Speech Information Research Articles

An Innovative Novel Transformer Model and Datasets for Safeguarding Religious Sensitivities in Online Social Platforms

Nowadays, blasphemy, inter-religious hatred, and intolerance toward other people's faith become increasing problems in different social media. To effectively tackle these problems, it is necessary to develop efficient techniques for automated detection. With the increasing concern about hate speech on the internet, this research study focuses on creating and testing advanced models to identify and categorize texts related to religious hatred. Leveraging two meticulously curated datasets, the first comprising binary classification tasks to discern hate speech directed towards religion and the second featuring a nuanced five-class categorization including Blasphemy, Ethnic Hatred, Ethnic Group Hatred, Undetermined, and Non-Hatred. The main focus of the research is examining the effectiveness of different established machine learning techniques, deep learning methods, and pre-trained transformer models. Notably, the study introduces hatebnBERT, a novel BERT-based model tailored to the Bangla language's specific nuances and linguistic intricacies specifically built for hate speech detection. Through rigorous evaluation and comparative analysis, the results demonstrate the superior accuracy and precision of hatebnBERT, achieving an accuracy of 98.8% in the binary-class dataset and 98.6% in the multiclass dataset, in effectively identifying and classifying various forms of hate speech within the realm of religious discourse. The outcomes underscore the significance of leveraging specialized language models, such as hatebnBERT, to combat the proliferation of abusive speech and promote a more courteous and inclusive online social media environment within the Bangla-speaking community, considering the potentially devastating consequences of unchecked religious hatred propagation. Insert here your abstract text.

The hatred of all against all? Evidence from online community platforms in South Korea.

Over several years of recent efforts to make sense and detect online hate speech, we still know relatively little about how hateful expressions enter online platforms and whether there are patterns and features characterizing the corpus of hateful speech. In this research, we introduce a new conceptual framework suitable for better capturing the overall scope and dynamics of the current forms of online hateful speech. We adopt several Python-based crawlers to collect a comprehensive data set covering a variety of subjects from a multiplicity of online communities in South Korea. We apply the notions of marginalization and polarization in identifying patterns and dynamics of online hateful speech. Our analyses suggest that polarization driven by political orientation and age difference predominates in the hateful speech in most communities, while marginalization of social minority groups is also salient in other communities. Furthermore, we identify a temporal shift in the trends of online hate from gender to age based, reflecting the changing sociopolitical conditions within the polarization dynamics in South Korea. By expanding our understanding of how hatred shifts and evolves in online communities, our study provides theoretical and practical implications for both researchers and policy-makers.

Construction and Evaluation of a Machine Learning-Based Conversation Model for English Spoken Tourism Conversation

Abstract The development of language macromodeling has led to the widespread adoption of spoken English conversation systems in various industries. Similarly, it has had an impact on the tourism industry, where the involvement of machine learning has taken a big step forward. The purpose of this paper is to explore the use of machine learning methods in spoken dialog systems to construct a dialog model that is suitable for the tourism industry. Speech, map image, and tourism information text data are collected first, which form the database of the dialog model through processing, storage, and application. A bag-of-words model and word vectors represent the text, while speech activity and an English conversation system construct the speech and intention recognition model. On the test set, the best speech recognition rate using double-threshold endpoint detection is up to 98.73%, and the corresponding sentence recognition rate is also up to 93.67%. This paper’s model achieves the highest intention recognition accuracy, scoring 0.943 and 0.955 on the CCKS2018 and SMP2018 datasets, respectively. This paper proposes a machine-learning conversation model that provides technical support for conversational AI development in the tourism industry.

