Katia Lida Kermanidis

In recent years, there has been a notable resurgence of interest in the study of Ancient Greek within the context of the digital age. This paper presents an extended and systematic review of digital techniques, computational methodologies, resources, and tools developed specifically for the analysis and study of the Ancient Greek language. The review aims to highlight how contemporary computational approaches can support both linguistic research and educational applications for this historical language. The review begins with an outline of the morphological structure of Ancient Greek, followed by a discussion of key linguistic characteristics relevant to digital processing. Subsequently, the availability of corpora and computational resources for the development of Machine Learning applications in Ancient Greek is examined. The systematic review was conducted in accordance with the PRISMA 2020 guidelines. Identified research gaps are analyzed, and future research directions are proposed to advance work within the Ancient Greek Language Processing community. While previous study has offered general overviews of related work, this paper provides an expanded and focused synthesis that concentrates exclusively on techniques, resources, tools and methodologies tailored to Ancient Greek. The study demonstrates how a historical language such as Ancient Greek can be effectively investigated through digital means and made accessible to learners and researchers through a diverse range of computational approaches.

This research investigates topic-based sentiment classification in Greek educational-related data using transformer-based language models. A comparative evaluation is conducted on GreekBERT, XLM-r-Greek, mBERT, and Palobert using three original sentiment-annotated datasets representing parents of students with functional diversity, school directors, and teachers, each capturing diverse educational perspectives. The analysis examines both overall sentiment performance and topic-specific evaluations across four thematic classes: (i) Material and Technical Conditions, (ii) Educational Dimension, (iii) Psychological/Emotional Dimension, and (iv) Learning Difficulties and Emergency Remote Teaching. Results indicate that GreekBERT consistently outperforms other models, achieving the highest overall F1 score (0.91), particularly excelling in negative sentiment detection (F1 = 0.95) and showing robust performance for positive sentiment classification. The Psychological/Emotional Dimension emerged as the most reliably classified category, with GreekBERT and mBERT demonstrating notably high accuracy and F1 scores. Conversely, Learning Difficulties and Emergency Remote Teaching presented significant classification challenges, especially for Palobert. This study contributes significantly to the field of sentiment analysis with Greek-language data by introducing original annotated datasets, pioneering the application of topic-based sentiment analysis within the Greek educational context, and offering a comparative evaluation of transformer models. Additionally, it highlights the superior performance of Greek-pretrained models in capturing emotional detail, and provides empirical evidence of the negative emotional responses toward Emergency Remote Teaching.

This paper presents a comparative analysis of two state of the art Greek to English machine translation systems, ChatGPT-4 (paid version) and Gemini, focusing on their ability to handle specific linguistic phenomena. The phenomena under scrutiny include ambiguity, long sentences, modal verbs, idiomatic expressions, negations, and compositionality. By translating a carefully selected dataset from Greek to English, we evaluate the performance of both systems based on their ability to preserve meaning and produce grammatically correct translations. The results show significant differences between the way the two systems handle complex sentence structures and subtle contextual differences. Furthermore, we examine the strengths and weaknesses the two systems present in accuracy, fluency and overall quality. The study provides useful insights into how reliable modern machine translation systems are by exploring their strengths and limitations. The persuasiveness of these systems was tested in addressing various linguistic challenges and the paper suggests potential improvements for managing complex language patterns more accurately. These findings contribute to the on-going discussion about refining translation technologies for more practical and accurate real-world applications.

This study explores topic identification using text analysis techniques in Modern Greek interviews with parents of students with functional diversity during Emergency Remote Teaching. The analysis focused on identifying key educational themes and addressing challenges in processing Greek educational data. Machine learning models, combined with Natural Language Processing techniques, were applied for topic identification, utilizing cross-validation and data balancing methods to enhance reliability. The findings revealed the impact of linguistic complexity on topic modeling and highlighted the educational implications of analyzing qualitative data in this context. Among the models tested, the Naïve Bayes (Kernel) algorithm performed best when combined with lemmatization-based preprocessing, confirming that text normalization significantly enhances classification accuracy in Greek educational data. The proposed framework contributes to the analysis of qualitative educational data by identifying key parental concerns related to Emergency Remote Teaching. It demonstrates how text analysis techniques could support data-driven decision-making and help guide policy development for the inclusive and effective integration of Information and Communication Technology in education.

