Imaging Science Seminar: Understanding the Robustness in Machine Learning and its Importance in Imaging Science
Imaging Science Seminar
Understanding the Robustness in Machine Learning and its Importance in Imaging Science
Dr. Dimah Dera
Assistant Professor
Chester F. Carlson Center for Imaging Science
This presentation explores the concept of robustness in machine learning, focusing on its pivotal role within the realm of imaging science. We will shed light on the paramount importance of uncertainty quantification in bolstering the robustness of machine learning models.
Abstract:
Machine learning has gained significant prominence in the field of imaging science, revolutionizing tasks such as image recognition, medical diagnosis, and object detection. However, ensuring the reliability and effectiveness of machine learning models in imaging applications remains a critical challenge. Robustness in machine learning refers to a model's ability to maintain performance across diverse and challenging conditions, including noisy data, adversarial attacks, and domain shifts. The importance of robust machine learning cannot be overlooked in imaging science. Image data is inherently complex, with variations in lighting, scale, viewpoint, and heterogeneity. Robust models are essential to delivering consistent and accurate results, particularly in fields like medical imaging, autonomous vehicles, and surveillance systems. This presentation explores the concept of robustness in machine learning, focusing on its pivotal role within the realm of imaging science. We will shed light on the paramount importance of uncertainty quantification in bolstering the robustness of machine learning models. Uncertainty quantification is the process of estimating the confidence or uncertainty associated with model predictions. We will present state-of-the-art models for quantifying uncertainty and empowering the robustness of modern machine learning models towards real-world deployment in various applications.
Speaker Bio:
Dimah Dera is an Assistant Professor at the Chester F. Carlson Center for Imaging Science at RIT. She received her Ph.D. and M.S. in Electrical and Computer Engineering and M.A. in Mathematics from Rowan University. Dimah received the National Science Foundation (NSF) Computer and Information Science and Engineering Research Initiation Initiative (CRII) award in 2022 for her current research focusing on robust machine learning and time-series analysis. She won several research Awards at IEEE conferences and the Engineering community. The most recent was the University of Texas Rio Grande Valley Faculty Excellence Award in 2023. Her research interests are in robust and trustworthy modern machine learning solutions for real-world applications, including healthcare, cybersecurity, and surveillance systems. She publishes in the areas of trustworthy, reliable, and expandable machine learning, signal and image processing and optimization.
Intended Audience:
Undergraduates, graduates, and experts. Those with interest in the topic.
To request an interpreter, please visit myaccess.rit.edu
Event Snapshot
When and Where
Who
Open to the Public
Interpreter Requested?
No