Search Results (150)

This article investigates offline recognition of handwritten Kazakh text in the Latin script using a convolutional recurrent neural network. The relevance of the study is determined by the transition of the Kazakh language to the Latin alphabet and the need to automate the processing of handwritten documents. The proposed model consists of a convolutional neural network feature extractor, two bidirectional long short-term memory layers, and a Connectionist Temporal Classification decoder. The convolutional layers extract visual features from word images, the bidirectional recurrent layers model the sequential relationships between characters, and CTC enables end-to-end training without explicit character-level segmentation. A specialized dataset named KazEsim, containing 20,000 handwritten Kazakh name images, was created and divided into writer-independent training, validation, and test subsets. Experimental results showed a character accuracy rate of 96.5% and a word accuracy rate of 92.3%. Compared with a conventional CNN baseline, the proposed CRNN model improved character accuracy by 6.1 percentage points and word accuracy by 9.2 percentage points. The proposed model also outperformed the fine-tuned TrOCR-small comparative baseline while requiring fewer parameters and lower inference latency. These findings demonstrate the effectiveness of CNN–BiLSTM–CTC sequence modeling for offline recognition of handwritten Kazakh words in the Latin script. Full article

We introduce finite-observation conditional guarded language operators on the complete length-based ultrametric space of formal languages over a finite alphabet. A finite observation map selects a guarded branch according to the membership of prescribed test words. Although each branch is Banach-contractive, branch switching may destroy global Banach contractivity. We derive depth–residual conditions yielding a max-type Kannan inequality with an explicit constant $\alpha <1/2$. Consequently, the operator has a unique fixed language, and the Picard iteration converges from every initial language. We also construct multi-branch operators that are Kannan-contractive but not Banach contractive and obtain residual error estimates, explicit convergence bounds, finite-depth certification, and eventual stabilization of the selected branch. Full article

This article argues that Denise Levertov’s Vietnam-era poetry develops a poetics of language ethics: a formal practice through which words are tested against bodies, suffering, agency, and moral judgment. Focusing on “During the Eichmann Trial,” “Life at War,” and “Prologue: An Interim,” the essay examines how Levertov exposes the bureaucratic, legal, and military languages that render violence administrable and morally tolerable. Drawing on Levertov’s poetics, postwar American poetics, Hannah Arendt’s account of moral thoughtlessness, and critical discourse analysis, it shows how quotation, fragmentation, lineation, apostrophe, parataxis, and bodily imagery resist abstraction and restore ethical perception. By situating these poems within the Eichmann trial, the Vietnam War, the Boston Five prosecution, and antiwar protest culture, the article presents Levertov’s work as a historically specific model of linguistic vigilance and poetic witness, one that remains alert to the ethical vulnerability of public language today. Full article

Modernising historical literature into contemporary language is a form of diachronic intralingual translation that supports access to written cultural heritage. For low-resource languages such as Slovene, this task is hindered by orthographic, lexical and syntactic shifts, as well as the scarcity of parallel data. We present a two-stage pipeline that combines a fine-tuned No Language Left Behind (NLLB) model with Claude Opus 4.8 post-editing for the modernisation of 16th-century Slovene literature, retaining archaic words and phrases while normalising the alphabet, orthography and grammar. Using Jurij Dalmatin’s 1584 Bible and its 2017 modernised edition, we constructed an aligned parallel corpus of 14,876 sentence and clause-level pairs through Bohorič-to-Gaj normalisation and LaBSE-based embedding alignment. The hybrid pipeline achieved the best overall scores, reaching BLEU 45.78, CHRF 67.68 and METEOR 71.72, with TER 43.25 and CER 34.00. Its gains over the standalone LLM baseline were large and statistically significant across all the metrics, while its improvement over the fine-tuned NLLB model was smaller and significant mainly for overlap-based measures. We applied the pipeline further to Tulščak’s Kerszhanske leipe molitve (1579), producing the first preliminary modernisation of the earliest known Slovene prayer book assessed qualitatively and tested the generalisation on additional 16th-century texts, including an out-of-domain legal text. The results demonstrate that combining task-specific neural machine translation with controlled LLM post-editing offers a practical strategy for modernising low-resource historical texts and contributes a reusable methodology for digital cultural heritage preservation. Full article