INTERIOR ROOM INFLUENCE ON THE DEFENCE FROM LASER ACOUSTIC INTELLIGENCE SYSTEMS

One of the most vulnerable issues in the technical protection of information is the acquisition of private or confidential data due to leakage through the optoelectronic channel. Today, the most promising area of protection of speech information from laser acoustic intelligence systems is development of passive protection methods. The work is devoted to the influence study of interior elements in the room on the protection of language information from LAIS, that can be implemented effectively and with minimal costs at the design stage in order to comfortably conduct confidential negotiations. On the basis of the conducted research, a systematic approach to determining the security of premises against the leakage of language information by an optoelectronic channel is proposed, which includes technical, interior, technological and design solutions that will help ensure the protection of language information from laser acoustic intelligence systems. Detecting the penetration of laser radiation into the room is solved with the help of the following technical solutions: use of solar panels and lighting sensors, geometry of the arrangement of glass in window frames, application of special coatings. Detection of the influence of room elements on the probing beam is provided with the help of interior solutions: use of materials with a high attenuation coefficient and their location in the room, use of appropriate coloring and composition of absorbing surfaces. Detection of the effect of the masking ability of various glass types is achieved with the help of technological solutions: use of glass with different relief and scattering ability, use of relief protective films. Detection of the impact of room shielding is achieved with the help of design solutions: interior design, taking into account the size and location of windows and their soundproofing characteristics. The direction of further research is the study of the influence of external environmental factors on the acquisition of language information by laser acoustic intelligence systems.

Influence of the atmosphere on the spread of laser radiation

The article is devoted to the problem of protecting information from being read by an optoelectronic channel by studying the influence of atmospheric parameters on the spread of laser radiation. The paper considers key aspects of radiation attenuation due to scattering and absorption by air gases and aerosol particles. It is shown that the efficiency of the probing laser beam depends on extreme climatic conditions, and due to scattering and absorption, the power of the laser beam will decrease as it spreads through the atmosphere. Taking into account the peculiarities of the passage of a laser beam through the atmosphere and the influence of various types of interference on the probing signal, it is possible to predict its attenuation and blocking, namely, for reading confidential speech information, lasers with radiation located in the parts of the atmospheric spectrum occupied by wide windows of transparency should be taken for the appropriate weather conditions; to protect information from being read by an optoelectronic channel, it is necessary to use parameters of the external environment that prevent direct access of the laser beam to the object of information. The presented author's vision of research into the influence of the atmosphere on the spread of laser radiation will allow us to apply the results, ideas, and proposals summarized in this article to solve current problems in the field of information protection.

Speech recognition enhancement based on wireless network sensors application in interactive intelligent teaching system

The speech function used in the multi-dimensional speech recognition model based on MTL and RNN is the commonly used MFCC function, and in feature extraction, after various filtering and conversion, deleting part of the speech information is a requirement for 3D audio recognition. Use multidimensional information as much as possible. Based on this, this article not only brings the development of convolutional neural network intelligent teaching, but also proposes a new education model. Heuristic education like flipped classroom is still the content of education research in recent years, and it is important to the current education form in China. It is mainly based on the traditional lecture mode or the classroom form. This article focuses on undergraduate education. It is designed in the classroom based on the video data of the undergraduate interactive activities and the interactive system. It has important research significance and practicality. At the same time, the interactive content of the traditional teaching method is used to perform voice recognition, and the recognition effect of the main body of the interactive system is further improved through the voice information. The spatial perspective is used to divide the lens perspective into seminars and experimental forms, and the pixel change value of each frame provides a specific reference for it to determine the reflection of student interaction activities and the evaluation of student seminars. This article introduces the implementation process of interactive content and interactive activities of the teaching system in the classroom, introduces the speech recognition framework of the MARS Chinese teaching system, uses the model principle, and compares the general speech data set test with the actual ARST model audio. The results show that, MAS model performance is more advantageous.

Diversified Teaching of English-Chinese Bilingual Courses Based on Integrating Multimodal Discourse Analysis

Abstract This paper uses the BERT model for text feature extraction to capture semantic information more accurately. FBank feature extraction technique is applied for analyzing speech data to improve the accurate recognition of speech information. The 3D-CNN model is built and used to image feature extraction to obtain richer image information. The study combines long and short-term memory networks and attention mechanisms to enhance the interaction and integration between different modalities. The study results show that this multimodal teaching approach not only improves students’ motivation to learn, but also significantly enhances the classroom teaching effect. Especially in vocabulary teaching, using multimodal resources helps deepen students’ understanding and memorization of vocabulary. The model shows high accuracy in the empirical Analysis of pronunciation detection in noisy environments, especially after incorporating multimodal features. The study provides an effective multimodal fusion method for English-Chinese bilingual teaching, especially in improving students’ vocabulary learning efficiency. It has been proved that the multimodal teaching method can better mobilize students’ various senses and thus improve learning. Therefore, the methods and findings of this study are of great theoretical and practical significance to the field of bilingual education.

