Page 50 - tech fest 2025 ב״ש חוברת תקצירים

50
Online Analysis of Lebanese and Israeli Mindsets during Lebanon-Israel Conflict
CS-C-21
Bar Partush; barpa@ac.sce.ac.il Ron Yehuda; ronye2@ac.sce.ac.il
Advisor: Dr. Aviad Elyashar
SCE - Shamoon College of Engineering, Be’er-Sheva
The Lebanon-Israel conflict has been ongoing since October 2023, deeply impacting the populations on both sides. Analyzing the sentiments, beliefs, and attitudes of Lebanese and Israeli social media users during this period provides valuable insights into their experiences and perspectives. This project applied machine learning tools for text analysis of X posts (formerly Twitter tweets), extracting user features and deriving meaningful insights from the social media data. Special emphasis was placed on the heterogeneity within these two societies, which included their diverse ethnic and religious groups. By leveraging natural language processing techniques, the study uncovered patterns and trends that reflect the collective and individual responses to the conflict, contributing to a deeper understanding of its societal implications.
Keywords: conflict studies, multilingual discourse, natural language processing, sentiment analysis, social network analysis, Twitter analysis
Anomaly Detection in Industrial Control Systems Using Spiking Neural Networks
CS-D-22
Michael Kaminer; narkaminer@gmail.com Eviatar Cohen; eviatar12341@gmail.com
Advisor: Dr. Aviad Elyashar1, Prof. Shlomo Greenberg1 1SCE - Shamoon College of Engineering, Be’er-Sheva
Industrial control systems (ICSs) are critical to essential infrastructure. As such, they are frequent targets of cyberattacks that trigger system anomalies. However, anomaly detection methods, which are often based on neural networks, may struggle with the complex temporal patterns and high- dimensional data typical in ICS environments. This project investigated the use of spiking neural networks (SNNs) for anomaly detection in ICSs. Inspired by the human brain, SNNs are well-suited for processing temporal data and identifying patterns in time-sensitive environments. The project developed and trained SNN models to detect abnormal behavior and potential cyber threats in ICS data streams, leveraging their event-driven architecture and computational efficiency. The results improved detection accuracy, system responsiveness, and overall cyber resilience in real-time ICS monitoring.
Keywords: anomaly detection, cybersecurity, industrial control systems, spiking neural networks