Lulu Ge receives IEM doctoral fellowship
Doctoral candidate Lulu Ge has been awarded a fellowship by the Institute for Engineering in Medicine (IEM) for her work applying hyperdimensional computing for the diagnosis and treatment of neurological disorders. Titled Engineering in Medicine doctoral fellowship, it supports students whose work seeks to transform an area of medicine through innovative research, and entails collaboration with health or biological sciences and engineering or physical sciences faculty members. Ge earned her bachelor of science degree from Nanjing University of Posts and Telecommunications, and her master's degree from Southeast University, both in China. She is working on her dissertation under the guidance of Keshab Parhi, Distinguished McKnight University Professor and Erwin A. Kelen Chair in Electrical Engineering. We heard from Ge on her doctoral research, her motivation to pursue the particular line of research, and the impact of the fellowship.
Tell us about your research interests
My research interests include but are not limited to molecular computing, machine learning, hyperdimensional computing (HDC), and neuroscience. My main research interest is HDC. As a brain-inspired computing paradigm, HDC has its unique data representation, transformation, and interpretation as compared to traditional machine learning and deep learning. HDC continues to draw significant attention due to its comparable performance with traditional machine learning techniques, high noise immunity, one- or few-shot learning ability, high energy efficiency, and wide applications. Given its promising potential, and also driven by my interest in neuroscience, my dissertation research attempts to explore, theoretically and practically, the ability of HDC in classifying and clustering problems for two neurological disorders: epilepsy (my ongoing project) and major depressive disorder (MDD) (the IEM project).
What has motivated you to pursue this particular line of research?
HDC is still in its infancy, although current research has revealed its promising potential in addressing various cognitive tasks indicative of wide applications. However, several major questions still remain unanswered. For example, what are the optimal encoding methods in HDC for an arbitrary interest-driven task? What are the theoretical foundations to explain why HDC works and what limits it? What are the killer applications for HDC against traditional machine learning techniques? Inspired by these questions, I am interested in exploring the application of HDC in healthcare and neuroscience in particular.
How will the IEM fellowship support your work?
This IEM fellowship funds me for the whole year’s research work. Without this fellowship, if I were supported as a teaching assistant, I wouldn't have been able to devote as many hours to my research. Additionally, this fellowship engages me in challenging and interesting work. It also gives me invaluable access to professional networks.
What is your vision of where your dissertation research might lead you eventually?
After validating our experimental results, I plan to extend the transcranial magnetic stimulation (TMS) treatment for MDD to new experimental trials on patients with other mental disorders to further advance the field of neuromodulation. In terms of HDC, if the proposed research is successful, it will be the first time, to the best of my knowledge, that HDC is applied to the neuromodulation field. This enriches the application of hyperdimensional computing.