Search

This study proposes a data-driven personalized diagnostic feedback strategy (D-DPDFS) based on data, and its purpose is to explore the effects that this strategy brings on pupils' writing performance, epistemic network structure, and self-efficacy. In this study, the participating pupils were randomly divided into an experimental group (N = 39) and a control group (N = 39), in which the experimental group used the data-driven personalized diagnostic feedback strategy and the control group used the Experience-Driven Teacher Diagnostic Feedback Strategy (E-DTDFS), and the writing activity lasted for 6 weeks. The results of the study showed that the D-DPDFS strategy had higher performance in promoting the four aspects of students' writing and enhanced the self-efficacy of the pupils in the experimental group. Meanwhile, the results based on epistemic network analysis showed that pupils using the DDPDFS focused on describing the action, psychology, and emotional changes in vocabulary use, and their epistemic network structure was more complex. After the interviews, the pupils in the experimental group were able to accurately understand and reflect deeply on the writing problem, and their satisfaction was high. Therefore, we believe that this study brings new insights into improving the teaching and learning of writing. [ABSTRACT FROM AUTHOR]

This empirical investigation rigorously evaluated the teaching potential of Data-Driven Learning (DDL) for enhancing the ability of learners to apply 30 near-synonymous change-of-state verbs in two EFL Freshman English courses in a college setting across two semesters. In the study, 32 participants in the experimental group and 22 in the control group were involved in two distinct educational approaches: Data-Driven Learning (DDL) guided by constructivism, facilitated by the Corpus of Contemporary American English (COCA), and conventional rule-based instruction, respectively. The students' skill levels were measured at three separate stages: pre-intervention, post-intervention, and a delayed post-intervention (3 months). The quantitative findings revealed that DDL notably strengthened learners' collocational competence following the intervention, with the effect enduring even 3 months later. However, the impact was not substantial for high-complexity verbs. Additionally, DDL seemed to promote a varied utilization of collocates of a singular node word by students in the experimental group, a result maintained at the delayed posttest. The paper concludes with an exploration of the pedagogical consequences of DDL, specifically in relation to collocation instruction for English as a Foreign Language (EFL) learners. The insights emphasize the potential of DDL as an efficacious language education instrument, underscoring its importance in boosting the linguistic aptitude of EFL learners within academic contexts. Hence, the study accentuates DDL's anticipated function in the instructional framework of second language acquisition, calling for additional scrutiny to refine its execution. [ABSTRACT FROM AUTHOR]

This paper uses corpus data to analyse spoken expressions and discourse markers in Czech, applying these findings to corpus-based exercises for learners of Czech as a foreign language. The analytical section highlights the usefulness of parallel corpus in identifying suitable translation equivalents for prevalent Czech spoken vocabulary in English, French, and Polish as native languages from the learner’s perspective. The methodology outlines the process of finding appropriate translation equivalents in film subtitles, considering both meaning and spoken register. The pedagogical section introduces three corpus-based exercises designed to improve conversational skills, featuring authentic texts that familiarise learners with spoken vocabulary. This research builds on previous studies of the English language that did not use parallel corpora to identify translation equivalents in learners’ native languages — an essential factor for understanding a foreign language. In addition, tailor-made corpus-based exercises can be seamlessly integrated into everyday classroom activities to enhance language awareness among non-native speakers.

Abstract The purpose of the current study was twofold: examining the efficacy of data-driven learning (DDL) (hands-on and hands-off approaches) in the realization of interactional metadiscourse markers (IMMs) among English as a foreign language (EFL) learners and analyzing the learners’ perceptions of DDL. The participants consisted of 93 male and female advanced language learners randomly assigned to one of the three groups: hands-on, hands-off, and control. Throughout the duration of treatment lasting for 10 sessions, the hands-on group employed the use of Microsoft Copilot, artificial intelligence (AI) chatbot, on a computer screen to discuss and explore IMMs, but the hands-off group was exposed to IMMs through written texts that were physically printed on paper and articles to be examined through AntConc concordancing program. The control group received conventional instructional techniques including reading assigned course materials. The findings from a one-way analysis of covariance (ANCOVA) procedure indicated that both experimental groups outperformed the control group in the posttest of realizing and identifying IMMs. However, the post hoc comparisons showed statistically significant differences between the hands-on and hands-off groups, with the hands-on group performing more successfully in identifying IMMs. The results of the questionnaire data revealed that all the learners had positive perception of DDL. The results of the current study suggest using both hands-on and hands-off DDL methods helps learners develop their writing performance through metadiscourse realization.

