These labels were automatically added by AI and may be inaccurate. For details, see About Literature Database.
Abstract
With tools like GitHub Copilot, automatic code suggestion is no longer a
dream in software engineering. These tools, based on large language models, are
typically trained on massive corpora of code mined from unvetted public
sources. As a result, these models are susceptible to data poisoning attacks
where an adversary manipulates the model's training by injecting malicious
data. Poisoning attacks could be designed to influence the model's suggestions
at run time for chosen contexts, such as inducing the model into suggesting
insecure code payloads. To achieve this, prior attacks explicitly inject the
insecure code payload into the training data, making the poison data detectable
by static analysis tools that can remove such malicious data from the training
set. In this work, we demonstrate two novel attacks, COVERT and TROJANPUZZLE,
that can bypass static analysis by planting malicious poison data in
out-of-context regions such as docstrings. Our most novel attack, TROJANPUZZLE,
goes one step further in generating less suspicious poison data by never
explicitly including certain (suspicious) parts of the payload in the poison
data, while still inducing a model that suggests the entire payload when
completing code (i.e., outside docstrings). This makes TROJANPUZZLE robust
against signature-based dataset-cleansing methods that can filter out
suspicious sequences from the training data. Our evaluation against models of
two sizes demonstrates that both COVERT and TROJANPUZZLE have significant
implications for practitioners when selecting code used to train or tune
code-suggestion models.
External Datasets
Python code files sourced from 18,310 public repositories on GitHub