has only focused on a small group of parts of speech in Tamazight language verbs, nouns. Work is still on the adjectives, prounouns, adverbs and others.

Keywords: Tamazight (Berber) language, corpus linguistic, grammar rules, statistical methods

Procedia PDF Downloads 64

Authors: Ramuel Spirituel Mattathiah A. San Juan, Neil Ambasing

Abstract:

Background: Reversal of vision metamorphopsia is a transient form of metamorphopsia described as an upside-down alteration of the visual field in the coronal plane. Patients would describe objects, such as cups, upside down, but the tea would not spill, and people would walk on their heads. It is extremely rare as a stable finding, lasting days or weeks. We report a case wherein this type of metamorphopsia occurred only in written words and lasted for six months. Objective: To the best of our knowledge, we report the first rare occurrence of reversal of vision metamorphopsia described as inverted words as the sole initial presentation of an underlying stroke. Case Presentation: We report a 59-year-old male with poorly controlled hypertension and diabetes mellitus who presented with a 3-day history of difficulty reading, described as the words were turned upside down as if the words were inverted horizontally then with the progression of deficits such as right homonymous hemianopia and achromatopsia, prosopagnosia. Cranial magnetic resonance imaging (MRI) revealed an acute infarct on the left posterior cerebral artery territory. Follow-up after six months revealed improvement of the visual field cut but with the persistence of the higher cortical function deficits. Conclusion: We report the first rare occurrence of metamorphopsia described as purely inverted words as the sole initial presentation of an underlying stroke. The differential diagnoses of a patient presenting with text reversal metamorphopsia should include stroke in the occipitotemporal areas. It further expands the landscape of metamorphopsias due to its exclusivity to written words and prolonged duration. Knowing these clinical features will help identify the lesion locus and improve subsequent stroke care, especially in time-bound management like intravenous thrombolysis.

Keywords: rare presentation, text reversal metamorphopsia, ischemic stroke, stroke

Procedia PDF Downloads 59

Authors: Muhammad Irfan, Guillermo Requena, Jan Haubrich

Abstract:

Adhesively bonded aluminum joints are common in automotive and aircraft industries and are one of the enablers of lightweight construction to minimize the carbon emissions during transportation for a sustainable life. This study is focused on the effects of two thermal processing routes, i.e., by direct and induction heating, and their parameters on void formation in PA6 bonded aluminum EN-AW6082 joints. The joints were characterized microanalytically as well as by lap shear experiments. The aging resistance of the joints was studied by accelerated aging tests at 80°C hot water. It was found that the processing of single lap joints by direct heating in a convection oven causes the formation of a large number of voids in the bond line. The formation of voids in the convection oven was due to longer processing times and was independent of any surface pretreatments of the metal as well as the processing temperature. However, when processing at low temperatures, a large number of small-sized voids were observed under the optical microscope, and they were larger in size but reduced in numbers at higher temperatures. An induction heating process was developed, which not only successfully reduced or eliminated the voids in PA6 bonded joints but also reduced the processing times for joining significantly. Consistent with the trend in direct heating, longer processing times and higher temperatures in induction heating also led to an increased formation of voids in the bond line. Subsequent single lap shear tests revealed that the increasing void contents led to a 21% reduction in lap shear strengths (i.e., from ~47 MPa for induction heating to ~37 MPa for direct heating). Also, there was a 17% reduction in lap shear strengths when the consolidation temperature was raised from 220˚C to 300˚C during induction heating. However, below a certain threshold of void contents, there was no observable effect on the lap shear strengths as well as on hydrothermal aging resistance of the joints consolidated by the induction heating process.

