Overview
Broadly, my research interests are in computational and mathematical science. I have worked in areas such as non-convex optimization, deep learning, data compression, and algorithmic economics. I appreciate both the theory and application side of problems, which is to say I enjoy both proving a theorem and programming a software package.
Papers & Publications
-
Regularized Saddle-Free Newton: Saddle Avoidance and Efficient Implementation
- Cooper Simpson and Stephen Becker
- (2023).
- Pre-Print
-
QuadConv: Quadrature-Based Convolutions with Applications to Non-Uniform PDE Data Compression
- Kevin Doherty, Cooper Simpson, et al.
- (2023).
- Journal Publication
-
Generalizing the Proportional Response Dynamic for Exchange Economies
- Cooper Simpson
- (2022).
- Pre-Print
-
Regularized Saddle-Free Newton: Saddle Avoidance and Efficient Implementation
- Cooper Simpson
- (2022).
- Master's Thesis
-
Embedded Neural Networks for Robot Autonomy
- S. Aguasvivas, D. Hughes, C. Simpson, et al.
- (2022).
- Conference Paper
Software
Saddle-Free Newton
Julia implementation of the regularized saddle-free newton method for unconstrained non-convex optimization.
GitHubRandNLA
Julia algorithms for randomized numerical linear algebra.
GitHubQuadrature Convolutions (QuadConv)
A quadrature-based discrete convolution operator for use in deep learning tasks with unstructured data.
GitHubNeural Networks for Microcontrollers
Python and C++ packages for translating trained neural networks into C code for use in embedded systems.
GitHub