This study explores the application of transformer-based language models for automated Topic Classification in qualitative datasets from interviews conducted in Modern Greek. The interviews captured the views of parents, teachers, and school directors regarding Emergency Remote Teaching. Identifying key themes in this kind of interview is crucial for informed decision-making in educational policies. Each dataset was segmented into sentences and labeled with one out of four topics. The dataset was imbalanced, presenting additional complexity for the classification task. The GreekBERT model was fine-tuned for Topic Classification, with preprocessing including accent stripping, lowercasing, and tokenization. The findings revealed GreekBERT’s effectiveness in achieving balanced performance across all themes, outperforming conventional machine learning models. The highest evaluation metric achieved was a macro-F1-score of 0.76, averaged across all classes, highlighting the effectiveness of the proposed approach. This study contributes the following: (i) datasets capturing diverse educational community perspectives in Modern Greek, (ii) a comparative evaluation of conventional ML models versus transformer-based models, (iii) an investigation of how domain-specific language enhances the performance and accuracy of Topic Classification models, showcasing their effectiveness in specialized datasets and the benefits of fine-tuned GreekBERT for such tasks, and (iv) capturing the complexities of ERT through an empirical investigation of the relationships between extracted topics and relevant variables. These contributions offer reliable, scalable solutions for policymakers, enabling data-driven educational policies to address challenges in remote learning and enhance decision-making based on comprehensive qualitative evidence.

In the digital age, the rapid proliferation of misinformation and disinformation poses a critical challenge to societal trust and the integrity of public discourse. This study presents a comprehensive machine learning framework for fake news detection, integrating advanced natural language processing techniques and deep learning architectures. We rigorously evaluate a diverse set of detection models across multiple content types, including social media posts, news articles, and user-generated comments. Our approach systematically compares traditional machine learning classifiers (Naïve Bayes, SVMs, Random Forest) with state-of-the-art deep learning models, such as CNNs, LSTMs, and BERT, while incorporating optimized vectorization techniques, including TF-IDF, Word2Vec, and contextual embeddings. Through extensive experimentation across multiple datasets, our results demonstrate that BERT-based models consistently achieve superior performance, significantly improving detection accuracy in complex misinformation scenarios. Furthermore, we extend the evaluation beyond conventional accuracy metrics by incorporating the Matthews Correlation Coefficient (MCC) and Receiver Operating Characteristic–Area Under the Curve (ROC–AUC), ensuring a robust and interpretable assessment of model efficacy. Beyond technical advancements, we explore the ethical implications of automated misinformation detection, addressing concerns related to censorship, algorithmic bias, and the trade-off between content moderation and freedom of expression. This research not only advances the methodological landscape of fake news detection but also contributes to the broader discourse on safeguarding democratic values, media integrity, and responsible AI deployment in digital environments.

Simplicity in information found online is in demand from diverse user groups seeking better text comprehension and consumption of information in an easy and timely manner. Readability assessment, particularly at the sentence level, plays a vital role in aiding specific demographics, such as language learners. In this paper, we research model evaluation metrics, strategies for model creation, and the predictive capacity of features and feature sets in assessing readability based on sentence complexity. Our primary objective is to classify sentences as either simple or complex, shifting the focus from entire paragraphs or texts to individual sentences. We approach this challenge as both a classification and clustering task. Additionally, we emphasize our tests on shallow features that, despite their simplistic nature and ease of use, seem to yield decent results. Leveraging the TextStat Python library and the WEKA toolkit, we employ a wide variety of shallow features and classifiers. By comparing the outcomes across different models, algorithms, and feature sets, we aim to offer valuable insights into optimizing the setup. We draw our data from sentences sourced from Wikipedia’s corpus, a widely accessed online encyclopedia catering to a broad audience. We strive to take a deeper look at what leads to greater readability classification in datasets that appeal to audiences such as Wikipedia’s, assisting in the development of improved models and new features for future applications with low feature extraction/processing times.