Suggestions
Nidhal Carla Bouaynaya
Associate Dean, Henry M. Rowan College of Engineering, and Professor of Electrical and Computer Engineering
Nidhal Carla Bouaynaya is a distinguished Professor of Electrical and Computer Engineering specializing in statistical learning and data science.
She serves as the Director of the Machine & Artificial Intelligence and Virtual Reality (MAVRC) Center at Rowan University, where she leads the research and graduate studies initiatives as the Associate Dean for Research and Graduate Studies.
Additionally, Nidhal Carla Bouaynaya is the Rowan faculty Lead for the NSF I-Corps Hub Northeast Region and is actively involved in fostering innovation and entrepreneurship in the region.
With a strong entrepreneurial spirit, she is the Co-Founder and Chief Technology Officer (CTO) of MRIMath LLC, a forward-thinking startup focused on artificial intelligence applications in precision medicine, and the Founder of Deep AI Solutions, LLC.
Educationally, Nidhal Carla Bouaynaya holds a Doctor of Philosophy (PhD) in Electrical and Computer Engineering, a Master of Science (MS) in Pure Mathematics, and another MS in Electrical and Computer Engineering, showcasing her diverse academic background.
Her academic journey includes studies at esteemed institutions such as the University of Illinois Chicago, Illinois Institute of Technology, and Ecole nationale supérieure de l'Electronique et de ses Applications.
Throughout her career, she has held key positions including being the Associate Dean for Research and Graduate Studies, Professor of Electrical and Computer Engineering, and Associate Professor at Rowan University, as well as an Assistant Professor at the University of Arkansas at Little Rock, demonstrating her commitment to academic excellence and research.
Highlights
Our team approach to conferences makes them not only professional events but also family affairs. Here, at the World Congress on Computational Intelligence in Japan. Proud of Christopher Angelini - a PhD student and also an ML engineer at iWorks Corp. working directly under the Vice President for Emerging Technologies - who presented his basic research on continual learning to mitigate the issue of catastrophic forgetting in neural networks. Catastrophic forgetting in neural networks is like a chef learning a new recipe and, in the process, forgetting how to cook their favorite dish. Most solutions in the field present an inherent tradeoff between not forgetting (stability) and learning new tasks (plasticity), which leads - in my opinion - to a mediocre solution: neither very good at the new tasks nor completely overcoming the forgetting. In the chef example, this means the chef ends up being not very good at the new recipe and also forgets important parts of the old one. Our work with Chris, however, takes an entirely different approach. We leverage our previous NSF-funded work on uncertainty quantification to cleverly navigate the loss landscape during training; the network explores the space more when uncertainty is high and tends to converge when its uncertainty is low. This intuitive idea required substantial mathematical development but beautifully led to this initial framework of uncertainty-guided continual learning.