This study presents the design, development, and evaluation of iSign, an Android-based educational mobile application developed to support sign language learning among deaf and hard-of-hearing individuals in rural communities. The project was conducted using a research and development (R&D) approach guided by the ADDIE instructional design framework. A preliminary needs assessment, based on municipal records and field validation, revealed limited access to structured sign language education and assistive learning resources in selected barangays of Odiongan, Romblon. The iSign application was designed as a user-centered and accessible mobile learning tool incorporating alphabet learning (A–Z), number recognition (0–9), a dictionary containing 160 commonly used vocabulary words with definitions and corresponding sign language video demonstrations, and multimedia content such as nursery rhyme songs interpreted in sign language. The system was developed using Android Studio and structured to support offline accessibility to accommodate low-connectivity environments. The application was evaluated in terms of functional suitability, performance efficiency, compatibility, usability, reliability, security, maintainability, and portability. Twenty-one community participants and four information technology experts assessed the system using a five-point Likert scale. The overall weighted mean rating of 4.0 (“Agree”) indicates that the developed application met acceptable software quality standards and user satisfaction levels. The findings demonstrate that iSign is a functional and accessible mobile learning application that can serve as a supplementary tool for sign language education in underserved communities. Full article

Coptic represents the final stage of the Ancient Egyptian language and remains an important component of Christian and Mediterranean cultural heritage. Although several digital resources exist for Coptic textual corpora, Greek-oriented computational tools for the interpretation of Coptic inscriptions on artifacts remain limited. This study presents the design and early implementation of a semi-automated software tool for the computer-assisted translation of Coptic inscriptions into Greek, with optional English support. The tool combines a Coptic–Greek digital dictionary, an interactive character-selection interface, and two dictionary-search strategies: a length/alphabetically structured linear search and a weighted linear search based on expected word frequency. The application is intended to support scholars working with inscriptions on fragile or fragmented cultural heritage objects, where full automation is not realistic and human supervision remains essential. The paper describes the linguistic and material challenges of Coptic inscriptions, the structure of the lexical database, the interface design, and the planned use of Coptic corpora for improving retrieval efficiency. The proposed approach contributes to cultural heritage digitization by offering a practical, expandable, and user-oriented framework for supporting the study, interpretation, and preservation of Coptic inscriptions. Full article

Orthographic depth varies across alphabetic writing systems and plays a central role in spelling acquisition. In immersion education, a second language (L2) is used as a language of instruction for part of the curriculum, such that learners are primarily confronted with its writing system during the initial stages of literacy development. This early exposure may shape the spelling strategies subsequently deployed in the first language (L1), which also corresponds to the dominant language of the surrounding community. This article provides a structured review of key mechanisms involved in spelling acquisition, orthographic depth, and cross-linguistic influence in bilingual and immersion contexts. On this basis, it proposes a conceptual and predictive framework specifying how the orthographic depth of the instructional language modulates spelling strategies and spelling error profiles in L1. Focusing on French-speaking pupils enrolled in immersion programmes with L2s characterised by either predominantly phonemic or opaque orthographies, the framework integrates strategy-based models of orthographic development. The model distinguishes phonological, lexical, and morphographic components of orthographic knowledge and predicts that immersion in phonemic-dominant orthographies favours phonographic dominance and regularisation patterns, whereas immersion in opaque orthographies promotes greater reliance on lexical–orthographic strategies, resulting in distinct and systematic spelling error profiles in French. Full article

We study positive–negative guarded systems of language equations over a fixed finite alphabet. The ambient space is the complete ultrametric space of all formal languages equipped with a length-based distance, where two languages are close whenever they agree on all words up to a sufficiently large length. The systems considered here contain both positive recursive dependencies and negative dependencies expressed through language complements. To handle this mixed structure, we introduce a suitable product order on pairs of languages and prove that the associated system operator has the weak monotone property. We show that the complement is an isometry for the length-based ultrametric and establish a signed wrapping estimate for guarded positive and negative language terms. These estimates lead to an ordered contraction principle for comparable pairs. As a consequence, the canonical lower and upper Picard iterations converge to the same limit, which is the unique fixed pair of the system. We also derive an explicit convergence rate and a finite-depth certification result: after a prescribed number of iterations, the approximants agree with the fixed-point semantics on all words below a given length. Additional symmetry assumptions are shown to force the unique fixed pair to be diagonal, reducing the system to a single language equation. Finally, we discuss an application to trace-based policies for tool-using AI agents. In this interpretation, finite executions of an agent are represented as words over an alphabet of observable tool-events, and the two components of the fixed point provide a stable semantics for policy-defined admissible and risky trace classes. The resulting framework gives a mathematically certified method for finite-depth analysis of recursive trace-based policies based on ultrametric fixed-point techniques. Full article

