External Language Research Articles

Cross-lingual Natural Language Processing on Limited Annotated Case/Radiology Reports in English and Japanese: Insights from the Real-MedNLP Workshop.

Textual datasets (corpora) are crucial for the application of natural language processing (NLP) models. However, corpus creation in the medical field is challenging, primarily because of privacy issues with raw clinical data such as health records. Thus, the existing clinical corpora are generally small and scarce. Medical NLP (MedNLP) methodologies perform well with limited data availability. We present the outcomes of the Real-MedNLP workshop, which was conducted using limited and parallel medical corpora. Real-MedNLP exhibits three distinct characteristics: (1) Limited Annotated Documents: The training data comprises only a small set (approximately 100) of case reports (CRs) and radiology reports (RRs) that have been annotated. (2) Bilingually Parallel: The constructed corpora are parallel in Japanese and English. (3) Practical Tasks: The workshop addresses fundamental tasks, such as named entity recognition and applied practical tasks. We propose three tasks: named entity recognition (NER) of approximately 100 available documents (Task 1), NER based only on annotation guidelines for humans (Task 2), and clinical applications (Task 3) consisting of adverse drug effects (ADE) detection for CRs and identical case identification (CI) for RRs. Nine teams participated in this study. The best systems achieved 0.65 and 0.89 F1-scores for CRs and RRs in Task 1, whereas the top scores in Task 2 decreased by 50-70%. In Task 3, ADE reports were detected by up to 0.64 F1-score, and CI scored up to 0.96 binary accuracy. Most systems adopt medical-domain-specific pre-trained language models using data augmentation methods. Despite the challenge of limited corpus size in Tasks 1 and 2, recent approaches are promising because the partial match scores reached approximately 0.8-0.9 F1-scores. Task 3 applications revealed that the different availabilities of external language resources affected the performance per language.

Vocabulary Development and Kinship Level of Javanese Dialect: Dialect Geography Studies

This study aims to explore the extent of vocabulary and kinship changes in Javanese dialects in Yogyakarta Special Region between 2019 and 2023. Data tapping was conducted at a number of observation points, namely in the villages of Kebonrejo (Kulonprogo), Pakem Binangun (Sleman), Segoroyoso (Bantul), Planjan (Gunungkidul), and Ngringin (Gunungkidul). The data were obtained using the basic Morris Swadesh vocabulary, the universal vocabulary of the world's languages. By looking at the data obtained from the same observation points in 2019 and 2023, it is possible to compare the vocabulary developed in Javanese dialects. From the data comparison, it is also possible to compare the level of kinship between dialects to see if there is a shift in their status from different speech, subdialect, dialect, or language. The data were obtained by tapping method with recording and note-taking techniques, then analyzed by comparison, introspection, and dialectometry methods, and presented by formal methods in the form of tables and figures and informal methods in the form of ordinary narratives. The results show changes in vocabulary and kinship levels caused by internal elements, namely vocabulary development, heterogloss, and dialectometry, and external language, namely factors of dialect supporting communities, dialect area development, and the Covid-19 pandemic.

X-Phishing-Writer: A Framework for Cross-lingual Phishing E-mail Generation

Cybercrime is projected to cause annual business losses of $10.5 trillion by 2025, a significant concern given that a majority of security breaches are due to human errors, especially through phishing attacks. The rapid increase in daily identified phishing sites over the past decade underscores the pressing need to enhance defenses against such attacks. Social Engineering Drills (SEDs) are essential in raising awareness about phishing yet face challenges in creating effective and diverse phishing e-mail content. These challenges are exacerbated by the limited availability of public datasets and concerns over using external language models like ChatGPT for phishing e-mail generation. To address these issues, this article introduces X-Phishing-Writer, a novel cross-lingual Few-shot phishing e-mail generation framework. X-Phishing-Writer allows for the generation of e-mails based on minimal user input, leverages single-language datasets for multilingual e-mail generation, and is designed for internal deployment using a lightweight, open-source language model. Incorporating Adapters into an Encoder–Decoder architecture, X-Phishing-Writer marks a significant advancement in the field, demonstrating superior performance in generating phishing e-mails across 25 languages when compared to baseline models. Experimental results and real-world drills involving 1,682 users showcase a 17.67% e-mail open rate and a 13.33% hyperlink click-through rate, affirming the framework’s effectiveness and practicality in enhancing phishing awareness and defense.

