Page 177 - tech fest 2025 ב״ש חוברת תקצירים

An AI-Based System for Enhanced Self-Learning
SE-C-10
Doron Swisa; seller4500@gmail.com Yan Frenklakh; yanfr2@ac.sce.ac.il
Advisor: Dr. Natalia Vanetik
SCE - Shamoon College of Engineering, Be’er-Sheva
This project introduces a learning system that incorporates the application of Large Language Model (LLM) and Retrieval Augmented Generation (RAG) for effective self-learning in deep subjects in discrete mathematics and formal logic. The model addresses three different issues: unreliable AI- generated information and “hallucinations,” and solving Natural Language Processing (NLP) tasks with the complex morphological structures of Hebrew. The model ensures the correctness, understandability, and accessibility of the sources to the end user. The system not only significantly increases the number of correct answers to questions by providing learners with many types of stimulation, but it also helps the student improve their ability to ask the right questions.
Keywords: discrete mathematics, formal logic, hallucinations, Hebrew nlp, logic learning model, retrieval-augmented generation, self-learning
Geospatial Community Mapping System for Emergency and Routine Preparedness
SE-C-11
Ofek Atayak; ofekat@ac.sce.ac.il Sahar Gabay; saharga@ac.sce.ac.il
Advisors: Dr. Hadas Chassidim1, Dr. Batel Yossef Ravid1 1SCE - Shamoon College of Engineering, Be’er-Sheva
This project presents a geospatial community mapping system designed to enhance urban planning by integrating local community perspectives. A targeted survey was distributed to residents to capture their perceptions of the community, with responses combined with demographic, geographic, and logistical data for comprehensive profiling. Using a GIS-based Plotly interface, the system enables users to visualize community structures, identify key characteristics, and match resources to local needs. A prototype of the system is currently being piloted in the city of Beer Sheva to demonstrate its functionality in a real urban environment. Combining data analysis and interactive design, the platform helps municipalities and emergency services make efficient decisions in both routine and emergency situations.
Keywords: community, mapping, gis, interactive visualization, urban planning
Book of Abstracts | 2025
177