MainRunModel.py: code of main run model for mnist, start from here.

RunArtificial.py: code of main run model for synthetic, start from here.

./Models: code of robust bounded aggregation rules

Attacks.py: Different Byzantine attacks, include gaussian attacks, sign-flipping attacks, sample-duplicating attacks

Config.py: Configurations of these rules under Byzantine attack for mnist experiment. All hyper parameters like learning rate and penalty parameter can be tuned here

Config0.py: Configurations of these rules without attack. All hyper parameters like learning rate and penalty parameter can be tuned here

Config_artificial.py: Configurations of these rules under Byzantine attack for synthetic experiment. All hyper parameters like learning rate and penalty parameter can be tuned here

Config0_artificial.py: Configurations of these rules without attack for synthetic experiment. All hyper parameters like learning rate and penalty parameter can be tuned here

draw/draw_all.py: Plot the curve of mnist experiment results

draw/draw_artificial.py: Plot the curve of synthetic experiment results

LoadMnist.py: Load MNIST dataset

FatherModel.py: Solver of softmax regression, includes the calculation functions of loss, regret, variance and accuracy.