The purpose of the present mixed methods study was to investigate if enhancing Focus on Form (FonF) tasks with data-driven learning (DDL) affects English as a Foreign Language (EFL) learners' task motivation from the self-determination theory (SDT) perspective. Following a quasi-experimental comparison group pretest-posttest and sequential explanatory mixed methods designs, 76 female English-major university students were randomly assigned to comparison and intervention groups and exposed to the control and treatment, respectively. The control involved completing 15 non-DDL tasks, whereas the treatment was completing 15 DDL-enhanced tasks over a month. The results of a task motivation questionnaire and semi-structured interviews showed that DDL-enhanced tasks increase learners' task motivation from the SDT perspective. In other words, enhancing FonF tasks with DDL leads to more self-determined motivation to perform English tasks. That is, discovery learning which is a kind of self-regulated learning, along with awareness raising, in tasks with authentic language data where the primary attention is on meaning, tends to promote learners' agency, and result in more motivation to learn English. This kind of motivation which seems to be the result of increased achievement in English language skills, autonomous language learning, association with others, internal reward, and usefulness, is most influential, as it is self-determined. [ABSTRACT FROM AUTHOR]

The general purpose of this study is to investigate the effect of data-driven learning (DDL) on the development of English language learners' writing skills. "Writing skills" in the present study are limited to the formal or structural aspects of language, the way linguistic elements, words, phrases, clauses, and sentences form larger units of language to convey ideas and concepts. The objectives of the study were to compare the learning effects of DDL method with the conventional teaching method's effects on the measures of learners' declarative knowledge of the taught materials, and analytic scoring of their written products. A pre-test and post-test control group research design was used to collect the required data. Two groups of students who participated in the "Paragraph Writing" course were compared in terms of writing skills at the micro level. The control group was trained with the conventional method of using textbooks, teacher's explanations and classroom exercises. The experimental group, in addition to textbooks, received lessons prepared based on concordance lines. Statistical analyzes showed that there is a significant difference between the two groups in terms of declarative knowledge. This result can be interpreted as an indication of superiority of DDL-based courses over conventional textbooks in terms of learners' (declarative knowledge) writing skills development. The results of the analytical scores showed that there is no statistically significant difference between the two groups in terms of "content", "vocabulary" and "organization". However, the DDL group showed more knowledge in "Language Use", which indicates the greater advantage of DDL-based lessons in learning and applying grammar patterns. The practical implications of this research are presented in three separate but related areas of using corpora in language instruction, applying concordancers in language education, and preparing DDL-based materials in language classes. Textbook developers are also recommended specifying certain sections of class activities to DDL-based materials. [ABSTRACT FROM AUTHOR]

The purpose of the current study was twofold: examining the efficacy of data-driven learning (DDL) (hands-on and hands-off approaches) in the realization of interactional metadiscourse markers (IMMs) among English as a foreign language (EFL) learners and analyzing the learners' perceptions of DDL. The participants consisted of 93 male and female advanced language learners randomly assigned to one of the three groups: hands-on, hands-off, and control. Throughout the duration of treatment lasting for 10 sessions, the hands-on group employed the use of Microsoft Copilot, artificial intelligence (AI) chatbot, on a computer screen to discuss and explore IMMs, but the hands-off group was exposed to IMMs through written texts that were physically printed on paper and articles to be examined through AntConc concordancing program. The control group received conventional instructional techniques including reading assigned course materials. The findings from a one-way analysis of covariance (ANCOVA) procedure indicated that both experimental groups outperformed the control group in the posttest of realizing and identifying IMMs. However, the post hoc comparisons showed statistically significant differences between the hands-on and hands-off groups, with the hands-on group performing more successfully in identifying IMMs. The results of the questionnaire data revealed that all the learners had positive perception of DDL. The results of the current study suggest using both hands-on and hands-off DDL methods helps learners develop their writing performance through metadiscourse realization. [ABSTRACT FROM AUTHOR]

Introduction: In multiplexed positron emission tomography (mPET) imaging, physiological and pathological information from different radiotracers can be observed simultaneously in a single dynamic PET scan. The separation of mPET signals within a single PET scan is challenging due to the fact that the PET scanner measures the sum of the PET signals of all the tracers. The conventional multitracer compartment modeling (MTCM) method requires staggered injections and assumes that the arterial input functions (AIFs) of each tracer are known. Methods: In this work, we propose a deep learning-based method to separate triple-tracer PET images without explicitly knowing the AIFs. A dynamic tripletracer noisy MLEM reconstruction was used as the network input, and dynamic single-tracer noisy MLEM reconstructions were used as training labels. Results: A simulation study was performed to evaluate the performance of the proposed framework on triple-tracer ([18F]FDG+82Rb+[94mTc]sestamibi) PET myocardial imaging. The results show that the proposed methodology substantially reduced the noise level compared to the results obtained from single-tracer imaging. Additionally, it achieved lower bias and standard deviation in the separated single-tracer images compared to the MTCM-based method at both the voxel and region of interest (ROI) levels. Discussion: As compared to MTCM separation, the proposed method uses spatiotemporal information for separation, which improves the separation performance at both the voxel and ROI levels. The simulation study also demonstrates the feasibility and potential of the proposed DL-based method for the application to pre-clinical and clinical studies. [ABSTRACT FROM AUTHOR]

