Abstract: Deep code models are vulnerable to adversarial attacks, making it possible for semantically identical inputs to trigger different responses. Current black-box attack methods typically ...
Writing clean, bug-free code is a point of pride for any developer. For decades, tools that measure code quality have been a staple of the software development lifecycle, helping teams eliminate bugs, ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results