Robust Models Are More Interpretable Because Attributions Look Normal.

Zifan Wang Matt Fredrikson Anupam Datta

Published in: ICML (2022)

Keyphrases

data sets
statistical model
probabilistic model
machine learning algorithms
neural network model
database
data mining
learning algorithm
evolutionary algorithm
prior knowledge
maximum likelihood
model selection
regression model
process model
partial occlusion
parameter tuning