Thesis writing (used here as an umbrella term to cover both master's and doctoral postgraduate-level writing) is a high-stakes genre for postgraduate students. This important student genre has been well-researched from a corpus-based perspective. Corpora of theses and also research articles have been used for data-driven learning (DDL) of this key genre. The purpose of this article is to critically examine key DDL initiatives, some of which take a 'research into practice' orientation. Importantly, the discussion is framed around the notion of 'specificity' in the context of needs analysis, and whether the initiatives take a wide-angle, narrow-angle or move from a wide-angle to a narrow angle approach. Accounts which focus on DIY (do-it-yourself) mini-corpus compilation and use by students are also reviewed. The final section of the article presents a critique of current pedagogic applications, taking a closer look at the issue of 'specificity' within the wider context of needs analysis and mapping out areas for future consideration. It is suggested that an ethnographic perspective may be particularly useful for conceptualising specificity relating to students' present situation needs. The article also considers the impact of AI/ChatGPT on future corpus-based pedagogy of thesis writing. • Disciplinary corpora can usefully be used for teaching thesis writing. • Thesis writing courses using corpora can be viewed along a continuum of a wide-angle and a narrow-angle approach. • Do-it-yourself corpora are an ideal means of achieving 'specificity' in thesis writing courses. • Ethnographic methods can also shed light on the issue of 'specificity' in thesis writing pedagogy. [ABSTRACT FROM AUTHOR]

Most corpus tools commonly used for corpus-based data-driven learning (DDL) are designed for research rather than teaching purposes, with much DDL research suggesting learners and their teachers often stop DDL after initial training due to tool-related issues like complex user interfaces and system settings. Based on feedback from secondary-age language learners and their teachers in the Australian context, we present CorpusMate (https://corpusmate.com), a new, user-friendly corpus tool that incorporates several publicly available written and spoken corpora across 20 disciplinary subjects. It offers a range of flexible concordancing, n-gram and data visualisation options to ensure a fast, smooth and simple DDL experience for end users. [ABSTRACT FROM AUTHOR]

Purpose: This study attempts to evaluate the effect of the corpus-based inductive teaching approach with multiple academic corpora (PICA, CAEC and Oxford Corpus of Academic English) and conventional deductive teaching approach (i.e., multiple-choice items, filling the gap, matching and underlining) on learning academic collocations by Iranian advanced EFL learners (students learning English as a foreign language). Design/methodology/approach: This is a quasi-experimental, quantitative and qualitative study. Findings: The result showed the experimental group outperformed significantly compared with the control group. The experimental group also shared their perception of the advantages and disadvantages of the corpus-assisted language teaching approach. Originality/value: Despite growing progress in language pedagogy, methodologies and language curriculum design, there are still many teachers who experience poor performance in their students' vocabulary, whether in comprehension or production. In Iran, for example, even though mandatory English education begins at the age of 13, which is junior and senior high school, students still have serious problems in language production and comprehension when they reach university levels. [ABSTRACT FROM AUTHOR]

This article discusses the affordances and limitations of an automatic text evaluator in the context of the online teaching/learning of composition writing skills within a specialized linguistic domain, namely, English for Tourism. The system, named G-Rubric, was designed and built by an interdisciplinary team of linguists, psychologists, educationalists, and computer engineers to explore the applicability of data-driven language learning in education, for which it subsequently obtained several awards and distinctions. This article describes the adaptation process of G-Rubric to English for Tourism, contextualized in a distance learning university degree, and analyses its potential to substitute or complement frontline tutors in the task of revising and assessing student compositions. Two types of textual evaluation are provided by G-Rubric: numerical grading and focused feedback on form (writing) and function (content). Content evaluation is based on pattern-matching and machine reasoning against a specialized corpus and associated knowledge previously inserted in the tool as appropriate. The paper compares the performance of both tutors and system and proposes specific lines of research to gain insights into their optimal integration. [ABSTRACT FROM AUTHOR]

Text corpora provide authentic material for language instruction and an insight into the development of learner language use. This pedagogical potential can be enhanced by accompanying user-friendly text analysis tools designed for researchers, teachers, and learners alike. We introduce the Estonian Language Learning and Analysis Environment (ELLE) that combines a growing corpus of Estonian learner writings (Estonian Interlanguage Corpus - EIC) with various applications for linguistic analysis and automated text evaluation. The toolkit can be employed to analyse EIC and other corpora as well as study materials or users' own texts. ELLE's ongoing implementation follows a prototype which was created using interaction and participatory design methods, involving members of different target groups. The paper outlines the system architecture and presents the functionalities of each tool, highlighting their unique features compared to alternative web applications. [ABSTRACT FROM AUTHOR]

Data-driven learning (DDL) form-focused tasks are a relatively new concept. These tasks involve using concordance lines to teach language in a way that integrates discovery learning, authentic language use, consciousness-raising, and the communicative use of language. Given their novelty, there haven't been many studies on how they impact learners' engagement. Therefore, this study sought to study whether DDL form-focused tasks influence English as a foreign language (EFL) learners' task engagement. A total of 114 Iranian EFL learners were randomly divided between comparison and intervention groups in a study that utilized an experimental (comparison group, pretest, and post-test) design within a sequential explanatory mixed-methods design. The comparison group completed 10 non-DDL form-focused tasks, whereas the intervention group completed 10 DDL form-focused tasks. The results of t -tests and repeated-measures ANOVA indicated that incorporating DDL form-focused tasks into English classes enhanced EFL learners' task engagement in the short run. However, the impact of DDL form-focused tasks on EFL learners' task engagement was not durable. Moreover, analyzing semi-structured interview data suggested that using DDL-enhanced tasks with a form-focused approach increases EFL learners' task engagement by triggering their curiosity, improving their autonomy, enhancing their concentration and interest, and facilitating their discovery learning. The present study lends more credence to the application of such tasks. The paper ends with implications for English language teaching and materials development. [ABSTRACT FROM AUTHOR]

