Department of Mathematics and Statistics

Department of Mathematics and Statistics
Department of Mathematics and Statistics
Subscribe to RSS - Topological Data Analysis

Interdisciplinary Topological Data Analysis Learning Seminar

Topological data analysis (TDA) is a class of topological techniques for understanding large, noisy, possibly incomplete data. The techniques have a particular focus on understanding the large-scale coordinate-independent "shape" of the data. While perhaps the most famous application of TDA is the discovery of a new type of breast cancer from an old dataset, the techniques are finding use in everything from medical imaging to progression analysis of disease, signal analysis, viral evolution, complex networks, AI, & much more (wikipedia has a more extensive list of applications). The aim of this seminar is to build a common language for understanding TDA & then hopefully to shift to a focus on applications. The tentative structure involves learning and sharing knowledge on dimensionality reduction, clustering, persistent homology, probabilistic analysis and homology inference in a format that includes different seminar members sharing knowledge through prepared presentations, coupled with group discussion.

Topological Data Analysis - Jeffrey Gauthier (Swarthmore)

Monday, March 9th, 2020

Time: 2:30-4:00 p.m. Place: Goodes Hall 120

Speaker: Jeffrey Gauthier (Swarthmore)

Title:  Are we there yet? How hippocampus neurons help us navigate the world.

Jeffrey Gauthier is an Assistant Professor of Biology at Swarthmore. Before thathe was most recently a postdoctoral research associate at the Princeton Neuroscience Institute, examining in vivo measurements of neural activity in awake mice navigating a virtual environment. He also did postdoctoral research at the Salk Institute for Biological Studies, including multielectrode recordings in primate retina to see how color opponency arises from the sampling of individual cones.

Topological Data Analysis - Catherine Pfaff (Queen's University)

Monday, March 2nd, 2020

Time: 2:30-4:00 p.m. Place: Goodes Hall 120

Speaker: Catherine Pfaff (Queen's University)

Topics:  Past Presentations & Directions Forward.
We've had a series of recent presentations on applications that have led to really great discussions (often cut short by time). This week, I'll lead a recollection of our thoughts on these applications (& directions forward), record them, & continue discussion on further directions forward (possibly to be explored by undergrad & grad students over the summer & next year).

All are welcome!

Topological Data Analysis - David Riegert & Troy Zeier

Monday, February 24th, 2020

Time: 2:30-4:00 p.m. Place: Goodes Hall 120

Speaker: David Riegert & Troy Zeier (Queen's University)

Title: A View from the Pitcher's Mound: The Statistics of Persistence Landscapes

Topics:  We (re)introduce a topological summary for data called the "persistence landscape" and discuss how this summary can be viewed as a random variable. Next, we provide a brief review of hypothesis testing and demonstrate how to apply standard statistical tests to persistence landscapes from the safety of simulations. Finally, we take these methods for a spin; applying them to Major League Baseball pitchers using a data set with measurements for ~3 million pitches across 2015-18.

All are welcome!

Topological Data Analysis - Multiple Speakers

Monday, November 18th, 2019

Time: 2:30 p.m. Place: Goodes Hall 120

Speaker: Jordan Kokocinski, Arne Kuhrs, Catherine Pfaff, David Riegert Luke Steverango

Topics:  The graduate students have a well-prepared presentation on the Bubenik worksheet. This is highly recommended, even if you've missed a few meetings &/or are confused about homology. With leftover time Pfaff will provide some supplement to last week's presentation.

All are welcome!