Topic Modelling: Going beyond Token Outputs

Topic modelling is a text mining technique for identifying salient themes from a number of documents. The output is commonly a set of topics consisting of isolated tokens that often co-occur in such documents. Manual effort is often associated with interpreting a topic’s description from such tokens. However, from a human’s perspective, such outputs may not adequately provide enough information to infer the meaning of the topics; thus, their interpretability is often inaccurately understood. Although several studies have attempted to automatically extend topic descriptions as a means of enhancing the interpretation of topic models, they rely on external language sources that may become unavailable, must be kept up to date to generate relevant results, and present privacy issues when training on or processing data. This paper presents a novel approach towards extending the output of traditional topic modelling methods beyond a list of isolated tokens. This approach removes the dependence on external sources by using the textual data themselves by extracting high-scoring keywords and mapping them to the topic model’s token outputs. To compare how the proposed method benchmarks against the state of the art, a comparative analysis against results produced by Large Language Models (LLMs) is presented. Such results report that the proposed method resonates with the thematic coverage found in LLMs and often surpasses such models by bridging the gap between broad thematic elements and granular details. In addition, to demonstrate and reinforce the generalisation of the proposed method, the approach was further evaluated using two other topic modelling methods as the underlying models and when using a heterogeneous unseen dataset. To measure the interpretability of the proposed outputs against those of the traditional topic modelling approach, independent annotators manually scored each output based on their quality and usefulness as well as the efficiency of the annotation task. The proposed approach demonstrated higher quality and usefulness, as well as higher efficiency in the annotation task, in comparison to the outputs of a traditional topic modelling method, demonstrating an increase in their interpretability.

Leveraging the AAPPL to promote positive washback in K‐12 language teaching

AbstractIn recent years, the use of external proficiency‐based tests to measure learners' language ability has increased significantly, in part thanks to the Seal of Biliteracy movement. However, there exists very limited research investigating the influence of such tests and their subsequent outcomes data on teachers' practices. As such, this study examines the washback effects of the ACTFL Assessment of Performance toward Proficiency in Languages (AAPPL) on teachers' instruction and planning in an urban Midwest K‐12 district. Survey and interview data revealed that teachers increased their use of ACTFL key documents, placed more emphasis on target language usage, and sought to balance their use of the three modes of communication to better meet learners' needs and facilitate their proficiency development. This evidence of positive washback was observed among teachers across languages and levels and was enhanced by the opt‐in approach to test implementation and the close link between testing and teaching. For some, the AAPPL served as a catalyst to further align their curricula and instruction. Recommendations are offered for those who may wish to implement a standards‐based external language assessment such as the AAPPL as a means by which to foster professional growth and promote teachers' adoption of proficiency‐oriented practices.

Inner Voice in Silent Reading and Speaking

Current research reveals that individuals use two languages: external language and internal language. The external language is a loud, noisy language that everyone can hear. The internal language is a silent, mysterious, personal language that merely the individual hears. This phenomenon, called inner language, inner voice or inner speech, is a mental activity seen in most people. This activity plays important roles in language and cognitive processes, helps the processes such as reading, writing, understanding, questioning, learning, thinking, controlling emotions, motivation, and so on. The inner voice emerges during silent reading and contributes significantly to understanding words, sentences, and text. Expert readers read and understand the words by hearing the small inner sound of the words in their brain. Controversy, internal speech is when an individual talks silently to himself. Inner speech is referred when planning daily tasks, making decisions, solving problems, and leading social and emotional issues. Nonetheless, research on the inner voice and inner speech, which are effective in the development of the individual, is yet insufficient. There is a need for clearer and more comprehensive research on the role and function of inner speech in processes such as reading, speaking, understanding, and thinking. It is recommended that this phenomenon, which exists in every individual, be handled scientifically and used to develop language and cognitive skills.

