Awesome AI for Software Reverse Engineering

A curated collection of learning-based reverse engineering research, focusing on recovering and understanding low-level code using machine learning, deep learning, and large language models (LLMs).

🎯 Overview

Reverse engineering (RE) has entered a new phase. While traditional RE relies heavily on handcrafted heuristics and symbolic analysis, recent advances in learning-based methods have fundamentally changed how we recover and understand low-level code.

This repository tracks this paradigm shift — from early machine learning approaches, to deep neural models, and most recently, large language models (LLMs) — highlighting how learning-based techniques enable scalable, semantic, and cross-platform reverse engineering.

Core Focus: Learning-based Code Understanding & Recovery

This repository focuses on research whose primary goal is to recover or enhance human understanding of low-level or opaque code representations using learning-based techniques.

Included topics:

Structure Recovery, e.g., function boundary identification, control flow graph recovery, and modular decomposition
Semantic Recovery, e.g., type inference, variable naming, function naming and summarization
Code Reconstruction, e.g., neural decompilation, decompiler output refinement

Supported representations:

Native binaries (x86, ARM, MIPS)
Bytecode (JVM, EVM, .NET, Android/Dalvik)
WebAssembly (Wasm)

Excluded topics:

Analysis performed directly on high-level source code
Vulnerability detection, malware classification, or exploit generation unless program understanding is the primary contribution

📌 Table of Contents

📚 Top Conferences

S&P

(26, 25, 24, 23)

sp2026

[Decompilation] Oxidizer: Toward Concise and High-fidelity Rust Decompilation. [repo]

sp2025

[Data-Code Separation] You Can't Judge a Binary by Its Header: Data-Code Separation for Non-Standard ARM Binaries Using Pseudo Labels.
[Decompilation] PyLingual: Toward Perfect Decompilation of Evolving High-Level Languages. [repo].
[Composite Type Inference] TypeForge: Synthesizing and Selecting Best-Fit Composite Data Types for Stripped Binaries.

sp2024

[Variable Naming] Len or index or count, anything but v1": Predicting Variable Names in Decompilation Output with Transfer Learning. [repo].

sp2023

[Control Flow Recovery] Callee: Recovering Call Graphs for Binaries with Transfer and Contrastive Learning. [repo].
[Function Naming] XFL: Naming Functions in Binaries with Extreme Multi-label Learning. [repo].

CCS

(26, (25, 24, 22, 21)

ccs2026

[Function Naming] Hieronym: Leveraging Hierarchical Multi-Source Information for Function Renaming in Stripped Binary.

ccs2025

[Instruction Recovery, Function Recovery, Control Flow Recovery] Disa: Accurate Learning-based Static Disassembly with Attentions. [repo].

ccs2024

[Variable Naming, Type Inference] ReSym: Harnessing LLMs to Recover Variable and Data Structure Symbols from Stripped Binaries. [repo].

ccs2022

[Function Naming] SymLM: Predicting Function Names in Stripped Binaries via Context-Sensitive Execution-Aware Code Embeddings. [repo].

ccs2021

[Pretrained-model] PalmTree: Learning an Assembly Language Model for Instruction Embedding. [repo].

USENIX Security

(25, 24, 22)

sec2025

[Function Naming] BLens: contrastive captioning of binary functions using ensemble embedding. [zenodo].
[Instruction Recovery] Tady: A Neural Disassembler without Structural Constraint Violations. [repo].

sec2024

[Type Inference] TYGR: type inference on stripped binaries using graph neural networks. [repo].

sec2022

[Variable Naming, Type Inference] Augmenting Decompiler Output with Learned Variable Names and Types. [repo].
[Instruction Recovery, Function Recovery] DeepDi: Learning a Relational Graph Convolutional Network Model on Instructions for Fast and Accurate Disassembly. [repo].

NDSS

(26, 25, 24, 21)

ndss2026

[Code Refinement] FidelityGPT: Correcting Decompilation Distortions with Retrieval Augmented Generation. [repo].
[Decompilation, Composite type inference, Variable Naming] Idioms: Neural Decompilation With Joint Code and Type Definition Prediction. [repo].

ndss2025

[Variable Naming] Unleashing the Power of Generative Model in Recovering Variable Names from Stripped Binary. [repo].
[Function Naming] Beyond Classification: Inferring Function Names in Stripped Binaries via Domain Adapted LLMs. [repo].

ndss2024

[Variable Naming, Binary Summarization, Code Refinement] DeGPT: Optimizing Decompiler Output with LLM. [repo].

ndss2021

[Instruction Recovery, Function Recovery] XDA: Accurate, Robust Disassembly with Transfer Learning. [repo].