Keywords: adhesive, aluminium, convection oven, induction heating, mechanical properties, nylon6 (PA6), pretreatment, void

Procedia PDF Downloads 122

Authors: Sangita Pokhrel, Nalinda Somasiri, Rebecca Jeyavadhanam, Swathi Ganesan

Abstract:

Tourism is a booming industry with huge future potential for global wealth and employment. There are countless data generated over social media sites every day, creating numerous opportunities to bring more insights to decision-makers. The integration of Big Data Technology into the tourism industry will allow companies to conclude where their customers have been and what they like. This information can then be used by businesses, such as those in charge of managing visitor centers or hotels, etc., and the tourist can get a clear idea of places before visiting. The technical perspective of natural language is processed by analysing the sentiment features of online reviews from tourists, and we then supply an enhanced long short-term memory (LSTM) framework for sentiment feature extraction of travel reviews. We have constructed a web review database using a crawler and web scraping technique for experimental validation to evaluate the effectiveness of our methodology. The text form of sentences was first classified through Vader and Roberta model to get the polarity of the reviews. In this paper, we have conducted study methods for feature extraction, such as Count Vectorization and TFIDF Vectorization, and implemented Convolutional Neural Network (CNN) classifier algorithm for the sentiment analysis to decide the tourist’s attitude towards the destinations is positive, negative, or simply neutral based on the review text that they posted online. The results demonstrated that from the CNN algorithm, after pre-processing and cleaning the dataset, we received an accuracy of 96.12% for the positive and negative sentiment analysis.

Keywords: counter vectorization, convolutional neural network, crawler, data technology, long short-term memory, web scraping, sentiment analysis

Procedia PDF Downloads 88

Authors: Ana Maria Reis d'Azevedo Breda, Catarina Maria Neto da Cruz

Abstract:

The role of images/illustrations in communicating meanings and triggering emotions assumes an increasingly relevant role in contemporary texts, regardless of the age group for which they are intended or the nature of the texts that host them. It is no coincidence that children's books are full of illustrations and that the image/text ratio decreases as the age group grows. The vast majority of children's books can be considered multimodal texts containing text and images/illustrations interacting with each other to provide the young reader with a broader and more creative understanding of the book's narrative. This interaction is very diverse, ranging from images/illustrations that are not essential for understanding the storytelling to those that contribute significantly to the meaning of the story. Usually, these books are also read by adults, namely by parents, educators, and teachers who act as mediators between the book and the children, explaining aspects that are or seem to be too complex for the child's context. It should be noted that there are books labeled as children's books that are clearly intended for both children and adults. In this work, following a qualitative and interpretative methodology based on written productions, participant observation, and field notes, we will describe the perceptions of future teachers of the 1st cycle of basic education, attending a master's degree at a Portuguese university, about the role of the image in literary and non-literary texts, namely in mathematical texts, and how these can constitute precious resources for emotional regulation and for the design of creative didactic situations. The analysis of the collected data allowed us to obtain evidence regarding the evolution of the participants' perception regarding the crucial role of images in children's literature, not only as an emotional regulator for young readers but also as a creative source for the design of meaningful didactical situations, crossing other scientific areas, other than the mother tongue, namely mathematics.

Keywords: children’s literature, emotions, multimodal texts, soft skills

Procedia PDF Downloads 94

Authors: Prashanth Pillai, Purnaprajna Mangsuli

Abstract:

Oil exploration and production companies invest considerable time and efforts to extract essential well attributes (like well status, surface, and target coordinates, wellbore depths, event timelines, etc.) from unstructured data sources like technical reports, which are often non-standardized, multimodal, and highly domain-specific by nature. It is also important to consider the context when extracting attribute values from reports that contain information on multiple wells/wellbores. Moreover, semantically similar information may often be depicted in different data syntax representations across multiple pages and document sources. We propose a hierarchical multi-source fact extraction workflow based on a deep learning framework to extract essential well attributes at scale. An information retrieval module based on the transformer architecture was used to rank relevant pages in a document source utilizing the page image embeddings and semantic text embeddings. A question answering framework utilizingLayoutLM transformer was used to extract attribute-value pairs incorporating the text semantics and layout information from top relevant pages in a document. To better handle context while dealing with multi-well reports, we incorporate a dynamic query generation module to resolve ambiguities. The extracted attribute information from various pages and documents are standardized to a common representation using a parser module to facilitate information comparison and aggregation. Finally, we use a probabilistic approach to fuse information extracted from multiple sources into a coherent well record. The applicability of the proposed approach and related performance was studied on several real-life well technical reports.