Artificial Intelligence has evolved in the educational field as a key tool for personalizing assessment and providing real-time feedback, allowing for more precise adaptation to individual student needs and thereby improving learning outcomes in the classroom. The aim of this study was to interpret, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 methodology, the role of Artificial Intelligence in data-driven real-time assessment and feedback in the classroom. The PRISMA 2020 methodology was employed to conduct a systematic literature review on this topic, selecting and analyzing scientific articles published between 2020 and 2024 that address the implementation and effectiveness of these technologies in the classroom. The included studies were critically evaluated using the Critical Appraisal Skills Programme tool to ensure the quality and relevance of the findings. This study has demonstrated that the implementation of Artificial Intelligence in data-driven real-time assessment and feedback can transform education by enhancing the precision and personalization of evaluations, promoting critical and creative thinking in students. It was found that Artificial Intelligence facilitates more adaptive learning, particularly in vocational contexts. However, challenges in its implementation were highlighted, emphasizing the need for more rigorous studies that ensure the equity and effectiveness of these technologies in educational environments. Future research should focus on overcoming these limitations and ensuring ethical integration. [ABSTRACT FROM AUTHOR]

Technological advancements have not only introduced new tools for language education but also tailored apps and resources for specific language skills. These include mobile apps (e.g., Drops: Language Learning Games for vocabulary), websites (e.g., SpanishDict for Spanish dictionary and grammar), and reference tools (e.g., Grammarly for writing assistance). One such resource is PlayPhrase.me, an online and mobile tool serving as a database of movie clips for language practice, particularly vocabulary. Despite the relative attention similar tools (e.g., YouGlish and Voscreen) have received, only two studies were conducted about the PlayPhrase.me a website. Accordingly, there is a need for a review that outlines its general characteristics and pedagogical affordances. To this end, this review evaluated the website and revealed considerable potential for vocabulary, grammar, and pronunciation learning. Citing the possible pitfalls, such as short video durations, lack of filtering, and absence of evaluative and competitive means, the researcher offered recommendations for improvement. The review concludes with the researcher’s reflections as a teacher and urges further research to validate the findings.

Corpus linguistics can be seen as a set of methods applicable to different branches of linguistics. This article deals with the application of corpus linguistics (including learner corpus research) to second language acquisition (SLA) research and foreign language teaching (FLT). It describes the existing links between these frameworks, arguing that the combinations usually work in pairs. By contrast, the concept of corpus-based theory-to-practice paths is introduced as a way of combining the three frameworks. More precisely, SLA research and FLT are brought together through the prism of corpora, which can provide data to test and possibly refine SLA hypotheses and then form the basis for pedagogical materials that can help solve some of the issues brought to light by the corpus analysis. Illustrations of such paths are given that centre around language input, the natural order hypothesis and language transfer.

As a corpus-assisted method for language pedagogy, DDL (data-driven learning) may have the potential to enhance language exposure and promote active learner engagement. Concordancing, or KWIC (Key Words in Context), has been the traditional method used in DDL to retrieve numerous language examples, while the method has limitations with overreliance on individual words to search. This paper aims to propose and promote concordancing alternatively based on grammar patterns, a multi-word concept in corpus linguistics. The conceptualised method is named as PIC (Patterns in Context), an extended form of KWIC. An empirical study was conducted to investigate whether the PIC method has any advantages over the traditional KWIC method, using two custom-built Android apps. The research involved 18 pre-university intermediate learners (and six pilot study participants), who used the apps in a self-directed way for two weeks. Then the assessment of the two apps and methods was conducted based on data from automatic logs and responses from questionnaires and interviews. The results suggest that, compared to KWIC, PIC could be slightly advantageous in efficiently helping learners find the target language use, while this approach seems not strong in user engagement and perceived effectiveness. The implications for DDL are discussed, and further investigation is also planned. [ABSTRACT FROM AUTHOR]

Corpus-based approaches to language and literacy education, commonly known as data-driven learning (DDL), are increasing in prominence. However, the vast majority of DDL interventions involve adult tertiary level learners, leaving a dire need for comprehensive DDL studies for secondary education. The present study reports on a half-year DDL experiment conducted at an all-girls secondary school in Australia, focusing on the development of receptive and productive knowledge of passive voice constructions used when writing scientific research reports for a physical science class. Pre/post-tests were conducted testing learners' receptive knowledge and productive use of the passive, alongside data on learners' autonomous use of corpora within a written research report. Learners' perceptions of corpora and DDL were also collected through questionnaire survey and interview data taken both immediately post-training and three months after training. The results suggest learners' corpus consultation was effective in improving use of the passive voice for science writing with pre-tertiary learners, although clear preferences for (and criticisms of) certain corpus tools, functions and usage was apparent, and continued uptake post-training was relatively weak. Generally however, the implications of these findings paint a positive picture of what is possible regarding DDL with younger learners, and provide a model of how a DDL intervention with younger learners can be successfully managed and integrated in a context where secondary content teachers, rather than solely the applied linguist, can be the main stakeholders in a DDL intervention. [ABSTRACT FROM AUTHOR]

