Paper published in IEEE Access

The paper with the title "Smoothing Methods for Automatic Differentiation Across Conditional Branches", co-authored by Justin Kreikemeyer and Philipp Andelfinger, has been published in IEEE Access (https://doi.org/10.1109/ACCESS.2023.3342136). It presents and evaluates novel methods to allow the gradient-based optimization of imperative programs and simulations with discrete jumps, where automatic differentiation alone does not suffice.