Data Crimes and Dynamic Imaging: Frontiers in Medical AI
Efrat Shimron, PhD, Department of Electrical Engineering and Computer Sciences, UC Berkeley-
The University of Texas at Austin
Gates Dell Complex (GDC 6.302)
Abstract: Although Machine learning (ML) algorithms have recently made a huge impact on medical imaging, their development and deployment for clinical applications must be conducted carefully. This seminar will introduce both potential pitfalls and novelty offered by ML for medical imaging. The first part will introduce bias and data crimes that stem from naïve training using public databases. Although open-access databases are an important resource in the current ML era, they are sometimes used “off label”: data published for one task are used for training algorithms for a different one. I will show how this common practice leads to biased, overly optimistic results of well-known inverse problem solvers. Moreover, this could lead to algorithmic failure for clinical real-world data. While this phenomenon is general, examples will focus on algorithms developed for magnetic resonance imaging (MRI) reconstruction. In the second part I will focus on areas where ML can be impactful and introduce a novel technique for dynamic MRI, named BladeNet. This technique addresses some of the current barriers in abdominal imaging: motion-blurring and limited resolution due to long scan times. BladeNet offers built-in motion correction, rapid imaging, and reconstruction of videos with high spatio-temporal resolution, which are highly beneficial for clinical applications.
Speaker Bio: Efrat Shimron completed her PhD at the Technion – Israel Institute of Technology in the field of computational imaging. Since 2020 she is a postdoc at the Department of Electrical Engineering and Computer Sciences (EECS) at UC Berkeley, working with Prof. Michael (Miki) Lustig. Her research interests are in developing machine learning techniques for rapid and robust medical imaging, focusing on Magnetic Resonance Imaging (MRI). She is the recipient of several national career awards, including the 2023 EECS Rising Stars award, and many other international excellence awards. Her work on identifying “data crimes” in medical AI has recently received wide media coverage.Event Registration