Modern methods of teaching foreign languages

The purpose of the study is to assess the possibility of developing and further using in practice a comprehensive concept of teaching foreign languages based on traditional and modern pedagogical methods and characterized by the greatest effectiveness in terms of adaptation to the requirements of the external environment. The subject of the study is the problem of finding new methods of teaching languages that could fully correspond to the dynamic structure of the modern educational concept. Special attention is paid to the methods of motivating the process of learning the language system. New pedagogical technologies have been selected as progressive forms of teaching, the essence of the implementation of which is given in the study. The concept of the term "Smart education" is disclosed, since its introduction and widespread use contributes to improving the effectiveness and accessibility of educational programs for a large number of users. As an evidence-based advantage, this study accepts the idea of universal digitalization of linguistic knowledge, the ubiquity of mobile technologies, which inevitably creates stable prerequisites for the transition from book sources of information to the active development of digital capabilities of the external language environment. In this regard, the objective expediency of using new educational methods, in particular, related to the process of digitalization and universal accessibility of mobile phones, as well as the possibilities of the Internet, attracts the close attention of the modern educational community.

A sociolinguistic discourse analysis of assimilated English words: a usage-based model of language acquisition

This study will explore how some words from a different language get assimilated into the dialogues of a community or group which speaks a totally different language or dialect. The Angami Nagas community of Nagaland is adopted as the case in point for this study, for they use a number of English words in their regular speaking irrespective of being English educated or not. The study is qualitative based research and the methodology includes the research methods of interviews and document analyses. The study reveals that the understanding of an external language through the use of assimilated words by a community or group with diverse associations with the source language, can lead to a usage-based model of language acquisition. The study discloses that diversity in the articulation of assimilated words cannot be strictly ruled out even in the presence of precision in usage and meaning. It is also found that, overtime, some assimilated words, in its original essence, get attached and identified with a group of things by the user community. The study concludes on the success of using assimilated words within a community or group with accuracy or near accuracy and assigns it as a usage-based model of basic language assimilation which can get remodeled in future studies of more formal language learning. It infers that assimilated words from an external language, used in the correct context, can lead to successful conversation or speech in varying contexts as the words become part of the communication to convey concepts and thoughts. The numerous works on the Angami Nagas, by and large, deal with their socio-cultural, economic and political status as an ethnic community and particularly with mostly their contribution and standing during the British times. In a diverging trend, this study is a pioneering exploration of the sociolinguistic pattern among the Angami Nagas in employing assimilated English words and phrases in their everyday usage.

Internet-Based Lifelong Learning: Implications for Language Teaching and Semiotic Structure Principles in Media Sociology and Religion Sociology

In the course of science theory, when we deal with the subject of language teaching in the context of semiotic structure principle developments as well as in the course of the media culture and religious sciences, it shows that in numerous contributions concerning this subject area there is a complaint that the aporetic treatment and integration of language and religious history [Sprachgeschichte/Religionsgeschichte] in the sense of an internal and external language and religious history should be taken into account [1]. The associated diachrony of historical linguistics is said to be not only relevant but also sufficiently socially systemic and imminently codependent for its constitution [2]. There is talk of a "step child," of "marginal permanence," and even of a malicious deletion from facultative curricula since the international history of language and language teaching is said to be hardly fruitful, especially in view of a problem-oriented approach, obviously not having any relevance for the present constantly transforming, media- and religion-sociology-oriented reality of life and work (paper presented at the Ecumenical and Pedagogy of Religion Colloquium, "Dimensions of Education," at the Nicolaus Copernicus University in Torun, on October 11, 2007, in Torun). The author of this paper does not wish to join in this lament about the "Geschichtslosigkeit und Religionslosigkeit" [historical unawareness] regarding time and language, with its sustainable effect on international language and religion teaching, since in recent years, there have been some astonishing changes in perspective, signaling a big opportunity in terms of a substantial paradigm shift for international language teaching – the subject of lifelong learning.