A data‐driven model for predicting the surface temperature using neural networks was proposed to alleviate the computational burden of numerical weather prediction (NWP). Our model, named TPTNet uses only 2 m temperature measured at the weather stations of the South Korean Peninsula as input to predict the local temperature at finite forecast hours. The turbulent fluctuation component of the temperature was extracted from the station measurements by separating the climatology component accounting for the yearly and daily variations. The effect of station altitude was then compensated by introducing a potential temperature. The resulting turbulent potential temperature (TPT) data at irregularly distributed stations were used as input for predicting the TPT at forecast hours through three trained networks based on convolutional neural network, Swin Transformer, and a graph neural network. By comparing the prediction performance of our network with that of persistence and NWP, we found that our model can make predictions comparable to NWP for up to 12 hr. Plain Language Summary: We developed a new model called TPTNet to predict local temperatures using only temperature measurements from weather stations in South Korea. This model aims to reduce the heavy computational demands of traditional numerical weather prediction (NWP) methods. By analyzing 20 years of data, we separated the regular temperature patterns that change yearly and daily from the more unpredictable fluctuations. We also adjusted the model for the altitude of each weather station. We trained three different types of neural networks for this purpose: a convolutional neural network, a Swin Transformer, and a graph neural network. When we tested our model with new data from 2020, it made reliable temperature predictions for up to 12 hr, performing as well as, or even better than, traditional NWP methods. This new approach could help improve weather forecasting by making it faster and less computationally intensive. Key Points: Only 2 m temperature data measured at the weather station is used as input for the prediction of temperature at finite forecast hoursTurbulent fluctuation temperature relative to the climatological yearly and daily periodic variations with the altitude adjustment using potential temperature is considered in data‐driven learningFor 12 hr of forecast hour, TPTNet produces forecasts comparable to those of the numerical weather prediction, with the less scattered errors over stations [ABSTRACT FROM AUTHOR]

The prognosis of organic light‐emitting diodes (OLEDs) not only requires early detection of a bearing defect, but also the capability to predict their life data under all operational scenarios. The use of sophisticated machine learning (ML) algorithms is undoubtedly becoming an increasingly exciting research direction, as these algorithms can yield high predictive models with minimal domain expertise. The central question of this perspective is: how well can ML models advance our ability to forecast the lifetime of OLEDs compared to the physics‐based models? In this paper, data‐driven methods, feed‐forward neural networks (FFNN), support vector machines (SVMs), k‐nearest neighbors (KNNs), partial least squares regression (PLSR), and decision trees (DTs), are used to predict the lifetime and reliability of OLEDs through analyzing the lumen degradation data collected from the accelerated lifetime test. The final predicted results indicate that both the data‐driven and our physics‐based OLED lifetime models fit well the experimental data. The main drawback of the former method is that their efficacy is highly contingent on the quantity and quality of the operational dataset. Among all these methods, much more reliability information (time to failure) and the highest prediction accuracy can be achieved by FFNN. [ABSTRACT FROM AUTHOR]

This study investigates the impact of Data-Driven Learning (DDL) on the motivation of vocational high school students in Indonesia, particularly those studying automotive engineering. Traditional grammar-translation methods used in vocational schools often fail to engage students because they do not connect language learning with practical, real-world applications. This gap is significant in vocational education, where students view English as a tool for their careers rather than an academic subject. DDL, which engages learners with authentic language data from their vocational fields, offers a promising alternative by making language learning more relevant and interactive. Using a quasi-experimental, mixed-methods design, the study involved 60 students divided into an experimental group (DDL-based instruction) and a control group (traditional instruction). Motivation was measured using Dornyei's L2 Motivation Self System, focusing on Ideal L2 Self, Ought-to L2 Self, and L2 Learning Experience. Results showed that students in the experimental group experienced significant gains in all three motivational components, particularly in Ideal L2 Self and L2 Learning Experience, highlighting how DDL made English more relevant to their career goals. In contrast, the control group saw only minor improvements in motivation. Qualitative findings from open-ended responses and classroom observations emphasized the role of learner autonomy and the relevance of materials in enhancing motivation. The study concludes that DDL can effectively address the gap in vocational education by making English learning more meaningful, though it requires additional support to help students manage complex language data. [ABSTRACT FROM AUTHOR]

This paper uses corpus data to analyse spoken expressions and discourse markers in Czech, applying these findings to corpus-based exercises for learners of Czech as a foreign language. The analytical section highlights the usefulness of parallel corpus in identifying suitable translation equivalents for prevalent Czech spoken vocabulary in English, French, and Polish as native languages from the learner's perspective. The methodology outlines the process of finding appropriate translation equivalents in film subtitles, considering both meaning and spoken register. The pedagogical section introduces three corpus-based exercises designed to improve conversational skills, featuring authentic texts that familiarise learners with spoken vocabulary. This research builds on previous studies of the English language that did not use parallel corpora to identify translation equivalents in learners' native languages -- an essential factor for understanding a foreign language. In addition, tailor-made corpus-based exercises can be seamlessly integrated into everyday classroom activities to enhance language awareness among non-native speakers. [ABSTRACT FROM AUTHOR]

