Preliminary code for the paper Learning Deterministic Surrogates for Robust Convex QCQPs.
We introduce a double implicit layer strategy for obtaining robust decision losses for convex quadratically constrained quadratic programs (QCQPs) with uncertain constraints.
The experiments were done in a spartan fashion in Jupyter notebooks. This is likely to be improved in the future.