Word Level Research Articles

Guessing errors made by children with dyslexia in word and text reading.

Children with dyslexia face persistent difficulties in reading acquisition, which results in poor reading accuracy. In addition to the commonly studied reading errors such as omissions, additions, substitutions, and letter reversals, they also make guessing errors characterized by replacing a word with an orthographic neighbor. These errors, which occur in the context of isolated words and sentence or text reading, might be linked to the inhibition issues that have already been demonstrated in connection to dyslexia. However, to the best of our knowledge, there is no experimental evidence that children with dyslexia make more guessing errors than typically developing children, as is the case for sequential errors. The current study aimed to evaluate whether children with dyslexia made more guessing errors than typically developing children and whether these errors were more frequent in an isolated word or sentence context. Twenty-eight children with dyslexia from Grade 4 were matched with typically developing children by either chronological age or reading level. Reading was assessed through word and text reading tasks. Error types were classified into seven categories and analyzed. A repeated-measure ANOVA showed that children with dyslexia made more guessing errors, particularly visual and morphemic errors, than typically developing children. Moreover, these errors were found at both the single word and sentence levels. These findings suggest that children with dyslexia tend to use a global treatment of words, relying on incomplete information to compensate for their difficulties. The findings have practical implications for pedagogical and therapeutic approaches.

Explainable paper classification system using topic modeling and SHAP

The exponential growth of academic papers necessitates sophisticated classification systems to effectively manage and navigate vast information repositories. Despite the proliferation of such systems, traditional approaches often rely on embeddings that do not allow for easy interpretation of classification decisions, creating a gap in transparency and understanding. To address these challenges, we propose an innovative explainable paper classification system that combines Latent Semantic Analysis (LSA) for topic modeling with explainable artificial intelligence (XAI) techniques. Our objective is to identify which topics significantly influence the classification outcomes, incorporating Shapley additive explanations (SHAP) as a key XAI technique. Our system extracts topic assignments and word assignments from paper abstracts using latent semantic analysis (LSA) topic modeling. Topic assignments are then employed as embeddings in a multilayer perceptron (MLP) classification model, with the word assignments further utilized alongside SHAP for interpreting the classification results at the corpus, document, and word levels, enhancing interpretability and providing a clear rationale for each classification decision. We applied our model to a dataset from the Web of Science, specifically focusing on the field of nanomaterials. Our model demonstrates superior classification performance compared to several baseline models. Ultimately, our proposed model offers a significant advancement in both the performance and explainability of the system, validated by case studies that illustrate its effectiveness in real-world applications.

Textual Cohesion in Surah Al- Kahf through the Element of Repetition.

The study of textual cohesion is a recent trend in the linguistic lesson. Recurrence is one of the topics of textual cohesion at the lexical level through which the existence of textual cohesion can be determined through the effect that repetition has on textual cohesion. Studies have been striving to investigate the effect of repetition on textual cohesion through text grammar, so the importance of this study lies in the fact that it made a contribution in this field. This study examines the textual cohesion in Surah Al-Kahf through the element of repetition, as it is one of the prominent aspects of textual cohesion. The study relied on choosing Surah Al-Kahf specifically due to the abundance and variety of repetition found within this surah, encompassing repetition at the level of letters, words, sentences, and entire Quranic verses.

Incorporating Frequency Effects in the Lexical Access of Mandarin Tone 3 Sandhi.

Mandarin tone 3 sandhi refers to the phenomenon whereby a tone 3 syllable changes to a tone 2 when followed by another tone 3. This phonological process creates a deviation between the tonal forms realized at morphemic (/tone3-tone3/) and word ([tone2-tone3]) levels, posing questions in terms of how disyllabic tone 3 sandhi words are represented and accessed. The current study conducted three cross-modal lexical decision priming experiments to investigate this issue. Experiment 1 manipulated the frequencies of the initial morpheme and whole word, showing that the higher initial-character frequency against the whole word gives stronger activation to the underlying representation and the lower frequency of the initial character leads to stronger activation of the surface tone. Experiments 2 and 3 operationalized the relative frequency of the initial tone 3 morpheme's realization as a sandhi tone, finding that the competition between the two tonal realizations also influences how T3 sandhi words are accessed. Specifically, the more frequently the T3 morpheme surfaces as a T2 allomorph, the less activated the underlying representation becomes in the mental lexicon. Our results indicate a complex interplay between morpheme, word, and the associated tonal representations in the mental lexicon and that these factors co-determine the lexical access of tone 3 sandhi.