Efficient Conformer for Agglutinative Language ASR Model Using Low-Rank Approximation and Balanced Softmax

Recently, the performance of end-to-end speech recognition has been further improved based on the proposed Conformer framework, which has also been widely used in the field of speech recognition. However, the Conformer model is mostly applied to very widespread languages, such as Chinese and English, and rarely applied to speech recognition of Central and West Asian agglutinative languages. There are more network parameters in the Conformer end-to-end speech recognition model, so the structure of the model is complex, and it consumes more resources. At the same time, we found that there is a long-tail problem in Kazakh, i.e., the distribution of high-frequency words and low-frequency words is not uniform, which makes the recognition accuracy of the model low. For these reasons, we made the following improvements to the Conformer baseline model. First, we constructed a low-rank multi-head self-attention encoder and decoder using low-rank approximation decomposition to reduce the number of parameters of the multi-head self-attention module and model’s storage space. Second, to alleviate the long-tail problem in Kazakh, the original softmax function was replaced by a balanced softmax function in the Conformer model; Third, we use connectionist temporal classification (CTC) as an auxiliary task to speed up the model training and build a multi-task lightweight but efficient Conformer speech recognition model with hybrid CTC/Attention. To evaluate the effectiveness of the proposed model, we conduct experiments on the open-source Kazakh language dataset, during which no external language model is used, and the number of parameters is relatively compressed by 7.4% and the storage space is relatively reduced by 13.5 MB, while the training speed and word error rate remain basically unchanged.

일식조리사의 커뮤니케이션 유형이 자기효능감 및 조직시민행동에 미치는 영향

This study aimed to verify how the communication type of Japanese chefs affects their self-efficacy and organizational citizenship behavior due to the improvement in satisfaction caused by smooth communication between visiting customers and colleagues. To this end, an empirical analysis was conducted targeting Japanese cooks working at Japanese restaurants in Bueul-gyeong (Gyeongnam, Busan, Ulsan). The results and implications of the empirical analysis are presented as follows. First, the type of communication (verbal communication, body language, spatial language, external language) of Japanese chefs had a positive (+) effect on self-efficacy. Second, Japanese cooks

A customized residual neural network and bi-directional gated recurrent unit-based automatic speech recognition model

Speech recognition aims to convert human speech into text and has applications in security, healthcare, commerce, automobiles, and technology, just to name a few. Inserting residual neural networks before recurrent neural network cells improves accuracy and cuts training time by a good margin. Furthermore, layer normalization instead of batch normalization is more effective in model training and performance enhancement. Also, the size of the datasets presents tremendous influences in achieving the best performance. Leveraging these tricks, this article proposes an automatic speech recognition model with a stacked five layers of customized Residual Convolution Neural Network and seven layers of Bi-Directional Gated Recurrent Units, including a logarithmic softmax for the model output. Each of them incorporates a learnable per-element affine parameter-based layer normalization technique. The training and testing of the new model were conducted on the LibriSpeech corpus and LJ Speech dataset. The experimental results demonstrate a character error rate (CER) of 4.7 and 3.61% on the two datasets, respectively, with only 33 million parameters without the requirement of any external language model.

Fikrah Rusydi Ahmad Th’aimah wa Mahmud Kamil al-Naqah fi Ta’lim al-Lughah al-‘Arabiyyah ‘Inda Dhu’i al-Madkhal al-Itthishal

This article is written to find out and analyze Rusydi Ahmad Thu’aimah’s and Mahmud Kamil An-Naqah’s thoughts on teaching the Arabic language using a communication approach. This article is written resting upon a library research and a content analysis approach. The findings indicate that: 1) There are four kinds of learning elements according to Rusydi Ahmad Thu’aimah and Mahmud Kamil An-Naqah. They are messages, the sender, the receiver, and learning media. 2) There are three kinds of learning strategies according to Rusydi Ahmad Thu’aimah and Mahmud Kamil An-Naqah. They are metacognitive strategy, cognitive strategy, and sosio-affective strategy. 3) There are two kinds of problems in learning the Arabic language using the communication approach according to Rusydi Ahmad Thu’aimah and Mahmud Kamil An-Naqah. They are a) the problem in the same language, which is divided into sound, vocabulary, word, sentence, and writing; and b) the problem of external language, which is divided into the educational materials, the teacher, and the learning media.