Keywords: natural language processing, deep learning, transformers, information retrieval

Procedia PDF Downloads 193

Authors: Sarah Amodeo

Abstract:

Drawing on the author’s research about the experiences of female immigrant students in academic Adult Education, in Montreal, Quebec, this paper deconstructs the benefits of autoethnography as a methodology for educators in Adult Education. Autoethnography is an advantageous methodology for teachers in Adult Education as it allows for deep engagement, allowing for educators to reflect on student experiences and their day-to-day realities, and in turn, allowing for professional development, improved andragogy, and changes to classroom practices. Autoethnography is a qualitative research methodology that cultivates strategies for improving adult learning. The paper begins by outlining the context that inspired autoethnography for the author’s work, highlighting the emergence of autoethnography as a method, while examining how it is evolving and drawing on foundational work that continues to inspire research. The basic autoethnographic methodologies that are explored in this paper include the use of memory work in episode formation, the use of personal photographs, and textual readings of artworks. Memory work allows for the researcher to use their professional experience and the lived/shared experiences of their students in their research, drawing on episodes from their past. Personal photographs and descriptions of artwork allow researchers to explore images of learning environments/realities in ways that compliment student experiences. Major findings of the text are examined through the analysis of categories of autoethnography. Specific categories include realism, impressionism, and conceptualism which aid in orientating the analysis and emergent themes that develop through self-study. Finally, the text presents a discussion surrounding the limitations of autoethnography, with attention to the trustworthiness and ethical issues. The paper concludes with a consideration of the implications of autoethnography for adult educators in juxtaposition with youth sector work.

Keywords: artwork, autoethnography, conceptualism, episode formation, impressionism, memory work, personal photographs, and realism, realism

Procedia PDF Downloads 193

Authors: Q. R. Cheah, Y. F. Tan

Abstract:

In modern times, the push for a low carbon footprint entails achieving carbon neutrality as a goal for future generations. One possible step towards carbon footprint reduction is the use of more durable materials with longer lifespans, for example, silicone data cableswhich show at least double the lifespan of similar plastic products. By having greater durability and longer lifespans, silicone data cables can reduce the amount of trash produced as compared to plastics. Furthermore, silicone products don’t produce micro contamination harmful to the ocean. Every year the electronics industry produces an estimated 5 billion data cables for USB type C and lightning data cables for tablets and mobile phone devices. Material usage for outer jacketing is 6 to 12 grams per meter. Tests show that the product lifespan of a silicone data cable over plastic can be doubled due to greater durability. This can save at least 40,000 tonnes of material a year just on the outer jacketing of the data cable. The facility in this study specialises in compounding of liquid silicone rubber (LSR) material for the extrusion process in jacketing for the silicone data cable. This study analyses the carbon emissions from the facility, which is presently capable of producing more than 1,000 tonnes of LSR annually. This study uses guidelines from the World Business Council for Sustainable Development (WBCSD) and World Resources Institute (WRI) to define the boundaries of the scope. The scope of emissions is defined as 1. Emissions from operations owned or controlled by the reporting company, 2. Emissions from the generation of purchased or acquired energy such as electricity, steam, heating, or cooling consumed by the reporting company, and 3. All other indirect emissions occurring in the value chain of the reporting company, including both upstream and downstream emissions. As the study is limited to the compounding facility, the system boundaries definition according to GHG protocol is cradle-to-gate instead of cradle-to-grave exercises. Malaysia’s present electricity generation scenario was also used, where natural gas and coal constitute the bulk of emissions. Calculations show the LSR produced for the silicone data cable with high fire retardant capability has scope 1 emissions of 0.82kg CO2/kg, scope 2 emissions of 0.87kg CO2/kg, and scope 3 emissions of 2.76kg CO2/kg, with a total product carbon footprint of 4.45kg CO2/kg. This total product carbon footprint (Cradle-to-gate) is comparable to the industry and to plastic materials per tonne of material. Although per tonne emission is comparable to plastic material, due to greater durability and longer lifespan, there can be significantly reduced use of LSR material. Suggestions to reduce the calculated product carbon footprint in the scope of emissions involve 1. Incorporating the recycling of factory silicone waste into operations, 2. Using green renewable energy for external electricity sources and 3. Sourcing eco-friendly raw materials with low GHG emissions.

