search

vtjeng / MIPVerify.jl

Evaluating Robustness of Neural Networks with Mixed Integer Programming
MIT License
113 stars 31 forks source link

Bugfix: Guarantee that `find_adversarial_example` output is in target category. #152

Closed vtjeng closed 8 months ago

vtjeng commented 8 months ago

This change is inspired by #34.

Prior to the PR, it was possible for the find_adversarial_example to find an "adversarial example" that was not, in fact, classified in the target_selection. This is contrary to what the documentation promised.

This PR fixes that and also makes some incidental formatting changes.

codecov[bot] commented 8 months ago

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (80559b1) 91.42% compared to head (79e0554) 91.30%.

Files Patch % Lines
src/MIPVerify.jl 0.00% 1 Missing :warning:
Additional details and impacted files ```diff @@ Coverage Diff @@ ## master #152 +/- ## ========================================== - Coverage 91.42% 91.30% -0.12% ========================================== Files 25 25 Lines 816 817 +1 ========================================== Hits 746 746 - Misses 70 71 +1 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.