Optimalisasi Tindak Tutur Direktif Guru Terhadap Pembentukan Karakter Siswa dalam Proses Pembelajaran

This study aims to describe the use of teachers' directive speech acts in students' character-building during the learning process. This research uses a qualitative approach with a case study method. Primary data was collected in the form of speech from teachers in the form of directive speech acts. Primary data sources were obtained through direct observation or observation in the field. Secondary data were collected by recording. The subjects in this study were 11 subject teachers. Secondary data sources were obtained through documentation and direct recording. Data analysis went through six steps, including data collection, data simplification, data presentation, data classification, data interpretation, and conclusion drawing. In this study, the interaction conditions observed include: when entering the classroom, when starting the lesson, during the lesson, when the teacher gives assignments to students when the teacher asks for assignments from students, when the teacher asks students to correct the assignments, when concluding the lesson, and when ending the lesson. The conclusion is that the most dominant directive speech acts used are invitations, orders, and requests. The characteristics formed include discipline, religion, respect, curiosity, enthusiasm, responsibility, independence, hard work, collaboration, caring, and honesty.
 Keywords: directive speech acts, character, learning.

Expressive Speech Act on Pamphlet Demonstration Rejecting the Omnibus Law Bill

Human expression in venting its emotions through language becomes a very interesting thing to discuss and research. In this study, the expression of language contained in the protest pamphlet on the rejection of the Omnibus Law Bill will be classified and analyzed based on its nature and function in expressive speech. In the research method used is qualitative dextriptif. The data that will be examined in the form of writings contained in pamphlets on the omnibus law rejection demonstrations that took place in the center and the region. The techniques used in the processing and analysis of data in this study are with the stages of data collection, data reduction, and data presentation. The results of the study found a form of expressive speech in the pamphlet demonstration of omnibus law rejection. The expressive action on the demonstration pamphlet Reject Omnibus Law based on its nature is divided into two, namely constructive and destructive speech. Both traits are based on speech that builds or decreases the public perception of the existing speech. Expressive speech also has the function of giving greetings, apologizing, praising, insulting, blaming, mocking, and condolences. Abstracts are written in a single paragraph consisting of 200 words, without libraries/citations.

Korean Mothers’ Speech to Young Children with Cochlear Implants in Parent-Child Interaction

Objectives: Infant-directed speech is the particular form of spontaneous speech observed in interactions between parents and their young children. There are reasons to believe that infant-directed speech may help to make language acquisition easier for young children. Thus, this study investigates the effects of cochlear implantation on mothers’ speech to their young children.Methods: Fourteen children with cochlear implants (CIs) and 14 agematched children with normal hearing (NH), aged 12 to 35 months, participated in this study. We recorded mothers’ utterances through a parent-child interaction task. Mothers’ speech features such as fundamental frequency, utterance length, speech rate, and pause duration were measured across speech samples.Results: There were no significant differences between the CI and NH groups in fundamental frequency, utterance length, and pause duration variables. However, the speech of mothers in the CI group was significantly slower than that of the NH group.Conclusion: Mothers used typical infant-directed speech styles when speaking to their children with CIs. Additionally, mothers of the CI group tended to speak more slowly than mothers of the NH group in the parent-child interactions. Given that the language skills of children with CIs are significantly lower than those of children with NH in this study, mothers of the CI group may speak to their children at a slower rate than those of the NH group to provide their linguistic input more efficiently. These findings suggest that children with CIs are exposed to infant-directed speech with similar acoustic qualities as children with NH.

