Bachelor's thesis presentation. Ren is advised by Richard Milbradt and Prof. Dr. Christian Mendl.
Previous talks at the SCCS Colloquium
Ren Xinyu: Minimisation of the bath correlation sum expansion via gradient descent
SCCS Colloquium |
Simulating open quantum systems requires efficient representations of the bath correlation function (BCF), which encodes environmental memory effects. Traditional approximation methods often introduce unnecessary complexity and lack control over sparsity.
This thesis presents a gradient-based optimization framework for constructing compact multi-exponential expansions of the BCF. By combining differentiable optimization techniques with sparsity-promoting regularization, the method achieves high approximation quality with minimal computational overhead. The approach is applicable to general spectral densities and supports efficient simulation within the Hierarchical Equations of Motion (HEOM) formalism.