Matt O'Shaughnessy

About me

I am a Ph.D. student in the School of Electrical and Computer Engineering and Center for Machine Learning at Georgia Tech. I'm very fortunate to be advised by Mark Davenport and Chris Rozell, and to be supported by the NDSEG Fellowship.

My research spans statistics, signal processing, and machine learning; broadly speaking, it involves developing methods for statistical and causal inference that exploit low-dimensional and temporal structure. My past and current projects involve (1) Bayesian methods for tracking sparse signals; (2) the use of low-dimensional models in theory and algorithms for causal inference; and (3) applications to public policy and economics.

I am also interested in public policy — both the statistical analysis of policy and policy for science and technology. I am a current fellow in the Sam Nunn Security Program, where my research involves societal impacts of machine learning.

In Fall 2019, Greg Canal and I organized a petition regarding mandatory graduate student fees at Georgia Tech, which are the highest in the country.

Outside of research, I enjoy hiking and cycling, music (especially classical), reading, and baking. I have an only slightly tenuous claim to an Erdös-Bacon-Sabbath number of nine: my Erdös number is 4, my Bacon number (from a role as an extra in a major movie) is 2, and my Sabbath number (from my time as a violist) is 3.

Publications

Journal papers

M. O'Shaughnessy, M. Davenport, and C. Rozell, "Sparse Bayesian learning with dynamic filtering for inference of time-varying sparse signals," IEEE Transactions on Signal Processing, December 2019. [ungated]

Conference papers

M. O'Shaughnessy, M. Davenport, and C. Rozell, "Dynamical system implementations of sparse Bayesian learning," to appear in Proc. Int. Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), Dec 2019.
G. Canal*, M. O'Shaughnessy* (equal contribution), C. Rozell, and M. Davenport, "Joint estimation of trajectory and dynamics from paired comparisons," to appear in Proc. Int. Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), Dec 2019.
M. O'Shaughnessy, M. Davenport, and C. Rozell, "Robust incorporation of signal predictions into the sparse Bayesian learning framework," in Proc. Workshop on Signal Processing with Adaptive Sparse Structured Representations (SPARS), July 2019.
M. O'Shaughnessy and M. Davenport, "Localizing users and items from paired comparisons," in Proc. Int. Workshop on Machine Learning for Signal Processing (MLSP), September 2016.
R. Ortman, D. Carr, R. James, D. Long, M. O'Shaughnessy, C. Valenta, and G. Tuell, "Real-time, mixed-mode computing architecture for waveform-resolved lidar systems with total propagated uncertainty," in Proc. Laser Radar Technology and Applications XXI, May 2016.

Other

Work Experience

Intern, MIT Lincoln Laboratory (Open and Embedded Systems Group), Summer 2016.
Co-op, Georgia Tech Research Institute (Remote Sensing Group), 2014—2016 (three semesters full-time; two semesters part-time).
Intern, Boeing Satellite Systems (DSP Algorithms Group), Summer 2015.

Teaching

Undergraduate Research Mentor, Fall 2019 — Spring 2020.
I mentor three undergraduate students working on research involving ordinal comparisons and generative models.
Teaching Assistant, Fall 2013 — Spring 2016.
I was an undergraduate teaching assistant for CS 1371, Georgia Tech's introductory computing class for engineers. I taught a weekly recitation section and led a team developing software tools for students.

Based on theme by orderedlist