---
title: "SEMPRE: Semantic-Aware Protocol Reverse Engineering via Consistency-Driven Joint Optimization"
url: https://awesomerea.org/papers/paper-sempre/
updated: 2026-11-10
type: paper
---

# SEMPRE: Semantic-Aware Protocol Reverse Engineering via Consistency-Driven Joint Optimization

> SEMPRE uses dynamic analysis and LLM reasoning to recover hierarchical message fields, infer their types and roles, and refine field boundaries. It evaluates the method on mainstream protocols and a botnet case study.

- Authors: Yihua Xu; Weijie Huang; Xiaofeng Wang; Hang Zhang; Binghui Wang; Yue Duan
- Venue: arXiv
- Status: Preprint
- Publication / posting date: October 22, 2026
- Date source: First arXiv submission date.
- First arXiv submission: October 22, 2026
- arXiv: 2610.31844
- Read in a terminal: `hf papers read 2610.31844`
- Reviewed arXiv version: v1
- Sources reviewed: 2026-11-09
- Added to the catalog: 2026-11-10
- Original paper: <https://arxiv.org/abs/2610.31844>
- PDF: <https://arxiv.org/pdf/2610.31844>
- BibTeX: <https://awesomerea.org/papers/citations/paper-sempre.bib>
- Catalog page: <https://awesomerea.org/papers/paper-sempre/>
- Record JSON: <https://awesomerea.org/papers/paper-sempre/record.json>
- Chinese version: [简体中文](https://awesomerea.org/zh/papers/paper-sempre/index.md)

## Scope and limitations

The evaluation covers mainstream protocols, comparisons with six baselines, component ablations, and a botnet case study.

## Why it fits this collection

Inferring undocumented protocol message layouts and field meanings is protocol reverse engineering.

## Publication and review notes

The date and venue come from the arXiv record.

## BibTeX

```bibtex
@misc{paper-sempre,
  title = {{SEMPRE: Semantic-Aware Protocol Reverse Engineering via Consistency-Driven Joint Optimization}},
  author = {{Yihua Xu} and {Weijie Huang} and {Xiaofeng Wang} and {Hang Zhang} and {Binghui Wang} and {Yue Duan}},
  year = {2026},
  url = {https://arxiv.org/abs/2610.31844},
  howpublished = {arXiv:2610.31844},
  note = {Preprint},
  eprint = {2610.31844},
  archivePrefix = {arXiv},
  month = oct
}
```

## Primary sources

1. <https://arxiv.org/abs/2610.31844>

## 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.
