Star on GitHubStar3.2k stars

2026 · Preprint

VulDec: Vulnerability Detection in Binaries via Neural Decompilation

VulDec pairs real-world assembly and C code with CWE labels and uses neural decompilation to detect binary vulnerabilities. It evaluates code recovery across architectures and detection on a realistic test set.

Scope and limitations

The evaluation covers multiple architectures and an in-distribution test set with a 3% vulnerability rate.

Why it fits this collection

Neural decompilation recovers source-level code from binaries, and the dataset supports that reverse-engineering task.

Cite

@misc{paper-vuldec,
  title = {{VulDec: Vulnerability Detection in Binaries via Neural Decompilation}},
  author = {{Francesco Crescenzo Grasso} and {Pietro Liguori} and {Roberto Natella} and {Domenico Cotroneo}},
  year = {2026},
  url = {https://arxiv.org/abs/2610.33716},
  howpublished = {arXiv:2610.33716},
  note = {Preprint},
  eprint = {2610.33716},
  archivePrefix = {arXiv},
  month = oct
}

Read in a terminal: hf papers read 2610.33716

Publication and review details

Venue
arXiv
Status
Preprint
Publication / posting date
October 24, 2026
Date source
First arXiv submission date.
First arXiv submission
October 24, 2026
arXiv
2610.33716
Reviewed arXiv version
v1

The date and venue come from the arXiv record.

Sources reviewed

Reviewed against the paper and its primary sources. Cite the original authors using the linked paper or BibTeX.

Report an outdated entry Suggest a translation fix

Primary sources

  1. https://arxiv.org/abs/2610.33716
↑↓ move↵ open⌘↵ new tab