Keywords: carbon footprint, liquid silicone rubber, silicone data cable, Malaysia facility

Procedia PDF Downloads 95

Authors: Saleh Dabil

Abstract:

This study investigates the level of existence of organized retail crime in supermarkets of Riyadh, Saudi Arabia. The store managers, security managers and general employees were asked about the types of retail crimes occur in the stores. Three independent variables were related to the report of organized retail theft. The independent variables are: (1) the supermarket profile (volume, location, standard and type of the store), (2) the social physical environment of the store (maintenance, cleanness and overall organizational cooperation), (3) the security techniques and loss prevention electronics techniques used. The theoretical framework of this study based on the social disorganization theory. This study concluded that the organized retail theft, in specific, organized theft is moderately apparent in Riyadh stores. The general result showed that the environment of the stores has an effect on the prevalence of organized retail theft with relation to the gender of thieves, age groups, working shift, type of stolen items as well as the number of thieves in one case. Among other reasons, some factors of the organized theft are: economic pressure of customers based on the location of the store. The dealing of theft also was investigated to have a clear picture of stores dealing with organized retail theft. The result showed that mostly, thieves sent without any action and sometimes given written warning. Very few cases dealt with by police. There are other factors in the study can be looked up in the text. This study suggests solving the problem of organized theft; first is ‘the well distributing of the duties and responsibilities between the employees especially for security purposes’. Second is ‘installation of strong security system’ and ‘making well-designed store layout’. Third is ‘giving training for general employees’ and ‘to give periodically security skills training of employees’. There are other suggestions in the study can be looked up in the text.

Keywords: organized crime, retail, theft, loss prevention, store environment

Procedia PDF Downloads 196

Authors: Sincy Wilson

Abstract:

This article focuses on the data scattered throughout the 2020 Report on migration for a variety of reasons. Despite the fact that climate migrants are no longer recognized on an international or national level, their situation remains unchanged, and many countries have already encountered the problem of people entering their country without permission. With the information presented in the paper, researchers are focusing on climate-induced displacement rather than conflict-related migration. The author finishes by stating that there is no time to waste in recognizing climate migrants.

Keywords: climate refugees, climatological factors, migration, slow-onset migration

Procedia PDF Downloads 213

Authors: Mahmoud Abualsoud

Abstract:

The era of Thutmosis III represents a transitional period between the art of the Thutmoside art and the Amarna period, so we intend to declare that it serves as the cradle of Amarna art. The study will examine the Statues of king Thutmose III that was discovered in Luxor by an Egyptian mission. These Statues have been transferred to the Conservation Center of the Grand Egyptian Museum (GEM) to be conserved and made ready to be displayed at the new museum (the project of the century). We focus on three Statues chosen because they relate to different years of the king's reign. These Statues were all made of granite. The first one is a Kneeling statue representing the god Amun showing king Thutmose III offering to the goddess Hathor. The second is decorated with king Thutmose III with the red crown, between the goddess Hathor and the royal wife, Nefertari. The third shows the king offering NW vessels and bread to the god Seker. Each statue is divided into registers containing a description and decorated with scenes of the king presenting offerings to gods. The proposed study will focus on the development which happened sequentially according to differences that occur in each statue. We will use comparative research to determine the workshops of these statues, whether one or several, and what are the distinguishing features of each one. We will examine what innovations the artisans added to royal art. The description and the texts will be translated with linguistic comments. This research focuses on text analyses and technology. Paleographic information found on these objects includes the names and titles of the king. This research focuses on text analyses and technology. The study aims to create a manual that may help in dating the artwork of Thutmosis III. This research will be beneficial and useful for heritage and ancient civilizations, particularly when we talk about opening museums like the Grand Egyptian Museum, which will exhibit a collection of statues. Indeed, this kind of study will open a new destination in order to know how to identify these collections and how to exhibit them commensurate with the nature of ancient Egyptian history and heritage.