Exploring the Social and Cultural Implications of Hate Speech on Facebook in Malaysia

The study provides insights into the prevalence of hate speech on Facebook among Malaysian users and highlights the need for greater efforts to counteract hate speech in Malaysia. Effective policies and interventions could be developed to address this issue based on the findings. This study examines hate speech trends among Facebook users in Malaysia and their attitudes towards regulating such speech. The study employs a four-part survey questionnaire: Part A covers demographics (gender, age, education); Part B examines social media engagement (platform choice, Facebook usage); Part C explores Facebook hate speech (locations, types, frequency); and Part D investigates responses (experiences, emotions, actions, perspectives). It was conducted with 225 respondents. The study found that race-related content is not the most commonly observed form of hate speech on Facebook in Malaysia; instead, users commonly encounter other types of hate speech. Additionally, users' level of awareness regarding hate speech is influenced by their level of activity on the platform. Notably, the duration of time spent on Facebook is not correlated with the frequency of exposure to hate speech. Experiences of being victimized by hate speech do not necessarily lead to support for regulating hate speech on Facebook. These findings provide valuable insights into the prevalence of hate speech on Facebook among Malaysian users and underscore the need for greater efforts to counteract hate speech in Malaysia. The study suggests the importance of developing effective policies and interventions to address this issue based on the findings.

Hate Speech Through Identity Politics On Social Media Ahead Of The 2024 Presidential Election

The debate surrounding the 2024 General Election in the virtual space, such as social media, has led netizens to exchange hate speech and heavy criticism regarding their respective support. Two groups have emerged: government supporters highlighting the successes of President Joko Widodo and those shifting their support to Anies Baswedan after Prabowo joined the government. Over time, this phenomenon has led to the proliferation of hate speech based on identities such as religion, ethnicity, and fanatical groups, potentially dividing public opinion. This study aims to explain the emergence of hate speech and the forms of hate speech based on identity politics leading up to the 2024 General Election. This study employs virtual ethnographic analysis (Virtual Ethnography) to observe tweets or statuses on selected accounts and the comments that arise from those tweets. The objects of analysis are the trends of words and phrases and the content within observed sentences or paragraphs. The research findings have been presented in the form of tabulation and categorization. The results show that comments and statuses containing hate speech always create bidirectional and multi-directional communication models among users. Hate speech often uses phrases and keywords that lead to insults, malicious accusations, curses, and accusing the opposing party of being stupid and evil. Based on this study, social media users must realize that the presidential campaign is a time to have different choices but remain aware of the same nation. Users need to develop an awareness of social media ethics. Enforcing rules against those who spread hate speech is key for every actor involved.

프랑스의 인터넷상 혐오표현 및 불법정보 규제에 대한 헌법적 검토

France's Hate Speech on the Internet Act introduced different regulatory mechanisms depending on the nature of the information, and provided for imprisonment or fines for information intermediary service providers who violated their obligations under the Hate Speech on the Internet Act.
 In an Ex ante review of the France's Hate Speech on the Internet Act, the French Constitutional Court found that the main provisions of the law violated the right to freedom of expression under the French Constitution. First, the French Constitutional Court found that the restrictions on terrorist or child pornography content violated the right to freedom of expression and communication because the illegality of the content in question was based solely on the judgment of the administrative authority, and because the restrictions could lead to the removal of legitimate content by targeting controversial content. In addition, the French Constitutional Court noted that the law's procedure for identifying illegal information is based solely on reports from internet users, and that internet platform operators face the difficult problem of assessing the content reported against the full range of offenses provided for in the law against hate speech on the internet, In particular, in the case of media crimes, the context of the content in question must be taken into account, and the 24 hours allotted for reviewing the reported content is too short, resulting in internet platform operators choosing to remove the reported content, which violates the right to freedom of expression and communication.
 In the context of recent national legislation to combat hate speech and illegal information on the Internet, the French Constitutional Court's decision and the process of enacting the French Law Against Hate Speech on the Internet may be of interest.

