Sasha Petrenko, a doctoral candidate in computer engineering, will defend their dissertation titled “Lifelong Machine Learning with Adaptive Resonance Theory.” Their advisor, Dr. Donald C. Wunsch II, is a professor in the electrical and computer engineering department. The dissertation abstract is provided below.

Lifelong learning (L2) is a challenging machine learning paradigm that both encompasses and formalizes the fields of continual learning and incremental learning.
The field is concerned with the mitigation of the phenomenon of catastrophic forgetting whereby learning agents that are faced with incrementally novel information deleteriously overwrite previous knowledge if that learning process is not regularized to counteract this consequence.
ART algorithms solve the this stability-plasticity dilemma by optimally assigning learning to categories or instantiating new knowledge when information is sufficiently novel.
While the appeal of deep neural networks is their capacity to learning useful feature manifolds simultaneously with the task at hand, they are especially subject to catastrophic forgetting both by their hierarchical architectures and by the current techniques used to train them.

The publications of this dissertation explore techniques for adapting ART algorithms from novel application domains in computer vision and biomedical data analysis to formulations of deep learning networks that complement and combine the unique strengths of adaptive resonance and deep learning to tackle the lifelong machine learning problem.