Secular Sciences in an Era of Cultural Change: India, 1206–1526—A Survey

The establishment of the Delhi Sultanate, 1206, marked the infusion of a new cultural stream to which the religiously neutral term ‘Persianate’ is now attached. Indeed, it added to Sanskrit another language of learning in India, namely, Persian. The change, however, was not just one of an addition to India’s list of literary languages. There was a recognisable extension of knowledge (and area of reflection) alongside the intrusion of an external language. The present article attempts a survey of the developments in different aspects of science (including practical activity related to it) during the period of the Sultanate (1206–1526), considered in two parts, according to the respective languages of the texts, namely, Sanskrit and Persian.

Momentum Pseudo-Labeling: Semi-Supervised ASR With Continuously Improving Pseudo-Labels

End-to-end automatic speech recognition (ASR) has become a popular alternative to traditional module-based systems, simplifying the model-building process with a single deep neural network architecture. However, the training of end-to-end ASR systems is generally data-hungry: a large amount of labeled data (speech-text pairs) is necessary to learn direct speech-to-text conversion effectively. To make the training less dependent on labeled data, pseudo-labeling, a semi-supervised learning approach, has been successfully introduced to end-to-end ASR, where a seed model is self-trained with pseudo-labels generated from unlabeled (speech-only) data. Here, we propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">momentum pseudo-labeling</i> (MPL), a simple yet effective strategy for semi-supervised ASR. MPL consists of a pair of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">online</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">offline</i> models that interact and learn from each other, inspired by the mean teacher method. The online model is trained to predict pseudo-labels generated on the fly by the offline model. The offline model maintains an exponential moving average of the online model parameters. The interaction between the two models allows better ASR training on unlabeled data by continuously improving the quality of pseudo-labels. We apply MPL to a connectionist temporal classification-based model and evaluate it on various semi-supervised scenarios with varying amounts of data or domain mismatch. The results demonstrate that MPL significantly improves the seed model by stabilizing the training on unlabeled data. Moreover, we present additional techniques, e.g., the use of Conformer and an external language model, to further enhance MPL, which leads to better performance than other semi-supervised methods based on pseudo-labeling.

MONIZA ALVI’S DIASPORIC SENSIBILITY IN CONSTRUCTING GENDER IN SPLIT WORLD POEMS AND BLACK BIRD BYE BYE

Dislocation form a foremost concern for all the colonized indigenous people, acting as a model for the phenomenon of diaspora. This results not just in cultural assignation but cultural rotation as well. The research derives guide from Ashcroft, Tiffin and Grifith’s seminal work, The Empire Writes Back in examining how dislocation from a place creates concerns concerning identity and authenticity on the behalf of the writer in question. This particular methodology emphasizes on the appropriateness of an external language for the description of indigenous people in postcolonial diaspora literature. Krippendorff’s textual analysis method, “Content Analysis” is used to explore and collect the themes in relativeness to women, men and place from the poetic works of Moniza Alvi along with a postcolonial theory in the background. The present study discusses how Moniza Alvi’s diasporic sensibility discursively constructs indigenous female in the Split World Poems and male in Black Bird Bye Bye, respectively. Keywords: diaspora, discursive, indigenous, place, gender, postcolonial.