ICSE

(22)

icse2022

[Binary Structure Recovery] Modx: Binary Level Partial Imported Third-Party Library Detection through Program Modularization and Semantic Matching.

FSE

(25, 24, 23, 22, 21)

fse2025

[Variable Naming, Decompilation] DiSCo: Towards Decompiling EVM Bytecode to Source Code using Large Language Models. [online].
[Binary Summarization] MiSum: Multi-modality Heterogeneous Code Graph Learning for Multi-intent Binary Code Summarization. [repo].
[Type Inference] Recasting Type Hints from WebAssembly Contracts. [repo].

fse2024

[Function Naming] Enhancing Function Name Prediction using Votes-Based Name Tokenization and Multi-task Learning. [repo].

fse2023

[Type Inference] DeepInfer: Deep Type Inference from Smart Contract Bytecode. [repo].
[Function Recovery] FunProbe: Probing Functions from Binary Code through Probabilistic Analysis. [repo].

fse2022

[Memory Dependence Analysis] NeuDep: neural binary memory dependence analysis.

fse2021

[Type Inference] StateFormer: fine-grained type recovery from binaries using generative state modeling. [repo].

ASE

(26, 25, 24, 23)

ase2026

[Evaluation] Beyond Text Matching: Towards Reference-Free Evaluation for Human-Oriented Binary Reverse Engineering. [artifacts]

ase2025

[Binary Structure Recovery] BinStruct: Binary Structure Recovery Combining Static Analysis and Semantics. [online].
[Code Refinement] PseudoFix: Refactoring Distorted Structures in Decompiled C Pseudocode. [repo].

ase2024

[Type Inference] TypeFSL: Type Prediction from Binaries via Inter-procedural Data-flow Analysis and Few-shot Learning. [repo].
[Decompilation] WaDec: Decompiling WebAssembly Using Large Language Model. [repo].

ase2023

[Pretrained-model, Binary Summarization, Function Naming, Variable Naming, etc] HexT5: Unified Pre-Training for Stripped Binary Code Information Inference. [repo].

ISSTA

(25, 24, 21)

issta2025

[Code Refinement] DecLLM: LLM-Augmented Recompilable Decompilation for Enabling Programmatic Use of Decompiled Code. [online].

issta2024

[Pretrained-model, Binary Summarization] CLAP: Learning Transferable Binary Code Representations with Natural Language Supervision. [repo].
[Binary Structure Recovery] DeLink: Source File Information Recovery in Binaries.
[Type Inference, Binary Summarization] Multi-modal Learning for WebAssembly Reverse Engineering.

issta2021

[Function Naming] A lightweight framework for function name reassignment based on large-scale stripped binaries. [repo].

PLDI

(22)

pldi2022

[Type Inference] Finding the Dwarf: Recovering Precise Types from WebAssembly Binaries. [repo].

📜 Top Journals

TOSEM

(26, 25, 23)

tosem2026

[Type Inference, Variable Naming] HyRES: Recovering Data Structures in Binaries via Semantic Enhanced Hybrid Reasoning. [repo].

tosem2025

[Binary Summarization] ByteCue: enhancing bytecode comment generation with API information. [repo].
[Type Inference] Beyond the Edge of Function: Unraveling the Patterns of Type Recovery in Binary Code. [repo].
[Binary Summarization] Automating Comment Generation for Smart Contract from Bytecode. [figshare].
[Function Naming] llasm: Naming Functions in Binaries by Fusing Encoder-only and Decoder-only LLMs. [repo].

tosem2023

[Variable Naming] DIRE and its Data: Neural Decompiled Variable Renamings with Respect to Software Class.

TSE

(25, 22)

tse2025

[Function Recovery, Type Inference, Code Refinement] Augmenting Smart Contract Decompiler Output Through Fine-Grained Dependency Analysis and LLM-Facilitated Semantic Recovery. [code].

tse2022

[Inline Function Recovery] Learning to Find Usages of Library Functions in Optimized Binaries.