SPEECH ACTS USED BY JO MARCH IN MOVIE “LITTLE WOMEN 2019”

This research studies the speech acts of Jo March the main character in the movie Little Women 2019 and aims to find out the type and form of speech acts in Jo's utterances and which one is the most dominant. The speech act is a theory that examines the meaning of language based on the relationship between utterances and actions performed by speakers and hearers to communicate something. This research uses a qualitative descriptive method to obtain data. The writer collected data by watching the movie of Little Women 2019. In investigating Jo's speech acts, this study applied Searle's (1979) theory of speech acts in analyzing the utterances produced by Jo in the movie Little Women. There are five types of speech acts used by Jo, including assertive with 5 utterances, directive with 10 utterances, commissive with 6 expressive utterances with 9 utterances, and declarative with 2 utterances. So, the amount of data is 32 data and the most dominant speech act used by Jo is the directive speech act.

Tindak Tutur Direktif Guru PAUD EBTA dalam Proses Belajar Mengajar Sebagai Pemerolehan Bahasa Anak 4 Tahun

A child's mastery of a language begins with the acquisition of the first language, which is often called the mother tongue. Apart from the mother tongue, which plays a role in children's language acquisition, the language of teachers at school also plays a role in the development of children's language acquisition and learning. In schools, language use is a reality of teacher-student communication interactions that take place in teaching and learning activities. The teacher's use of language with students in conversations in class, especially the teacher's speech, Speech acts are speech events because they involve parties speaking in certain situations and places. These two symptoms occur in one process, namely the communication process. The teacher's speech acts referred to in this research are directive speech acts. Directive speech acts have an important role in children's language acquisition because they are carried out by the speaker with the intention that the listener performs the action mentioned in the speech. This research aims to (1) describe the forms of directive speech acts used by teachers in the teaching and learning process in language acquisition for 4-year-old children at EBTA Paud and (2) describe teachers' directive speech strategies in the teaching and learning process at EBTA Paud. The data source in this research is the speeches of teachers who teach at EBTA Early Childhood Education. The data collection techniques used in this research are recording techniques and note-taking techniques. The results of the research found that the use of speech acts in the form of directive speech in the teaching and learning process by teachers at EBTA Early Childhood Education was orders, requests, or hopes, suggesting, advising, and challenging. There are two speaking strategies used by teachers in the teaching and learning process at Paud EBTA, namely the strategy of speaking frankly without preamble and the strategy of speaking frankly with positive politeness.

From Epic Tragedy to Tragic Epic: Marina Carr’s Hecuba, Girl on an Altar, and iGirl

This article examines the genre of “epic tragedy” in two of the most recent adaptations of Greek tragedy by Marina Carr, Hecuba (2015) and Girl on an Altar (2022), as well as its reversal to produce a “tragic epic” in iGirl (2021). In the Poetics, Aristotle establishes a stark distinction between the genres of epic and tragedy. Yet, this distinction has been challenged throughout history in dramatic practices – in Euripides and Shakespeare’s tragic plays for instance – and theories – most notably by the emergence of epic theatre coined by Bertolt Brecht. Drawing from all these traditions, this article highlights Carr’s process of “epicisation” of tragedy through the implementation of a new form of dramatic speech and dramatisation of a narrative rather than a plot in Hecuba and Girl on an Altar. These strategies have influenced the creation of iGirl, which is a postmodern epic brought on stage through the lens of tragedy.

Lexical skills in children with and without autism in the context of Arabic diglossia: Evidence from vocabulary and narrative tasks

ABSTRACT Arabic is characterized by diglossia, which involves the use of two language varieties within a single speech community: Spoken Arabic (SpA) for everyday speech and Modern Standard Arabic (MSA) for formal speech and reading/writing. Earlier research suggests that some Arabic-speaking children with Autism spectrum disorder (ASD) might favor MSA over SpA in informal settings, unlike those with Typical Language Development (TLD). Research on typical and atypical lexical abilities within the diglossic context of Arabic is limited. This study investigated the manifestation of diglossia in the lexical skills of Palestinian-Arabic speaking children with and without ASD. A total of 57 Palestinian-Arabic speaking children aged 4-11 were evaluated, comprising three groups: TLD, ASD with language Impairment (LI) (ASD+LI), and ASD with normal language skills (ASD+NL). Two vocabulary tasks were used: comprehension and production of nouns and verbs that manipulated the lexico-phonological distance between SpA and MSA. Moreover, a LITMUS-MAIN narrative production task in SpA was administered to all children. The vocabulary comprehension and production tasks were coded for accuracy of response. The production tasks (vocabulary and narrative) were also coded using a special coding rubric according to which all words used were categorized as either SpA or MSA. No differences emerged between the TLD and ASD+NL groups in any vocabulary measure, yet lower scores were observed in the ASD+LI group. In production tasks, all groups utilized SpA more than MSA. Nonetheless, some children with ASD overused MSA lexis. These findings highlight the heterogeneity of lexical skills among children with ASD. The results further underscore the role of diglossia in Arabic in lexical abilities in autism, and thus call for thorough language assessment procedures across different domains.