Mono-lingual text reuse detection for the Urdu language at lexical level

Text reuse is the process of creating new texts from pre-existing ones. In recent years, Urdu Text Reuse Detection (U-TRD) has garnered the attention of researchers due to the ready availability of digital text all over the internet, which can be copied or paraphrased from other sources without proper attribution, making it easier to reuse but challenging to detect. Previous studies have explored the issue of U-TRD at the phrasal, sentence/passage, and document levels, using benchmark corpora and methods. However, the problem of U-TRD has not been investigated at the lexical level in terms of corpora and methods. To address this research gap, our study has developed a large benchmark corpus manually annotated at the lexical level. This corpus consists of 22,184 text pairs categorized into two levels of rewrite: (1) Derived (8,660) and (2) Non-Derived (13,524). Additionally, our research has involved the development, application, evaluation, and comparison of a range of methods, including baseline methods (uni-gram overlap and word embedding-based methods), along with state-of-the-art transformer-based methods and feature-fusion-based methods, using the proposed UTRD-Lex-23 corpus. Our study concludes that one of our proposed feature-fusion methods outperforms all other methods. The model we propose, which combines seven different Sentence Transformers (ST) (each producing 768 dimension vectors) with one uni-gram (at word level) and sixteen different features extracted from four different Word Embedding (WE) based models (yielding 300 dimension vectors), achieves an F1 score of 0.70601 using 10-fold cross validation. To foster and promote research in Urdu (a low-resourced language) proposed corpus will be freely and publicly available for research purposes.

End-to-End Deep Learning Framework for Arabic Handwritten Legal Amount Recognition and Digital Courtesy Conversion

Arabic handwriting recognition and conversion are crucial for financial operations, particularly for processing handwritten amounts on cheques and financial documents. Compared to other languages, research in this area is relatively limited, especially concerning Arabic. This study introduces an innovative AI-driven method for simultaneously recognizing and converting Arabic handwritten legal amounts into numerical courtesy forms. The framework consists of four key stages. First, a new dataset of Arabic legal amounts in handwritten form (“.png” image format) is collected and labeled by natives. Second, a YOLO-based AI detector extracts individual legal amount words from the entire input sentence images. Third, a robust hybrid classification model is developed, sequentially combining ensemble Convolutional Neural Networks (CNNs) with a Vision Transformer (ViT) to improve the prediction accuracy of single Arabic words. Finally, a novel conversion algorithm transforms the predicted Arabic legal amounts into digital courtesy forms. The framework’s performance is fine-tuned and assessed using 5-fold cross-validation tests on the proposed novel dataset, achieving a word level detection accuracy of 98.6% and a recognition accuracy of 99.02% at the classification stage. The conversion process yields an overall accuracy of 90%, with an inference time of 4.5 s per sentence image. These results demonstrate promising potential for practical implementation in diverse Arabic financial systems.

Speechformer-CTC: Sequential modeling of depression detection with speech temporal classification

