Importance-Driven Deep Learning System Testing -ICSE 2020
About
This paper
presents DeepImportance, a systematic testing methodology accompanied by an Importance-Driven (IDC)
test adequacy criterion for DL systems. Applying IDC enables to
establish a layer-wise functional understanding of the importance
of DL system components and use this information to guide the
generation of semantically-diverse test sets. Our empirical evalua-
tion on several DL systems, across multiple DL datasets and with
state-of-the-art adversarial generation techniques demonstrates the
usefulness and effectiveness of DeepImportance and its ability to
guide the engineering of more robust DL systems.
Repository
This repository includes details about the artifact corresponding to implementation of DeepImportance.
Our implementation is publicly available in
DeepImportance repository.
This artifact allows reproducing the experimental results presented in the paper. Below we
describe how to reproduce results. Before going further, first, check
installation page (i.e. INSTALL.md).