Informal speech in primary school children in their leisure time: the influence of Spanish or French on their spontaneous expressivity

This study was funded by the Mintzola Project of the Royal Academy of the Basque Language (Euskaltzaindia), by the Basque School Association (Ikastolen Elkartea) and by Mondragon Unibertsitatea. Our purpose was to examine whether speech among friends ina formal setting can have a positive effect on increasing informal speech in Basque among students. We present the first steps of this research here. Ten schools (ikastolas) from different territories in the Basque Country were selected, and the language habits of students in their fourth year of Primary Education at home, outside the home and at school were initially examined via a questionnaire in order to obtain baseline information about the sample in as much detail as possible. We then observed the participating children in the recreational areas of the schools in order to determine which language(s) they speak with their friends during their free time.

Analisis Retorik Pidato Politik Megawati Soekarno Putri Pada Peringatan Ke-50 Tahun PDI-P

Megawati Soekarno Putri's existence on the Indonesian political stage has been going on for more than two decades. During this period, Megawati has managed to maintain her position as a charismatic female political figure. Despite public doubts about her figure, Megawati is the de facto leader of one of the largest parties, as well as the only female president who has ever led the Republic of Indonesia. As a politician who has a strong character, Megawati has never escaped the media spotlight. Through her various controversial statements, Megawati's remarks are cited as a hot topic of discussion in the media. Reflecting on his charisma and rhetorical power, Megawati's speech is an interesting rhetorical phenomenon to analyze. Speech by Megawati Soekarno Putri on the Birthday of the Indonesian Democratic Party – Struggle (PDI Perjuangan), was a form of ceremonial speech filled with political messages. Megawati's speech was delivered in a heated political atmosphere, ahead of the contestation moment for the 2024 Presidential Election. This study aims to analyze the rhetorical strategy used by Megawati to reveal the implicit meaning contained in it. Referring to the rhetorical analysis method, the results of the study show that Megawati's speech applies the typical communication style of female leaders. Apart from that, Megawati's speech also applied a rhetorical strategy that was oriented towards emphasizing aspects of personal identity as a descendant of Bung Karno, authority as chairman of the PDI-P, and his superiority over President Jokowi as his cadre. Megawati's rhetorical strategy also shows that she is not an inferior female leader.

Features of the speech processing network in post- and prelingually deaf cochlear implant users.

The onset of hearing loss can lead to altered brain structure and functions. However, hearing restoration may also result in distinct cortical reorganization. A differential pattern of functional remodeling was observed between post- and prelingual cochlear implant users, but it remains unclear how these speech processing networks are reorganized after cochlear implantation. To explore the impact of language acquisition and hearing restoration on speech perception in cochlear implant users, we conducted assessments of brain activation, functional connectivity, and graph theory-based analysis using functional near-infrared spectroscopy. We examined the effects of speech-in-noise stimuli on three groups: postlingual cochlear implant users (n = 12), prelingual cochlear implant users (n = 10), and age-matched individuals with hearing controls (HC) (n = 22). The activation of auditory-related areas in cochlear implant users showed a lower response compared with the HC group. Wernicke's area and Broca's area demonstrated differences network attributes in speech processing networks in post- and prelingual cochlear implant users. In addition, cochlear implant users maintain a high efficiency of the speech processing network to process speech information. Taken together, our results characterize the speech processing networks, in varying noise environments, in post- and prelingual cochlear implant users and provide new insights for theories of how implantation modes impact remodeling of the speech processing functional networks.