Speech-based automatic depression detection systems have been extensively explored over the past few years. Typically, each speaker is assigned a single label (Depressive or Non-depressive), and most approaches formulate depression detection as a speech classification task without explicitly considering the non-uniformly distributed depression pattern within segments, leading to low generalizability and robustness across different scenarios. However, depression corpora do not provide fine-grained labels (at the phoneme or word level) which makes the dynamic depression pattern in speech segments harder to track using conventional frameworks. To address this, we propose a novel framework, Speechformer-CTC, to model non-uniformly distributed depression characteristics within segments using a Connectionist Temporal Classification (CTC) objective function without the necessity of input–output alignment. Two novel CTC-label generation policies, namely the Expectation-One-Hot and the HuBERT policies, are proposed and incorporated in objectives on various granularities. Additionally, experiments using Automatic Speech Recognition (ASR) features are conducted to demonstrate the compatibility of the proposed method with content-based features. Our results show that the performance of depression detection, in terms of Macro F1-score, is improved on both DAIC-WOZ (English) and CONVERGE (Mandarin) datasets. On the DAIC-WOZ dataset, the system with HuBERT ASR features and a CTC objective optimized using HuBERT policy for label generation achieves 83.15% F1-score, which is close to state-of-the-art without the need for phoneme-level transcription or data augmentation. On the CONVERGE dataset, using Whisper features with the HuBERT policy improves the F1-score by 9.82% on CONVERGE1 (in-domain test set) and 18.47% on CONVERGE2 (out-of-domain test set). These findings show that depression detection can benefit from modeling non-uniformly distributed depression patterns and the proposed framework can be potentially used to determine significant depressive regions in speech utterances.

Clinical efficacy of pre-trained large language models through the lens of aphasia

The rapid development of large language models (LLMs) motivates us to explore how such state-of-the-art natural language processing systems can inform aphasia research. What kind of language indices can we derive from a pre-trained LLM? How do they differ from or relate to the existing language features in aphasia? To what extent can LLMs serve as an interpretable and effective diagnostic and measurement tool in a clinical context? To investigate these questions, we constructed predictive and correlational models, which utilize mean surprisals from LLMs as predictor variables. Using AphasiaBank archived data, we validated our models’ efficacy in aphasia diagnosis, measurement, and prediction. Our finding is that LLMs-surprisals can effectively detect the presence of aphasia and different natures of the disorder, LLMs in conjunction with the existing language indices improve models’ efficacy in subtyping aphasia, and LLMs-surprisals can capture common agrammatic deficits at both word and sentence level. Overall, LLMs have potential to advance automatic and precise aphasia prediction. A natural language processing pipeline can be greatly benefitted from integrating LLMs, enabling us to refine models of existing language disorders, such as aphasia.

Joint exome and metabolome analysis in individuals with dyslexia: Evidence for associated dysregulations of olfactory perception and autoimmune functions.

Dyslexia is a learning disability that negatively affects reading, writing, and spelling development at the word level in 5%-9% of children. The phenotype is variable and complex, involving several potential cognitive and physical concomitants such as sensory dysregulation and immunodeficiencies. The biological pathogenesis is not well-understood. Toward a better understanding of the biological drivers of dyslexia, we conducted the first joint exome and metabolome investigation in a pilot sample of 30 participants with dyslexia and 13 controls. In this analysis, eight metabolites of interest emerged (pyridoxine, kynurenic acid, citraconic acid, phosphocreatine, hippuric acid, xylitol, 2-deoxyuridine, and acetylcysteine). A metabolite-metabolite interaction analysis identified Krebs cycle intermediates that may be implicated in the development of dyslexia. Gene ontology analysis based on exome variants resulted in several pathways of interest, including the sensory perception of smell (olfactory) and immune system-related responses. In the joint exome and metabolite analysis, the olfactory transduction pathway emerged as the primary pathway of interest. Although the olfactory transduction and Krebs cycle pathways have not previously been described in dyslexia literature, these pathways have been implicated in other neurodevelopmental disorders including autism spectrum disorder and obsessive-compulsive disorder, suggesting the possibility of these pathways playing a role in dyslexia as well. Immune system response pathways, on the other hand, have been implicated in both dyslexia and other neurodevelopmental disorders.

