Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Where can the user get the adversarial example when ab-crown return unsafe result? #25

Closed
CRhapsody opened this issue Mar 27, 2023 · 3 comments

Comments

@CRhapsody
Copy link

When I test a robustness property using this tool, the result of program return 'unsafe', but I can not find the corresponding adversarial example.

@huanzhang12
Copy link
Member

Counterexamples can be saved via setting "general"->"save_adv_example" to true.

save_adv_example: false # Save returned adversarial example in file.

(the above file is a documentation showing where you should change in your own config file. Do not change abcrown_all_params.yaml directly)

However, there are some cases the counterexamples were not saved due to an implementation glitch. In these cases, you can set a breakpoint when counterexample is found and dump it manually. We will fix it in the next release.

@kamdoumloich
Copy link
Contributor

I tested it, but the output file (out.txt) that I received is a binary file. How can I read its content?

@shizhouxing
Copy link
Member

Hi @kamdoumloich , the counterexample is saved to a file according to --cex_path you specify. By default, it's ./test_cex.txt:

https://github.com/Verified-Intelligence/alpha-beta-CROWN/blob/main/complete_verifier/arguments.py#L740

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants