JA
JA
Research projects
CV
Light
Dark
Automatic
diagonalization
Differentiable isospectral flows for matrix diagonalization
Implementing continuous isospectral flows in a differentiable manner to compute derivatives efficiently and accurately, enabling the solution of inverse problems.
JA in collaboration with Frank Schäfer, Heinz Krummenacher, Stefan Kehrein, and Christoph Bruder
Short talk at JuliaCon (2023)