التكرار الصوتي في شعر بدوي الجبل

The research seeks to study one of the important phenomena of the Arabic language, which is the phenomenon of vocal repetition in the poetry of Badawi al-Jabal. It is no secret that the voice has great importance in Arabic, as language is nothing but “sounds with which every people expresses their purposes,” and repetition is a distinctive stylistic phenomenon in poetry. Semantic value and condensing energies that complete the images and perform their suggestive function in a consistent manner with the general purpose intended by the speaker. It is also considered one of the most prominent stylistic phenomena for understanding the literary text. Its aesthetics are highlighted in the artistic text through alternating words and repeating them in the context of poetic expression, and forming a melody. Musically, it is intended by the composer in his poetry or prose, as repetition highlights something important to the poet that he reveals by repeating it, in addition to the fact that repetition is one of the tools that increase the beauty and beauty of the text. This study has clarified the meaning of phonetic repetition in language and terminology and the opinions of ancient and modern Arabic scholars regarding it, and its most important benefits in the poetic text - in the collection of the poet Badawi Al-Jabal - such as emphasis, increased alertness, and repetition of the related to negate heedlessness and others, in addition to mentioning the most important levels of repetition, which are repetition at the level of the letter and at the letter level. Word level, sentence level, and passage level. The researcher relied on a descriptive and analytical approach, and the study concluded with a conclusion that included the most important results reached by the researcher.

A New Method for Documenting Sign Language Productions in Schools

Purpose: There are well-established guidelines for the recording, transcription, and analysis of spontaneous oral language samples by researchers, educators, and speech pathologists. In contrast, there is presently no consensus regarding methods for the written documentation of sign language samples. The Handshape Analysis Recording Tool (HART) is an innovative method for documenting and analyzing word level samples of signed languages in real time. Fluent sign language users can document the expressive sign productions of children to gather data on sign use and accuracy. Method: The HART was developed to document children's productions in Australian Sign Language (Auslan) in a bilingual–bicultural educational program for the Deaf in Australia. This written method was piloted with a group of fluent signing Deaf educational staff in 2014–2016, then used in 2022–2023 with a group of fluent signing professionals to examine inter- and intrarater reliability when coding parameters of sign accuracy. Results: Interrater reliability measured by Gwet's Agreement Coefficient, was “good” to “very good” across the four phonological parameters that are components of every sign: location, movement, handshape, and orientation. Conclusions: The findings of this study indicate that the HART can be a reliable tool for coding the accuracy of location, orientation, movement, and handshape parameters of Auslan phonology when used by professionals fluent in Auslan. The HART can be utilized with any sign language to gather word level sign language samples in a written form and document the phonological accuracy of signed productions.

The Effect of PA Intervention through Remote Collaboration for CCK

Objectives: This study aimed to examine the effects of a phonological awareness and phonics intervention through remote collaboration between SLPs and teachers on the decoding skills of children from multicultural families with poor reading in the lower grades of elementary school. Methods: In this study, a multi-probe design was used to verify the effect of the intervention. The subjects were 3 multicultural children with poor reading skills (grades 1-2) and 3 teachers for each child. Their reading skills were examined through tests of phonological awareness and decoding (word level) in the baseline, intervention, and maintenance phases. Additionally, tests of decoding (sentence level) were conducted in the generalization phase to examine generalization at the sentence level. In this study, SLPs provided instruction to each child twice a week and also provided guide videos to each teacher once a week for 2 months. Results: All children were found to have high improvements in decoding skills (word level) and phonological awareness skills. Additionally, as a result of standardized reading and writing test (QRW), all children’s post-intervention raw scores were improved in all areas. In the generalization phase, all children showed improvements in decoding skills (sentence level). In the maintenance phase, all children were found to have a large effect size. Conclusion: The results show that phonological awareness and phonics intervention through remote collaboration between two experts had positive effects on the decoding skills of multicultural children with poor reading. Also, this intervention improved teachers’ teaching efficacy. Thus, the implementation of studies for collaboration between experts is suggested.

