PNNL

Brenda Praggastis graduated from the University of Washington in 1991 with a PhD in mathematics. Her research area was symbolic dynamics, which studied the invariant properties of dynamical systems by representing them as Shifts of Finite type. Her postdoc was at the Mathematical Sciences Research Institute in Berkeley, CA. A survey of her research was published in the Transactions of the American Mathematical Society in 1996. After teaching for a couple of years she took a long break to raise and homeschool her four children. She returned to teaching in 2010 and to research in 2015 when she joined PNNL as a data scientist.

Shortly after coming to PNNL she joined the Topological Data Analysis team. Her research focused on local homology, simplicial homology, and sheaf theory. She led the development and the 2019 release of PNNL's open-source library, HyperNetX (HNX), a Python library for the analysis and visualization of data modeled as hypergraphs. 

In 2017 she bootstrapped herself into machine learning research, acting as PI in a small project under the Deep Science initiative, and participating in the mathematical analysis of deep neural nets under Deep Skies. In 2020 she led the development and release of a second open-source library, DeepDataProfiler (DDP), a Python library that models the decision process of a model data pairing as graph-like structures.

She is PI for the FY21 MARS initiative project, Deep Data Profiler: A Platform and Methodology for the Analysis and Interpretation of Neural Networks, which will apply computational topology to DDP generated representations of model data pairings.

Her current interests are divided between the analysis and modeling of hypergraphs and the mathematics of deep neural networks.

• University of Washington, Doctor of Philosophy, mathematics
• SUNY at Stonybrook, Master of Arts, mathematics
• SUNY at Stonybrook, Bachelor of Science, mathematics