TIFS

(25)

tifs2025

[Binary Summarization] Malsight: Exploring Malicious Source Code and Benign Pseudocode for Iterative Binary Malware Summarization.

📌 Other Papers

ICLR

(26, 25)

iclr2026

[Function Naming, Variable Naming, Decompilation] SK2Decompile: LLM-based Two-Phase Binary Decompilation from Skeleton to Skin. [repo].

iclr2025

[Decompilation] Nova: Generative Language Models for Assembly Code with Hierarchical Attention and Contrastive Learning. [repo].

ICML

(24)

icml2024

[Decompilation, Variable Naming, Binary Summarization] StackSight: Unveiling WebAssembly through Large Language Models and Neurosymbolic Chain-of-Thought Decompilation.

EMNLP

(24, 23)

emnlp2024

[Decompilation] LLM4Decompile: Decompiling Binary Code with Large Language Models. [repo].

emnlp2023

[Binary Summarization] CP-BCS: binary code summarization guided by control flow graph and pseudo code. [repo].

ACSAC

(25, 22)

acsac2025

[Binary Summarization] AndroByte: LLM-Driven Privacy Analysis through Bytecode Summarization and Dynamic Dataflow Call Graph Generation. [repo].

acsac2022

[Decompilation] Boosting neural networks to decompile optimized binaries. [repo].

ASIACCS

(24, 23)

asiaccs2024

[Function Naming] BinAdapter: Leveraging Continual Learning for Inferring Function Symbol Names in a Binary. [repo].

asiaccs2023

[Function Naming] A Transformer-based Function Symbol Name Inference Model from an Assembly Language for Binary Reversing. [repo].

SANER

(23)

saner2023

[Binary Summarization] Extending Source Code Pre-Trained Language Models to Summarise Decompiled Binaries. [zenodo].

arXiv

(26, 25, 24, 23, 22)

arxiv2026

[Decompilation] Binary Decompilation LLM with Feedback-Driven Multi-Turn Refinement. [hugging face]
[Code Refinement] PCodeTrans: Translate Decompiled Pseudocode to Compilable and Executable Equivalent.
[Code Refinement] From Transactions to Exploits: Automated PoC Synthesis for Real-World DeFi Attacks.
[Control Flow Recovery] iResolveX: Multi-Layered Indirect Call Resolution via Static Reasoning and Learning-Augmented Refinement.

arxiv2025

[Function Naming] AGNOMIN -- Architecture Agnostic Multi-Label Function Name Prediction. [repo].
[Decompilation] Beyond C/C++: Probabilistic and LLM Methods for Next-Generation Software Reverse Engineering.
[Decompilation] The CodeInverter Suite: Control-Flow and Data-Mapping Augmented Binary Decompilation with LLMs. [repo].
[Control Flow Recovery] Resolving Indirect Calls in Binary Code via Cross-Reference Augmented Graph Neural Networks. [zenodo].
[Pretrained-model, Decompilation, Function Naming, Variable Naming, Type Inference, Binary Summarization] ReCopilot: Reverse Engineering Copilot in Binary Analysis. [online].
[Decompilation] ReF Decompile: Relabeling and Function Call Enhanced Decompile. [repo].
[Decompilation] SALT4Decompile: Inferring Source-level Abstract Logic Tree for LLM-Based Binary Decompilation. [repo].
[Decompilation] Context-Guided Decompilation: A Step Towards Re-executability.
[Decompilation] Decompiling Smart Contracts with a Large Language Model.

arxiv2024

[Instruction Recovery] Disassembling Obfuscated Executables with LLM.
[Type Inference, Variable Naming] STRIDE: Simple Type Recognition In Decompiled Executables.
[Function Naming, Variable Naming, Code Refinement] Exploring the Efficacy of Large Language Models (GPT-4) in Binary Reverse Engineering. [repo].