Keywords: archaeological study, Giza, new kingdom, statues, royal art

Procedia PDF Downloads 70

Authors: Jaruwan Chutrtong, Tanakwan Bussabun

Abstract:

Study production of tempeh inoculums powder by freeze-drying comparison with dry at 50°C and the sun bask for developing efficient tempeh inoculums for tempeh producing. Rhizopus oligosporus in PDA slant cultures was incubated at 30°C for 3-5 days until spores and mycelium. Preparation spores suspension with sterilized water and then count the number of started spores. Fill spores suspension in Rice flour and soy flour, mixed with water (in the ratio 10: 7), which is steamed and sterilized at 121°C 15min. Incubated at room temperature for 4 days, count number of spores. Then take the progressive infection and full spore dough to dry at 50°C, sun bask, and lyophilize. Grind to powder. Then pack in plastic bags, stored at 5°C. To investigate quality of inoculums which use different methods, tempeh was fermented every 4 weeks for 24 weeks of the experiment. The result found that rice flour is not suitable to use as raw material in the production of powdered spores. Fungi can growth rarely. Less number of spores and requires more time than soy flour. For drying method, lyophilization is the least possible time. Samples from this method are very hard and very dark and harder to grind than other methods. Drying at 50°C takes longer time than lyophilization but can also set time use for drying. Character of the dry samples is hard solid and brown color, but can be grinded easier. The sun drying takes the longest time, can’t determine the exact time. When the spore powder was used to fermented tempeh immediately, product has similar characters as which use spores that was fresh prepared. The tempeh has normal quality. When spore powder stored at low temperature, tempeh from storage spore in weeks 4, 8 and 12 is still normal. Time spending in production was close to the production of fresh spores. After storage spores for 16 and 20 weeks, tempeh is still normal but growth and sporulation were take longer time than usual (about 6 hours). At 24 week storage, fungal growth is not good, made tempeh looks inferior to normal color, also smell and texture.

Keywords: freez drying, preparation, spores powder, tempeh

Procedia PDF Downloads 202

Authors: Jia-Jane Shuai

Abstract:

Purpose of this research aims to improve understanding of the nature of online religion by examining the religious social websites. What motivates individual users to use the online religious social websites, and which factors affect those motivations. We survey various online religious social websites provided by different religions, especially the Taiwanese folk religion. Based on the theory of the Content Analysis and Social Network Analysis, religious social websites and religious web activities are examined. This research examined the folk religion websites’ presentation and contents that promote the religious use of the Internet in Taiwan. The difference among different religions and religious websites also be compared. First, this study used keywords to examine what types of messages gained the most clicks of “Like”, “Share” and comments on Facebook. Dividing the messages into four media types, namely, text, link, video, and photo, reveal which category receive more likes and comments than the others. Meanwhile, this study analyzed the five dialogic principles of religious websites accessed from mobile phones and also assessed their mobile readiness. Using the five principles of dialogic theory as a basis, do a general survey on the websites with elements of online religion. Second, the project analyzed the characteristics of Taiwanese participants for online religious activities. Grounded by social network analysis and text mining, this study comparatively explores the network structure, interaction pattern, and geographic distribution of users involved in communication networks of the folk religion in social websites and mobile sites. We studied the linkage preference of different religious groups. The difference among different religions and religious websites also be compared. We examined the reasons for the success of these websites, as well as reasons why young users accept new religious media. The outcome of the research will be useful for online religious service providers and non-profit organizations to manage social websites and internet marketing.