This contribution introduces the Swiss learner corpus SWIKO and provides examples on how this rich and near-authentic collection can be utilized in foreign language education, while also addressing some critical issues that corpus linguistic applications face in pedagogical contexts. SWIKO is a multilingual corpus currently being developed at the Institute of Multilingualism in Fribourg. The corpus contains written and spoken productions by Swiss lower secondary school students, both in their language of schooling and foreign languages learnt at school (English, French, and German). Participating students completed eight communicative tasks which systematically vary by rhetorical type (descriptive or argumentative), topic (personal or academic), and structure (more or less restrictive input). The resulting productions were analysed with a focus on linguistic features (e.g., Karges et al., 2022) and in relation to human ratings according to the levels of the CEFR (e.g., Studer & Hicks, 2022). Based on our findings, we present two scenarios on how SWIKO can be used in educational settings, i.e., teacher training and material development. First, the productions can serve as an illustration of learners’ abilities at the end of mandatory schooling. Our findings show how the length, complexity, and accuracy of learner texts heavily depend on the task, which can be addressed in teacher training. Second, an analysis of frequent errors in foreign language productions can shed light on challenging structures, while the language of schooling sub-corpus can serve as a peer-reference in the development of corresponding material. We exemplify this process focusing on negation in German, offering differentiated teaching material suitable for the secondary school classroom.

A growing body of research has examined the integration of corpora into foreign language classrooms in the last few decades. A sub-line of research concerning the interaction between the two fields has been student perceptions. Additionally, the growing integration of online learning into curricula, triggered by the COVID-19 pandemic, has aroused a necessity to explore the application of data-driven learning in online learning environments. The current study, therefore, attempted to elicit learner perceptions on learning vocabulary targeted particularly for speaking skill through direct data-driven learning in an online learning context. The data gathered from the participants (N=28) through a questionnaire and a semistructured interview revealed that the participants found the use of data-driven learning activities in conjunction with the coursebook and oral production tasks helpful to boost their learning. They did not evaluate corpus consultation as a very challenging task to tackle in general. Overall, the study revealed two significant findings: First, it highlighted that collaborative learning might play a significant role in data-driven learning applications. Second, the findings emphasized the necessity for integration of the audio data to the large, publicly available corpora, such as COCA, which could contribute to future research on exploring the efficiency of data-driven learning for speaking skills. [ABSTRACT FROM AUTHOR]

As a developing country with an agricultural economy as a pillar, Bangladesh is highly vulnerable to adverse effects of climate change, so the generation of high-resolution temperature maps is of great value for Bangladesh to achieve agricultural sustainable development. However, Bangladesh's weak economy and sparse meteorological stations make it difficult to obtain such maps. In this study, by mining internal features and links inside observed data, we developed an efficient data-driven downscaling technique to generate high spatial-resolution temperature distribution maps of Bangladesh directly from observed temperature data at 34 meteorological stations with irregular distribution. Based on these high-resolution historical temperature maps, we further explored a data-driven forecast technique to generate high-resolution temperature maps of Bangladesh for the period 2025–2035. Since the proposed techniques are very low-cost and fully mine internal links inside irregular-distributed observations, they can support relevant departments of Bangladesh to formulate policies to mitigate and adapt to climate change in a timely manner. [ABSTRACT FROM AUTHOR]

Introduction: Data-driven learning has emerged as a powerful method in language teaching and has drastically changed the traditional methods of teaching and learning, especially in academic writing. So, the present study was conducted to investigate the effect of data-driven learning on the English academic writing of medical students. Method: In this quasi-experimental intervention study, 29 MSc students of different fields of medical sciences participated in a training intervention program using a data-driven learning approach focusing on self-generated corpora of academic articles related to their fields of study. Students were asked to write an academic text related to their fields at the beginning and end of the course. Written texts were graded based on the Hemp-Lyons academic writing rubric, and the data collected from the pre-and post-test assessments were analyzed using the paired-sample t-test. Results: Comparison of pre-test scores with a mean of 59 (SD = 2.26) and post-test scores with a mean of 78.53 (SD = 2.33) indicated a significant improvement in the writing of the participants of this study (p < 0.05). Conclusion: In light of confirming the educational suitability of the data-driven approach and recognizing the need for students to engage in English article writing and publication, these findings serve as a valuable aid to MSc students studying medical sciences, facilitating their academic writing in English.

In this paper, an automated and real-time anomaly detection approach for sensors of autonomous vehicles called ARAD is presented. Automated vehicles gather environmental information through their diverse built-in sensors thus the correctness of this data affects the system’s reliability, directly. Accordingly, anomaly detection schemes are employed to guarantee the correctness of the sensors’ data. Moreover, due to the necessity of real-time operation in automated vehicles, the response time of the anomaly detection unit is important along with its precision. To this aim, in our proposed ARAD a lightweight and hierarchical architecture to detect and classify the anomalies based on their types is employed. Moreover, to enhance the detection capability, ARAD utilizes the data diversity property based on the sequence prediction scheme. After anomaly detection, ARAD mitigates and removes them from the system’s input by its rule-based engine. To meet the precision and real-time requirements of the anomaly detection unit in autonomous vehicles, ARAD has a lightweight sequence prediction structure based on statistical and data-driven methods. To evaluate the effectiveness of our proposed ARAD, several experiments are performed and a performance measurement metric called FoC is proposed to study the contradicted effects of precision and real-time operation in terms of computation overhead, simultaneously. Based on these experiments, ARAD is capable of detecting anomalies efficiently with precision and recall of $84.6~\%$ and 87%, respectively in real-time while applying low overhead to the system. It also shows 75.6% improvement in terms of computation cost over related methods.

