Mohamed Abdelhamid

MSc


Mohamed is a computer science graduate from Innopolis University, where he earned his bachelor’s degree with a thesis titled “Exploring the Impact of Conceptual Bottlenecks on Adversarial Robustness.” Originally from Egypt, Mohamed is currently pursuing a master’s degree in the Elite Data Science program, a joint initiative between LMU, TUM, and the University of Augsburg. His academic interests primarily focus on adversarial attacks and the safety of machine learning models.

Mohamed’s research explores how Concept Bottleneck Models (CBMs) can enhance AI interpretability and robustness, particularly against adversarial attacks. His bachelor’s research involved extensive experiments comparing the resilience of CBMs to conventional neural networks, demonstrating CBMs’ improved robustness.

Driven by a passion for AI safety, Mohamed aims to contribute to the development of reliable and secure AI systems. He is particularly interested in the intersection of machine learning theory, adversarial robustness, and real-world applications in critical domains.

relAI Research Areas: Mathematical and Algorithmic Foundations, Algorithmic Decision-Making

relAI Central Themes: Safety, Security, Responsibility