A simplified reinforcement and fracture mechanism analysis model of epoxy nanocomposites based on finite element simulation

Understanding the reinforcement and fracture mechanism of polymer matrix nanocomposites is crucial for the design of high-performance composite materials. By taking boron nitride@boronate polymer core-shell nanoparticles modified epoxy resin nanocomposites (EPNCs) as an example, a reinforcement and fracture mechanism analysis model of nanocomposites was established through the combination of experiments and simulations. Instead of the traditional programming method by using the external language, a new modeling method was introduced with the help of Digimat. The results indicated that the interface force between (within) EP matrix and nanoparticles, the shell thickness and nanoparticle content play important roles in the reinforcement and fracture of EPNCs. The bending strength increases with the raising of particle content, but decreases with the increasing of shell thickness. The simulated results were consistent with the experiments. The interface force evolves according to the order of interfaces within particles and their shell > matrix-shell > matrix interior. A proper interfacial force can make interfacial debonding and matrix failure occur at the same time, thus improving both the strength and toughness of nanocomposites. This model construction will provide theoretical guidance for nanocomposites design and the clarification of reinforcement mechanism. • Cross-examination of analytical, numerical and experimental results is conducted. • Appropriate interface, thin shell and particle content play important roles in composites reinforcement mechanism. • Finite element evaluation of core-shell nanoparticles reinforced epoxy resin composites wais established.

What good is a supercentral language?

The article focuses on shifts in justificatory repertoires within France for the promotion of the French language beyond the regions of the world where it is securely the core language. It is based on a corpus of twenty French public policy reports on external language and cultural policies published between 1998 and 2018. The corpus is characterised by a thick referential interplay, and its authors work within a loose and evolving advocacy coalition. The reports suggest that France’s position as a “supercentral” world language is characterised by significant fragility, and that its defenders must break with former representations of French as a prestigious language of culture and diplomacy. They argue that it must “equip” itself as a manifestly useful, broadly attractive and polycentric language in a new geolinguistic space. At an initial level, they attempt to justify its external projection by framing French with its specific attributes and reach as a sui generis global public good. They supplement this stock of justifications by invoking a privileged role for French in the defence of the second-order global public good of linguistic diversity as such. These discursive moves produce both ruptures and continuities in relation to the French cultural and language policy traditions of rayonnement and universalism.

일식레스토랑 조리종사자의 비언어적 커뮤니케이션이 대인매력과 관계지속의도에 미치는 영향

The purpose of this study is to examine the effect of non-verbal communication(body language, spatial language, auditory language, external language) on interpersonal attractiveness and relationship continuity intention of cooks in Japanese restaurants. This study conducted a survey on customers performed with SPSS 20.0 and AMOS 20.0. First, it was confirmed that non-verbal communication (body language, spatial language, auditory language, external language) of cooks at Japanese restaurants had a positive (+) effect on interpersonal attractiveness. Second, it was confirmed that interpersonal attractiveness had a positive (+) effect on relationship continuity intention. Third, it was confirmed that non-verbal communication (body language, spatial language, auditory language, external language) of cooks at Japanese restaurants had a positive (+) effect on relationship continuation intention. This study conducted a survey on customers who lived in Busan, Ulsan, and Gyeongsangnam-do and had visited Japanese restaurants twice or more within the last 6 months.

Ostension and the communicative function of natural language

Linguistic pragmatics has traditionally focused on the ways in which speaker-intended meanings are contextually inferred in solid communicative contexts. In this paper, I explore the possibility that all linguistic utterances necessitate (at least some) pragmatic processing, even when they are not transparently embedded in a speaker's overt communicative behavior. Following a relevance-theoretic rationale, I argue that, being externalized objects with special characteristics which set them apart from their internalized mental counterparts – most notably the linguistic underdeterminacy of their propositional form, linguistic utterances automatically trigger the dedicated mental module that is responsible for pragmatic inference, because they would otherwise not have evolved to invariably pre-empt our attention, as they typically do. In this picture, external language's evolutionary advantage lies precisely in its unique ability to reveal meaning with an amount of precision that cannot be reached through the use of other types of stimuli, while at the same time avoiding unnecessary expressive prolixity. Viewed in this way, contextual inference becomes an underlying feature of the way in which external languages link up with our internalized linguistic capacity, in an approach that reinforces the possibility of incorporating it in the study of meaning over and above the remit of overt communication.