Keywords: content analysis, online religion, social network analysis, social websites

Procedia PDF Downloads 167

Authors: Mehrnaz Mostafavi

Abstract:

The assessment and categorization of incidental lung nodules present a considerable challenge in healthcare, often necessitating resource-intensive multiple computed tomography (CT) scans for growth confirmation. This research addresses this issue by introducing a distinct computational approach leveraging radiomics and deep-learning methods. However, understanding local services is essential before implementing these advancements. With diverse tracking methods in place, there is a need for efficient and accurate identification approaches, especially in the context of managing lung nodules alongside pre-existing cancer scenarios. This study explores the integration of text-based algorithms in medical data curation, indicating their efficacy in conjunction with machine learning and deep-learning models for identifying lung nodules. Combining medical images with text data has demonstrated superior data retrieval compared to using each modality independently. While deep learning and text analysis show potential in detecting previously missed nodules, challenges persist, such as increased false positives. The presented research introduces a Structured-Query-Language (SQL) algorithm designed for identifying pulmonary nodules in a tertiary cancer center, externally validated at another hospital. Leveraging natural language processing (NLP) and machine learning, the algorithm categorizes lung nodule reports based on sentence features, aiming to facilitate research and assess clinical pathways. The hypothesis posits that the algorithm can accurately identify lung nodule CT scans and predict concerning nodule features using machine-learning classifiers. Through a retrospective observational study spanning a decade, CT scan reports were collected, and an algorithm was developed to extract and classify data. Results underscore the complexity of lung nodule cohorts in cancer centers, emphasizing the importance of careful evaluation before assuming a metastatic origin. The SQL and NLP algorithms demonstrated high accuracy in identifying lung nodule sentences, indicating potential for local service evaluation and research dataset creation. Machine-learning models exhibited strong accuracy in predicting concerning changes in lung nodule scan reports. While limitations include variability in disease group attribution, the potential for correlation rather than causality in clinical findings, and the need for further external validation, the algorithm's accuracy and potential to support clinical decision-making and healthcare automation represent a significant stride in lung nodule management and research.

Keywords: lung cancer diagnosis, structured-query-language (SQL), natural language processing (NLP), machine learning, CT scans

Procedia PDF Downloads 100

Authors: Julia Ive, Lucia Specia

Abstract:

Understanding and detecting toxicity is an important problem to support safer human interactions online. Our work focuses on the important problem of contextual toxicity detection, where automated classifiers are tasked with determining whether a short textual segment (usually a sentence) is toxic within its conversational context. We use “toxicity” as an umbrella term to denote a number of variants commonly named in the literature, including hate, abuse, offence, among others. Detecting toxicity in context is a non-trivial problem and has been addressed by very few previous studies. These previous studies have analysed the influence of conversational context in human perception of toxicity in controlled experiments and concluded that humans rarely change their judgements in the presence of context. They have also evaluated contextual detection models based on state-of-the-art Deep Learning and Natural Language Processing (NLP) techniques. Counterintuitively, they reached the general conclusion that computational models tend to suffer performance degradation in the presence of context. We challenge these empirical observations by devising better contextual predictive models that also rely on NLP data augmentation techniques to create larger and better data. In our study, we start by further analysing the human perception of toxicity in conversational data (i.e., tweets), in the absence versus presence of context, in this case, previous tweets in the same conversational thread. We observed that the conclusions of previous work on human perception are mainly due to data issues: The contextual data available does not provide sufficient evidence that context is indeed important (even for humans). The data problem is common in current toxicity datasets: cases labelled as toxic are either obviously toxic (i.e., overt toxicity with swear, racist, etc. words), and thus context does is not needed for a decision, or are ambiguous, vague or unclear even in the presence of context; in addition, the data contains labeling inconsistencies. To address this problem, we propose to automatically generate contextual samples where toxicity is not obvious (i.e., covert cases) without context or where different contexts can lead to different toxicity judgements for the same tweet. We generate toxic and non-toxic utterances conditioned on the context or on target tweets using a range of techniques for controlled text generation(e.g., Generative Adversarial Networks and steering techniques). On the contextual detection models, we posit that their poor performance is due to limitations on both of the data they are trained on (same problems stated above) and the architectures they use, which are not able to leverage context in effective ways. To improve on that, we propose text classification architectures that take the hierarchy of conversational utterances into account. In experiments benchmarking ours against previous models on existing and automatically generated data, we show that both data and architectural choices are very important. Our model achieves substantial performance improvements as compared to the baselines that are non-contextual or contextual but agnostic of the conversation structure.

Keywords: contextual toxicity detection, data augmentation, hierarchical text classification models, natural language processing

Procedia PDF Downloads 170