# Decaf: Improving Neural Decompilation with Automatic Feedback and Search

> Decaf samples multiple candidate decompilations, recompiles them and ranks them using compiler-derived byte distance or a learned reranker. The work studies candidate selection as a way to improve neural decompilation without an iterative source-repair conversation.

- Authors: Alexander Shypula; Osbert Bastani; Edward Schwartz
- Venue: arXiv
- Status: Preprint
- Publication / posting date: May 12, 2026
- Date source: First arXiv submission date; no separate proceedings or journal publication was verified.
- First arXiv submission: May 12, 2026
- arXiv: 2605.11501
- Reviewed arXiv version: v1
- Sources reviewed: 2026-10-20
- Added to the catalog: 2026-10-20
- Original paper: <https://arxiv.org/abs/2605.11501>
- PDF: <https://arxiv.org/pdf/2605.11501v1>
- Code: <https://github.com/AlexShypula/decaf>
- Project: <https://huggingface.co/decaf-usenix>
- BibTeX: <https://awesomerea.org/papers/citations/paper-decaf.bib>
- Catalog page: <https://awesomerea.org/papers/paper-decaf/>
- Record JSON: <https://awesomerea.org/papers/paper-decaf/record.json>
- Chinese version: [简体中文](https://awesomerea.org/zh/papers/paper-decaf/index.md)

## Scope and limitations

Byte matching assumes access to the reference compiler/configuration. Experiments subsample ExeBench and exclude examples Hex-Rays cannot decompile; stripped Synth is not execution-scored because call targets cannot be linked. Supplied tests can miss errors. Alternate architectures and obfuscated code are not evaluated, and Juliet results need not generalize to complex vulnerabilities.

## Why it fits this collection

Provides a compiler-feedback and candidate-ranking approach with clear distinctions between matching and test-based evidence.

## Publication and review notes

This May 2026 preprint is within the six-month fallback window and is explicitly under review. The pinned code root has no license file; inspected generator/test-set metadata exposes no license declaration.

## Primary sources

1. <https://arxiv.org/abs/2605.11501v1>
2. <https://arxiv.org/html/2605.11501v1>
3. <https://github.com/AlexShypula/decaf/blob/f33c8f4587b7a0055513d6396194015114de8e35/README.md>
4. <https://api.github.com/repos/AlexShypula/decaf/git/trees/0d6690fea728fae5d622afa771a7f6b7dd51df44>
5. <https://huggingface.co/api/models/decaf-usenix/Decaf-Gen-22b>
6. <https://huggingface.co/api/datasets/decaf-usenix/Decaf-Test-Sets>

## About this entry

Reviewed against the paper and its primary sources. Cite the original authors using the linked paper or BibTeX.
Original editorial content: MIT. Linked projects, papers and assets retain their own rights and licenses.