Despite considerable discussion in the literature (Flowerdew & Peacock, 2001; Hyland, 1998; Tang, 2012) competent English academic writing is still a problem which needs to be solved. English for Academic Purposes (EAP) teaching often focuses on specialized lexis, which may, however, be the area where academic writers need least help. The study of a small corpus of C2 level academic writing which consisted of the sub-genres of summary and discussion writing revealed that one key area which is problematic is collocation. This paper presents the results of this small corpus investigation into learner language and how it informed the classroom implementation of data-driven learning (DDL) to increase learner awareness of and ability to use collocations effectively in written academic English. The article briefly describes the corpus and the resulting teaching procedure adopted. The first step of this procedure is familiarization followed by experimentation using Sketch Engine (SkeLL).

Training-anomaly-based, machine-learning-based, intrusion detection systems (AMiDS) for use in critical Internet of Things (CioT) systems and military Internet of Things (MioT) environments may involve synthetic data or publicly simulated data due to data restrictions, data scarcity, or both. However, synthetic data can be unrealistic and potentially biased, and simulated data are invariably static, unrealistic, and prone to obsolescence. Building an AMiDS logical model to predict the deviation from normal behavior in MioT and CioT devices operating at the sensing or perception layer due to adversarial attacks often requires the model to be trained using current and realistic data. Unfortunately, while real-time data are realistic and relevant, they are largely imbalanced. Imbalanced data have a skewed class distribution and low-similarity index, thus hindering the model’s ability to recognize important features in the dataset and make accurate predictions. Data-driven learning using data sampling, resampling, and generative methods can lessen the adverse impact of a data imbalance on the AMiDS model’s performance and prediction accuracy. Generative methods enable passive adversarial learning. This paper investigates several data sampling, resampling, and generative methods. It examines their impacts on the performance and prediction accuracy of AMiDS models trained using imbalanced data drawn from the UNSW_2018_IoT_Botnet dataset, a publicly available IoT dataset from the IEEEDataPort. Furthermore, it evaluates the performance and predictability of these models when trained using data transformation methods, such as normalization and one-hot encoding, to cover a skewed distribution, data sampling and resampling methods to address data imbalances, and generative methods to train the models to increase the model’s robustness to recognize new but similar attacks. In this initial study, we focus on CioT systems and train PCA-based and oSVM-based AMiDS models constructed using low-complexity PCA and one-class SVM (oSVM) ML algorithms to fit an imbalanced ground truth IoT dataset. Overall, we consider the rare event prediction case where the minority class distribution is disproportionately low compared to the majority class distribution. We plan to use transfer learning in future studies to generalize our initial findings to the MioT environment. We focus on CioT systems and MioT environments instead of traditional or non-critical IoT environments due to the stringent low energy, the minimal response time constraints, and the variety of low-power, situational-aware (or both) things operating at the sensing or perception layer in a highly complex and open environment.

IntroductionIn multiplexed positron emission tomography (mPET) imaging, physiological and pathological information from different radiotracers can be observed simultaneously in a single dynamic PET scan. The separation of mPET signals within a single PET scan is challenging due to the fact that the PET scanner measures the sum of the PET signals of all the tracers. The conventional multi-tracer compartment modeling (MTCM) method requires staggered injections and assumes that the arterial input functions (AIFs) of each tracer are known.MethodsIn this work, we propose a deep learning-based method to separate triple-tracer PET images without explicitly knowing the AIFs. A dynamic triple-tracer noisy MLEM reconstruction was used as the network input, and dynamic single-tracer noisy MLEM reconstructions were used as training labels.ResultsA simulation study was performed to evaluate the performance of the proposed framework on triple-tracer ([F18]FDG+Rb82+[Tc99m]sestamibi) PET myocardial imaging. The results show that the proposed methodology substantially reduced the noise level compared to the results obtained from single-tracer imaging. Additionally, it achieved lower bias and standard deviation in the separated single-tracer images compared to the MTCM-based method at both the voxel and region of interest (ROI) levels.DiscussionAs compared to MTCM separation, the proposed method uses spatiotemporal information for separation, which improves the separation performance at both the voxel and ROI levels. The simulation study also demonstrates the feasibility and potential of the proposed DL-based method for the application to pre-clinical and clinical studies.

