LLM4CodeRE: Generative AI for Code Decompilation Analysis and Reverse Engineering
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911574022684672 |
|---|---|
| author | Jelodar, Hamed Bai, Samita Nwankwo, Tochukwu Emmanuel Hamedi, Parisa Meymani, Mohammad Razavi-Far, Roozbeh Ghorbani, Ali A. |
| author_facet | Jelodar, Hamed Bai, Samita Nwankwo, Tochukwu Emmanuel Hamedi, Parisa Meymani, Mohammad Razavi-Far, Roozbeh Ghorbani, Ali A. |
| contents | Code decompilation analysis is a fundamental yet challenging task in malware reverse engineering, particularly due to the pervasive use of sophisticated obfuscation techniques. Although recent large language models (LLMs) have shown promise in translating low-level representations into high-level source code, most existing approaches rely on generic code pretraining and lack adaptation to malicious software. We propose LLM4CodeRE, a domain-adaptive LLM framework for bidirectional code reverse engineering that supports both assembly-to-source decompilation and source-to-assembly translation within a unified model. To enable effective task adaptation, we introduce two complementary fine-tuning strategies: (i) a Multi-Adapter approach for task-specific syntactic and semantic alignment, and (ii) a Seq2Seq Unified approach using task-conditioned prefixes to enforce end-to-end generation constraints. Experimental results demonstrate that LLM4CodeRE outperforms existing decompilation tools and general-purpose code models, achieving robust bidirectional generalization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_06095 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | LLM4CodeRE: Generative AI for Code Decompilation Analysis and Reverse Engineering Jelodar, Hamed Bai, Samita Nwankwo, Tochukwu Emmanuel Hamedi, Parisa Meymani, Mohammad Razavi-Far, Roozbeh Ghorbani, Ali A. Cryptography and Security Artificial Intelligence Code decompilation analysis is a fundamental yet challenging task in malware reverse engineering, particularly due to the pervasive use of sophisticated obfuscation techniques. Although recent large language models (LLMs) have shown promise in translating low-level representations into high-level source code, most existing approaches rely on generic code pretraining and lack adaptation to malicious software. We propose LLM4CodeRE, a domain-adaptive LLM framework for bidirectional code reverse engineering that supports both assembly-to-source decompilation and source-to-assembly translation within a unified model. To enable effective task adaptation, we introduce two complementary fine-tuning strategies: (i) a Multi-Adapter approach for task-specific syntactic and semantic alignment, and (ii) a Seq2Seq Unified approach using task-conditioned prefixes to enforce end-to-end generation constraints. Experimental results demonstrate that LLM4CodeRE outperforms existing decompilation tools and general-purpose code models, achieving robust bidirectional generalization. |
| title | LLM4CodeRE: Generative AI for Code Decompilation Analysis and Reverse Engineering |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2604.06095 |