This paper proposes a formal framework for the study of mathematical language at the intersection of fixed point theory, formal language theory, and academic discourse analysis. Mathematical texts are modeled as languages over a finite alphabet of discourse tokens, combining natural-language expressions with symbolic content. To suppress irrelevant symbolic variation, we introduce a normalization procedure in which concrete mathematical expressions are replaced by an abstract placeholder while the surrounding linguistic structure is preserved. Within this framework, we define enrichment operators on phrases and the induced operators on languages, which model admissible stylistic and structural transformations of mathematical discourse. The collection of all languages over a fixed alphabet, ordered by inclusion, is shown to form a complete lattice, allowing the application of the Knaster–Tarski fixed point theorem. As a consequence, stable linguistic configurations can be interpreted as fixed points of the induced enrichment operator. We further show that different initial languages may lead to different fixed points under the same operator, reflecting the existence of multiple stable forms of mathematical expression. In addition, we introduce a notion of lexical distance based on frequency distributions of discourse units, which provides a quantitative tool for comparing languages. The illustrative analysis suggests a saturation effect: while enrichment increases the overall distance from the initial language, the incremental changes between successive stages remain bounded, indicating a tendency towards stabilization. A concrete illustrative example based on a classical theorem from mathematical analysis demonstrates how a proof evolves through successive levels of enrichment, from a minimal linguistic core to more elaborate stylistic realizations. The proposed framework thus provides a bridge between formal language models and the linguistic structure of mathematical discourse, offering a new perspective on the organization, stability, and variation of mathematical language. Full article

Recognition of the American Sign Language (ASL) alphabet is of utmost importance in bridging the communication gap between the hearing-impaired and the hearing. However, robust classification remains difficult because some hand gestures are morphologically very similar. To address this problem, this study presents CT-Net, a hybrid deep learning architecture that integrates ConvNeXt-Tiny with a lightweight Transformer encoder. CT-Net combines convolutional feature extraction and self-attention mechanisms, which enable it to capture fine-grained local patterns and long-range spatial dependencies effectively. The proposed model was extensively compared with various architectures including traditional CNNs, Transformer-based models, hybrid machine-learning approaches and recent lightweight hybrid networks. The experimental results show that CT-Net achieved the best overall performance with a peak accuracy of 95.67% on the enhanced ASL dataset. Ablation studies demonstrate the effectiveness of our design choice. CT-Net achieves a strong trade-off between recognition accuracy and computational efficiency with an inference rate of 163.55 Frames Per Second (FPS). These findings highlight the potential of hybrid frameworks as a powerful tool for fine-grained gesture recognition tasks. Full article

We study a class of wrapping operators acting on the space of formal languages over a fixed finite alphabet. The underlying space is equipped with a length-based ultrametric, in which two languages are close whenever they coincide on all sufficiently short words. We prove that every wrapping operator generated by a finite family of guards with positive total guard length is a contraction. As a consequence, Banach’s contraction principle yields existence and uniqueness of a fixed point for the corresponding recursive language equation, together with convergence of the Picard iteration from an arbitrary initial language. We also obtain an explicit quantitative estimate for the rate of convergence. This makes it possible to determine how many iterations are sufficient to recover the fixed point correctly on all words up to a prescribed length. Several examples illustrate the theory, including operators with different guard lengths and a case showing that convergence in the length-based ultrametric does not coincide with set-theoretic convergence. An application to recursive structures and document validation is also presented, including recursive data formats, abstract syntax trees, and a restricted fragment of JSON schemas. The results provide a formal foundation for validation together with explicit bounds for correctness on inputs of bounded length. Full article

