Shimaa Ahmed

Research Abstract:

Machine Learning (ML) systems power our everyday interactions with digital services. Social media platforms, smart devices, and governments employ ML for a myriad of tasks, including identity recognition, targeted advertising, facial recognition, speech recognition, and personalization of user experiences. The widespread adoption of ML systems, however, comes at a considerable societal cost in terms of privacy, bias, and trust. ML applications raise concerns about data privacy, as individuals’ information is collected, analyzed, and potentially shared without their consent. In the realm of public safety, the use of ML in surveillance can encroach upon individuals’ rights and amplify concerns about abuse of power. Furthermore, ML algorithms used in decision-making processes can perpetuate bias and discrimination, affecting the fairness and justice of outcomes. Finally, the spread of misinformation through ML-generated deepfakes challenges the integrity of free speech and trust in digital media. These examples emphasize the need to develop frameworks that ensure a responsible deployment of these technologies. Through my research, I examine the risks posed by ML algorithms and develop robust systems to mitigate these risks and empower users with agency over their personal data. Specifically, I have explored different ubiquitous ML applications encompassing various data modalities such as speech, text, and vision. In my future work, I will focus on understanding data privacy and ownership, interpretability, and model provenance in the generative AI era.

Bio:

Shimaa Ahmed is a Ph.D. dissertator in the Electrical and Computer Engineering Department at the University of Wisconsin-Madison under the supervision of Prof. Kassem Fawaz. She also works as a graduate student lecturer at UW-Madison teaching a Capstone course on machine learning applications. Shimaa's research interests include trustworthy machine learning, speech technologies, and privacy and security. Her dissertation focuses on developing systems and tools for the practical and responsible deployment of machine learning systems. She has presented her research through conferences and invited talks. Shimaa currently mentors three graduate students, two undergraduates, and a high school student. She is an active member of the Women in ECE group at UW-Madison and she hosted a session for the Expand Your Horizon career exploration day for girls and non-binary youth in middle school in April 2023. Upon the completion of her Ph.D. degree in December 2023, Shimaa intends to pursue a career in academic research and teaching. She received her B.Sc. and M.Sc. degrees with distinction in Electrical and Computer Engineering from Ain Shams University, Cairo, Egypt.