Mean length of utterance in words (MLUw) and lexical diversity among Urdu speaking children between 3 to 3.11 years of age

Background: Language is an excellent and convenient means of communication. Mean Length of Utterance (MLU) measures linguistic productivity in children. Objectives: to find the mean length of utterance in words and lexical diversity among children aged 3-3.11 years typically developing Urdu-speaking. Methods: The data of n=100 children of both genders of age between (3 to 3.11 years) who performed on blank levels 1 and 2 was collected through a cross-sectional survey. The data was collected from daycare centers and schools in Rawalpindi and Islamabad from November 2022 to June 2023. MLU calculation procedure was carried out through activities from ICW and Blank Level. For the calculation of MLU, the formula was used. Results: The Mean Length of Utterance at word level (MLUw)of children aged between 3 to 3.11 years was 3.43. In lexical diversity (LD) children have used Nouns (NN) most abundantly as 55% of Verbs (VB) with 35% Tense Auxiliaries (TA)30%. Conclusion: It is found that children's MLUw of 3.43 between the ages of 3 and 4 indicates an increase in speech complexity. Nouns, verbs, and tense auxiliaries have been recognized as the main elements via lexical analysis.

AI and the BoLC: Streamlining legal translation

Artificial Intelligence (AI) applied in the legal field has gained considerable ground in the recent years and is used in many fields, amongst which the legal one is of no exception. This paper wishes to explore the quality of the translation (from English into Italian) of an arbitration clause performed by the ChatGPT chatbot. To do so, the automatically generated target text is post-edited by consulting the BoLC (Bononia Legal Corpus) and the web as corpus in the gov.it Italian governmental domain. General and legal dictionaries are also used. The paper findings report some inaccuracies at word level which are easily tackled by corpus consultation. In view of the results obtained, however, complete reliance upon AI-driven solutions is not recommendable at the time being. Conversely, the use of ad hoc corpora and of targeted web searches are the most feasible and reliable solutions, although time-consuming.

Natural Language Processing Approaches to Text Data Augmentation: A Computational Linguistic Analysis

In the context of Natural Language Processing (NLP) tasks, problems such as insufficient or skewed data are frequently encountered. One practical solution to this problem is to generate additional textual data. Text Data Augmentation (TDA) refers to small changes made to accessible text at the character, word, or sentence level to generate synthetic data that is subsequently inserted into data loaders to train the model. By producing synthetic data, models can learn from a larger range of instances and, hence, enhance their resilience and generalization skills. Despite the fact that the entire NLP community has extensively studied many NLP DA approaches, recent research on the subject suggests that the relationship between the several DA techniques now in use is not entirely known in practice. Therefore, this study applies and extends the advances of TDA to encounter and cover varied tools on multiple settings or contexts. To carry out a thorough practical implementation of NLP DA approaches, comparing the way they perform and highlighting some of the significant similarities and differences in these various scenarios, this work depends on different tools of easy data augmentation and neural-based augmentation. This study suggests that some typical DA techniques might not be suitable in some circumstances or text environments. Specifically, according to the initial results, the context and word count of a text may have a significant impact on the quality of the synthetic data.

Lexical Density of Nursery Rhymes in Translation From English Into Indonesian

Cocomelon is a popular animated series for kids with billions of viewers on YouTube, filled with fun songs and stories that both entertain and teach. Because it is watched by children all over the world, Cocomelon has been translated into many languages, including Indonesian. Translating these videos isn't just about swapping words from one language to another, it involves careful thinking about the differences in language and culture to make sure the content stays meaningful and relatable for kids everywhere. This paper explains the concept of equivalence that exists at the word level in translations in the English and Indonesian versions of children's songs with a focus on the challenges faced by translators and the strategies used to overcome these challenges. Mona Bakar's framework on non-equivalence and studies that examine issues such as culturally specific concepts, grammatical differences, semantic complexity as well as the use of idiomatic expressions, comprehensive analysis of lexical density. Selected songs such as “Wheels on the Bus,” “Twinkle-Twinkle Little Star,” “Bath Song,” and “Old MacDonald” are revealing in the visible differences between the English versions. and Indonesia in its findings show that the Indonesian version has a high lexical density and is an indication that more words are needed to convey the same meaning.

