Use optimization_experiment/opt_code/Matlab/main.m in Matlab to run the optimization experiments.
These are written in python. You can open them using Jupyter notebook. dim_gaussian.ipynb is used to pick the dimension of Gaussian sketching on a small synthetic data. excess_risk.ipynb was used to test our sketching&solve paradigm. The results will not be exactly the same as you see in the paper, since all the datasets are generated randomly.
You can download a compressed zip at here