Συνέδρια ΔΙΠΑΕ

Μόνιμο URI για αυτήν την κοινότηταhttps://repository.ihu.gr/handle/11544/48408

Περιηγούμαι

Πλοήγηση Συνέδρια ΔΙΠΑΕ ανά Συγγραφέα "Delianidi, Marina"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω
Φιλτράρετε τα αποτελέσματα πληκτρολογώντας τα πρώτα γράμματα
Τώρα δείχνει 1 - 2 από 2
  • Μικρογραφία εικόνας
    Τεκμήριο
    Beyond the Chatbot: Co-Learning and Co-Teaching through a Dual-Persona Generative-AI Assistant
    (2025-12-12) Mizeli, Chaido; Delianidi, Marina; Diamantaras, Konstantinos; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
    Abstract—In this paper we present a generative-AI application developed to support both teachers and students in secondary education. The system employs two Large Language Models (LLMs) – Gemini and DeepSeek, and a Small Language model (SLM) – Gemma, integrated within a Retrieval-Augmented Generation (RAG) framework to create a pedagogically grounded, Greek-language assistant capable of adapting its reasoning and communication style to the user’s role. Unlike conventional chatbots, the assistant introduces Pedagogical Persona Switching, a dual-role mechanism that enables the same AI model to act as both a teaching companion and a learning guide. When interacting with a student, the assistant produces accessible, curriculum-grounded explanations that promote conceptual understanding while remaining faithful to the officially approved educational sources. When addressing a teacher, the assistant generates diverse artifacts, including assessment items, classroom activities, and structured lesson plans. This dual-persona approach bridges knowledge retrieval with pedagogical creativity, enhancing differentiated learning. Utilizing a RAG paradigm tailored to the Greek educational domain, the architecture segments official textbooks into coherent units. Enriched with specific metadata, these units preserve curricular structure and instructional context, demonstrating how generative AI optimizes modern instructional design. These units are represented by domain-specific sentence embeddings fine-tuned on Greek semantic similarity tasks. This linguistic adaptation ensures precise meaning alignment between user queries and content segments, maintaining fidelity to the official curriculum. All representations are organized within a retrieval index, guaranteeing transparent and source-grounded generation. The initial case study focuses on Home Economics in Greek lower-secondary education (Grades A–B), a cross-disciplinary subject that integrates elements of economics, health education, and social responsibility. The assistant has been developed to support both learners and educators in complementary ways. In future classroom implementations, students will be able to use it to clarify key concepts such as financial literacy, resource management, and healthy living, while teachers could employ it to design authentic instructional materials, formative assessments, and classroom activities aligned with the official curriculum. Home Economics (HE) was chosen as the pilot domain because it is a non-STEM, value-oriented subject focused on early adolescents, emphasizing ethical reasoning, social awareness, and everyday decision-making. This unique context allows us to explore generative AI's pedagogical potential beyond traditional core subjects, particularly its capacity to foster reflective reasoning and civic responsibility in young learners. The project introduces an innovative generative framework combining three key elements: (a) a localized, curriculum-specific Retrieval-Augmented Generation (RAG) environment, (b) semantic educational text segmentation, and (c) adaptive persona-driven prompting. More than just a technical solution, this framework demonstrates how generative AI can be pedagogically and linguistically aligned with national curricula and local educational contexts. The resulting model offers a reproducible foundation for future AI integration into authentic learning settings, supporting teacher agency, learner engagement, and values-based secondary education. The study elevates the concept beyond a simple chatbot, proposing a structured, contextually adaptive framework for pedagogical generative assistants that effectively bridge technology, curriculum, and human learning.
  • Μικρογραφία εικόνας
    Τεκμήριο
    Earthquaker-AI: A Retrieval-Augmented Generation Framework with Rubric-Based Assessment for Primary School Earthquake Education
    (2025-12-12) Kokkinou, Xanthi; Mizeli, Chaido; Koulaxidou, Nafsika; Delianidi, Marina; Diamantaras, Konstantinos; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
    Abstract—This paper presents Earthquaker-AI, a hybrid educational framework that builds upon a previously implemented educational robotics project by integrating a conversational artificial intelligence assistant based on Retrieval-Augmented Generation (RAG), aiming to enhance earthquake preparedness and conscious action among primary-school students. The system extends the earlier award-winning STEM project Earthquaker, which received 3rd place in the Hellenic WRO Competition (2020), and moves from mechanical simulation through Lego WeDo2 activities to cognitive and metacognitive processing. The robotics component, developed in the original Earthquaker project, employed Lego WeDo2-based automation to simulate seismic response, allowing students to interact with sensors and actuators that function as tangible representations of core protective actions. At the cognitive level, the conversational assistant adopts a Retrieval-Augmented Generation architecture, in which evidence retrieval precedes language generation, so that responses are grounded in official educational material and aligned with the intended pedagogical framework. The assistant serves a dual role: it operates as a guided learning mechanism that aligns students’ responses with institutional safety guidelines, while simultaneously providing rubric-based verbal feedback that supports self-regulated learning and the development of calmness under emergency conditions. Earthquaker-AI follows a progressive learning trajectory aligned with students’ cognitive development across age stages. In the early grades, the emphasis is placed on basic recognition of earthquake-safety actions through simple multiple-choice questions, in which students select the most appropriate answer, supporting orientation toward the recommended behavior during seismic events. Assessment at this stage is conducted using a two-dimensional rubric focused on action recognition and emotional regulation. In middle grades, learning activities become more demanding, requiring students to identify the correct sequence of actions through multiple-choice questions, evaluated using a three-axis rubric that captures organized thinking and decision-making. In upper grades, the approach shifts from recognition to verbal production, with students providing short written responses assessed through a four-dimensional rubric. The additional dimension of clarity of expression reflects increasing metacognitive maturity, as it requires structured reasoning, justification, and precise articulation. The system also includes a dedicated dialogic module that leverages RAG to retrieve and synthesize evidence-based answers. Student queries are semantically matched with selected excerpts from official earthquake-safety guidelines, from which pedagogically safe and accurate responses are generated. Experimental evaluation demonstrates high answer groundedness (0.84) and accuracy (0.85), together with a low hallucination rate (0.07), indicating stable and evidence-aligned behavior under the evaluated experimental conditions. Overall, Earthquaker-AI presents an integrated educational approach to earthquake preparedness, bringing together hands-on engagement, information processing, and reflective practice. The sequence of observation, physical interaction, interpretation, and verbal articulation provides a holistic learning experience. At the same time, the combined use of robotic processes, analytic rubrics, and artificial intelligence promotes technological literacy, self-regulation, and responsible use of digital systems in primary education, contributing meaningfully to earthquake preparedness and the early development of crisis-management skills.