The Contributions of Short-Term Memory to Writing at the Sublexical, Lexical, and Discourse Level in Beginning Writers

ABSTRACT Purpose The purpose of this study was to examine the direct and indirect relations between short-term memory (STM), reading, oral language, and writing at the letter, word, and discourse levels in young, developing writers both cross-sectionally and longitudinally. Method Participants were 449 English-speaking kindergarten students (52% female) from 41 schools in the U.S. Results The cross-sectional analysis indicated that STM was directly related to oral language (β = 0.62), letter naming fluency (β = 0.43), word reading (β = 0.54), and reading comprehension (β = 0.42), all of which were statistically significant. Results supported a complete mediation model with STM having an indirect effect on writing via reading and oral language. In the cross-sectional analysis, the total effects of STM on handwriting (β = 0.51), spelling (β = 0.57), and written composition (β = 0.47) were substantial and statistically significant. In the longitudinal analysis, the total effects of Fall STM on Spring writing outcomes, handwriting (β = 0.42), spelling (β = 0.70), and written composition (β = 0.44) were also substantial and statistically significant. Conclusion Findings confirm that STM is a foundational cognitive skill necessary for early writing; however, they provide a more nuanced understanding by showing that the contribution of STM is indirect through reading and oral language. Writing is complex, requiring interaction among higher-order domain-specific and domain-general components. Along with other recent studies, our findings highlight the need to examine the direct and indirect effects for a more precise understanding of skills that support writing development.

The development of English language connected speech perception skills: an empirical study on Chinese EFL children.

Connected speech processes (CSPs) occur randomly in everyday conversations of native speakers; however, such phonological variations can bring about challenges for non-native listeners. Looking at CSP literature, there seems to be very few studies that involved young foreign language learners. Therefore, the present study aimed to explore the development of connected speech perception skills by focusing on 201 9- to 12-year-old Chinese EFL children. It also incorporated systematic error analysis to further probe into the specific perceptual difficulties. The results indicate that: (1) Despite a significantly ascending trend for the overall growth of perception skills, no significant differences were found between 11 and 12 year olds in elision and contraction, which suggests that the developmental trend varied depending on different CSP types; (2) Although random errors decreased with age, the number of lexicon and syntax errors gradually increased, and the distribution of perceptual errors shifted from the level of words and syllables to that of phonemes; (3) The primary types of errors resulting in the perception difficulties for elision and contraction were consonant errors, grammatical errors and morphology errors. Ergo, this study enhances the understanding of connected speech perception among EFL children and provides some implications for EFL/ESL listening instructions.

Mixing Codes in Comments on Youtube Video Posts Debate for Presidential and Vice Presidential Candidates for the 2024 Election

This February, Indonesia will hold a democratic party to elect members of parliament and the president. Of course, the political and national atmosphere became warm with various political ripples and upheavals occurring, including political campaign activities, both political party campaigns and campaigns for Presidential and Vice Presidential Candidates. Various attributes and activities of political campaign speeches are interesting to study from various aspects, including the aspect of language use. This research aims to identify forms of code mixing found in KompasTV YouTube comments. This research method uses a descriptive method with a qualitative approach. The data for this research are comments from internet users (netizens) that researchers observed through screenshots from the comments column on the KompasTV YouTube channel. The data comes from the comments column of the KompasTV YouTube channel which contains elements of code mixing. Data collection techniques use listening, note-taking and documentation techniques. Data were analyzed using the Mills and Huberman model through the stages of data identification, data reduction, data analysis, and data presentation. Based on the research results, it was concluded that the code mixing that many netizens do in the comments column occurs at the word level. This is very possible because comments in word form are simpler, more practical and "to the point".

Next-Gen Linguistic Intelligence Image Summarizer for Documents Using Attention Model

Abstract: The process of producing a succinct synopsis of a text involves two primary techniques: both abstractive and extractive While abstractive summarising rephrases text in a way that is more human-like, extractive summarization chooses important statements straight. handbook Because it takes time and money to summarise because it requires reading through big texts, an automated text model with lots of features It is suggested to use both extractive and abstractive methods for summarising. Important sentences that are enhanced by linguistic features are identified by a feature-rich extractor. These sentences are then further enhanced by an abstracter that adds components such named entity tags and part-of-speech tags. Moreover, a loss function is presented to match attentions at the word and sentence levels. At last the summarized text is converted into image by using parsing and other relatable techniques.