A brain–computer interface (BCI) enables direct communication between the brain and external devices by translating neural activity into executable control commands. Among electroencephalography (EEG)-based paradigms, steady-state visual evoked potential (SSVEP) is widely adopted due to its high signal-to-noise ratio, robustness, and minimal calibration requirements. While SSVEP-based spellers have been extensively investigated, many existing systems rely on high-channel-density EEG recordings and computationally complex processing pipelines, and are primarily designed for alphabetic input structures. In this study, we present an SSVEP-based Korean speller that integrates the Cheonjiin keyboard layout to support intuitive composition of Hangul syllables. The proposed system adopts a simple configuration, employing only five visual stimulation frequencies (6.67–12 Hz) and two occipital EEG channels (O1 and O2), with real-time frequency recognition performed using canonical correlation analysis (CCA) within a 1.5 s sliding window. EEG signals were acquired at 200 Hz using an OpenBCI Ganglion board, band-pass filtered (5–45 Hz), and processed with harmonic sinusoidal reference templates for multi-frequency classification. The proposed interface generates five control commands (up, down, left, right, and select), enabling directional cursor navigation and character confirmation on a 4 × 4 virtual Cheonjiin keyboard. Experimental validation with three healthy participants demonstrated an average classification accuracy of approximately 82% and an information transfer rate (ITR) of 31.2 bits/min. Frequency-domain analysis revealed clear spectral peaks at the stimulation frequencies and their harmonics, indicating reliable SSVEP responses. The proposed system employs a simple two-channel configuration integrated with a Korean language-specific input structure, demonstrating that reliable SSVEP-based communication can be realized without computationally intensive algorithms or high-cost EEG acquisition systems. These findings demonstrate that reliable SSVEP-based communication can be achieved using a low-channel configuration without reliance on high-cost EEG equipment. Full article

Background/Objectives: With growing waves of migration, children speaking a home language different from the language of school literacy have become increasingly common in Western education systems. In this context, understanding and monitoring bilinguals’ reading development is crucial to inform both educational and clinical practices and ensure equitable services. The present study contributes to the literature by investigating learning patterns in single-word reading across primary school grades. Monolingual and bilingual children learning to read in an alphabetic orthography were examined. Methods: The sample consisted of 565 typically developing monolingual and bilingual primary school children from grades 1–5 (bilinguals = 162). Participants completed a computerised Lexical Decision task (LDT) recording accuracy and response times, and standardised tests of reading and cognition. A parental questionnaire was used to gather socio-demographic and linguistic information. Results: Response bias-corrected accuracy rates in the LDT revealed an increase in sensitivity across school years after correcting for potential confounds (SES, vocabulary, nonverbal intelligence). No significant effect of bilingualism was observed. Response times for correct responses also decreased consistently across grades after controlling for the same confounds. Although no significant main effect of bilingualism emerged, an interaction with grade revealed a greater decrease in response times for second-grade bilinguals compared to monolingual peers. Conclusions: Monolingual and bilingual children showed comparable sensitivity rates and reading times, suggesting similar decoding skill acquisition. However, an earlier decrease in response times for bilinguals points to a facilitatory effect in the early stages of reading development, consistent with a bilingual advantage during skill learning. Full article

Sarcasm detection remains a challenging task in Natural Language Processing (nlp), especially for low-resource and non-standardized languages. Hence, this study addresses Franco-Arabic, a widely used form of online communication where Arabic words are written with Latin characters and numerals. Its informal nature and orthographic variation complicate sarcasm identification and limit the applicability of existing nlp models. We propose an approach that integrates transformer-based representations with auxiliary linguistic features and rule-based cues to capture both contextual meaning and sentiment-driven inconsistencies. This research opens the door to practical applications. In particular, future work will investigate integrating sarcasm detection into the marketing sector, where accurate recognition of sarcastic reviews can enhance sentiment analysis, customer segmentation, and personalized communication strategies. Full article

A writing system $\mathfrak{W}$ on an alphabet $\mathcal{A}$ is a tuple $(\mathfrak{R},\mathfrak{M})$, where $\mathfrak{R}$ is a finite set of text formation rules and $\mathfrak{M}$ is a finite rule application mechanism that generates texts on $\mathcal{A}$. A natural writing system forms natural language texts on an alphabet such as Sanskrit on Devanagari. An artificial writing system generates formal language texts on an alphabet such as Lisp on Unicode. Let $\mathbb{N}=\{0,1,2,\dots \}$ be the set of natural numbers. A function on natural numbers $f:{\mathbb{N}}^k\mapsto \mathbb{N}$, $0<k\in \mathbb{N}$, is nameable by a writing system on an alphabet if, and only if, the system can generate a text on the alphabet that names f and no other function. We show that there exist functions on natural numbers unnameable in principle in that they cannot be named by any writing system on any alphabet. Our results imply the following computability-theoretic hierarchy of functions on natural numbers: $\mathrm{computable}⊊\mathrm{partially}\mathrm{computable}⊊\mathrm{nameable}⊊\mathfrak{F}$, where $\mathfrak{F}$ is the set of functions on $\mathbb{N}$. Full article