Introduction: Data -driven learning has emerged as a powerful method in language teaching and has drastically changed the traditional methods of teaching and learning, especially in academic writing. So, the present study was conducted to investigate the effect of data -driven learning on the English academic writing of medical students. Method: In this quasi -experimental intervention study, 29 MSc students of different fields of medical sciences participated in a training intervention program using a data -driven learning approach focusing on self-generated corpora of academic articles related to their fields of study. Students were asked to write an academic text related to their fields at the beginning and end of the course. Written texts were graded based on the Hemp -Lyons academic writing rubric, and the data collected from the pre - and post -test assessments were analyzed using the paired -sample t -test. Results: Comparison of pre -test scores with a mean of 59 (SD = 2.26) and post -test scores with a mean of 78.53 (SD = 2.33) indicated a significant improvement in the writing of the participants of this study (p < 0.05). Conclusion: In light of confirming the educational suitability of the data -driven approach and recognizing the need for students to engage in English article writing and publication, these findings serve as a valuable aid to MSc students studying medical sciences, facilitating their academic writing in English. [ABSTRACT FROM AUTHOR]

In this position paper, I argue that proficiency-rated learner corpora should gain a more prominent role in data-driven learning (DDL). With specific reference to DDL, proficiency-rated learner corpora can provide typical, atypical and erroneous target language data at different levels of proficiency, which can be meaningfully used in the design of learning activities. This makes them pivotal in expanding the scope of DDL to include mid- and lower-level proficiency learners more extensively. Although the field of learner corpus research has been promoting learner corpus use in DDL for a long time, only a small fraction of DDL studies make use of a learner corpus. As a contribution to overcome this hiatus, I will demonstrate how using a specific proficiency-rated learner corpus (i.e., the CELI corpus; Spina et al., 2022, 2023) can enrich the design of DDL activities, making them more adaptable to a wider range of learner needs. [ABSTRACT FROM AUTHOR]

This study investigates the differential effect of various noticing activities on grammatical accuracy development in EFL learners' written productions. We focus on different types of noticing activities based on an error-tagged learner corpus and report on effective practical experiments involving learner corpus data. A pretest/posttest quasi-experimental design is used with three experimental groups (receiving different treatments) and one control group. Error frequencies, at both group and individual levels, and proportions of learners producing errors on three specific error types (articles, verb tense, verb agreement) are compared. Our results suggest that accuracy in the use of articles and verb agreement could be more easily fostered through the comparison of learner output with native data (the BNC, in our case). As for verb tenses, the impact of a more traditional form of corrective feedback seems greater while the use of online machine translation tools does not seem to foster much accuracy development. [ABSTRACT FROM AUTHOR]

This article explores two under-researched types of corpora for use in data-driven learning (DDL): L2 corpora (i.e. in a second or foreign language) and multimodal corpora. It first outlines the development of FLEURON, a dedicated DDL platform designed to support interactional competence in French as a Foreign Language (FFL), based on multimodal corpora of both native and L2 speakers. It then presents an ecological study of how 19 international FFL learners interacted with the platform in a DDL approach at the University of Lorraine. The analysis highlights how L2 corpora in particular can help learners to improve their awareness of complex phenomena related to conversational narratives by engaging their meta-cognitive strategies during their time abroad. The study thus reveals the potential for integrating an L2 component among the range of resources available for teaching and learning spoken interaction. [ABSTRACT FROM AUTHOR]

Copyright of Texto Livre / Texto Livre: Linguagem e Tecnologia is the property of Revista Texto Livre: Linguagem e Tecnologia and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Training-anomaly-based, machine-learning-based, intrusion detection systems (AMiDS) for use in critical Internet of Things (CioT) systems and military Internet of Things (MioT) environments may involve synthetic data or publicly simulated data due to data restrictions, data scarcity, or both. However, synthetic data can be unrealistic and potentially biased, and simulated data are invariably static, unrealistic, and prone to obsolescence. Building an AMiDS logical model to predict the deviation from normal behavior in MioT and CioT devices operating at the sensing or perception layer due to adversarial attacks often requires the model to be trained using current and realistic data. Unfortunately, while real-time data are realistic and relevant, they are largely imbalanced. Imbalanced data have a skewed class distribution and low-similarity index, thus hindering the model's ability to recognize important features in the dataset and make accurate predictions. Data-driven learning using data sampling, resampling, and generative methods can lessen the adverse impact of a data imbalance on the AMiDS model's performance and prediction accuracy. Generative methods enable passive adversarial learning. This paper investigates several data sampling, resampling, and generative methods. It examines their impacts on the performance and prediction accuracy of AMiDS models trained using imbalanced data drawn from the UNSW_2018_IoT_Botnet dataset, a publicly available IoT dataset from the IEEEDataPort. Furthermore, it evaluates the performance and predictability of these models when trained using data transformation methods, such as normalization and one-hot encoding, to cover a skewed distribution, data sampling and resampling methods to address data imbalances, and generative methods to train the models to increase the model's robustness to recognize new but similar attacks. In this initial study, we focus on CioT systems and train PCA-based and oSVM-based AMiDS models constructed using low-complexity PCA and one-class SVM (oSVM) ML algorithms to fit an imbalanced ground truth IoT dataset. Overall, we consider the rare event prediction case where the minority class distribution is disproportionately low compared to the majority class distribution. We plan to use transfer learning in future studies to generalize our initial findings to the MioT environment. We focus on CioT systems and MioT environments instead of traditional or non-critical IoT environments due to the stringent low energy, the minimal response time constraints, and the variety of low-power, situational-aware (or both) things operating at the sensing or perception layer in a highly complex and open environment. [ABSTRACT FROM AUTHOR]