Other

[CNSSE-2026] From Heuristics to Transformers: A Comprehensive Survey of Type Inference from Stripped Binaries.
[BAR-2022] Beyond the C: Retargetable Decompilation using Neural Machine Translation. [repo].
[CGO-2024] SLaDe: A Portable Small Language Model Decompiler for Optimized Assembly. [zenodo].
[CODASPY-2026] Practical Type Inference: High-Throughput Recovery of Real-World Structures and Function Signatures.
[DTRP-2024] Asm2Seq: Explainable Assembly Code Functional Summary Generation for Reverse Engineering and Vulnerability Analysis. [repo].
[ESORICS-2023] AttnCall: Refining Indirect Call Targets in Binaries with Attention. [repo].
[ICSME-2024] Optimizing Decompiler Output by Eliminating Redundant Data Flow in Self-Recursive Inlining. [repo].
[ICSME-2024] How Far Have We Gone in Binary Code Understanding Using Large Language Models.
[IJCAI-2025] BinMetric: A Comprehensive Binary Code Analysis Benchmark for Large Language Models.
[ISC-2025] LLM-DAS: An LLM-Powered Deobfuscation System for ARM Binary Code.
[ISC-2024] Attention-Based Decompilation Through Neural Machine Translation.
[NLP4Prog-2021] DIRECT: A Transformer-based Model for Decompiled Variable Name Recovery. [repo].
[PST-2023] Investigating Neural-based Function Name Reassignment from the Perspective of Binary Code Representation. [repo].
[SmartSP-2024] RustBound: Function Boundary Detection over Rust Stripped Binaries. [code].
[TrustCom-2024] SBCM: Semantic-Driven Reverse Engineering Framework for Binary Code Modularization.
[WWW-2025] SuiGPT MAD: Move AI Decompiler to Improve Transparency and Auditability on Non-Open-Source Blockchain Smart Contract. [code].
[WWW-2023] Automated WebAssembly Function Purpose Identification With Semantics-Aware Analysis.

📊 Benchmarks、Study & SoKs

[CSUR-2026] A Survey of Symbol Name Recovery in Sotware Reverse Engineering.
[Arxiv-2026] SoK: AI-Augmented Binary Reversing.
[Github-2026] DecBench: An experimental living benchmark to measure how close we are to perfect decompilation on binaries.. [repo]
[NDSS-2026] Decompiling the Synergy: An Empirical Study of Human–LLM Teaming in Software Reverse Engineering.
[arXiv-2026] REBENCH: A Procedural, Fair-by-Construction Benchmark for LLMs on Stripped-Binary Types and Names. [zenodo].
[arXiv-2025] An Empirical Study on the Effectiveness of Large Language Models for Binary Code Understanding. [repo]
[arXiv-2025] BuildBench: Benchmarking LLM Agents on Compiling Real-World Open-Source Software.
[arXiv-2025] Decompile-Bench: Million-Scale Binary-Source Function Pairs for Real-World Binary Decompilation. [repo].
[arXiv-2025] Benchmarking Binary Type Inference Techniques in Decompilers. [repo].
[IJCAI-2025] BinMetric: A Comprehensive Binary Code Analysis Benchmark for Large Language Models.
[ASE-2025] R3-Bench: Reproducible Real-world Reverse Engineering Dataset for Symbol Recovery. [repo].
[arXiv-2025] SoK: Potentials and Challenges of Large Language Models for Reverse Engineering.
[arXiv-2025] An Empirical Study on the Effectiveness of Large Language Models for Binary Code Understanding.
[ICSME-2024] How Far Have We Gone in Binary Code Understanding Using Large Language Models.
[ISSTA-2024] Evaluating the Effectiveness of Decompilers.
[Security-2024] A taxonomy of C decompiler fidelity issues.
[arXiv-2023] Binary Code Summarization: Benchmarking ChatGPT/GPT-4 and Other Large Language Models. [repo].
[Security-2022] RE-Mind: a First Look Inside the Mind of a Reverse Engineer.
[Security-2022] Decomperson: How Humans Decompile and What We Can Learn From It.
[arXiv-2022] Pop Quiz! Can a Large Language Model Help With Reverse Engineering?. [repo].
[SP-2021] SoK: All You Ever Wanted to Know About x86/x64 Binary Disassembly But Were Afraid to Ask. [repo].

🛠️ Collected Tools

Tool	Last Commit	Stars
Noelo-Lab/kuna
cyberkaida/reverse-engineering-assistant
mrexodia/ida-pro-mcp
fosdickio/binary_ninja_mcp
mooncat-greenpy/KinGAidra
Forsworns/GhidrAI
mahaloz/DAILA
symgraph/BinAssist
radareorg/r2ai
r0ysue/r0idamcp
fdrechsler/mcp-server-idapro
ccq1/awsome_kali_mcpservers
zinja-coder/apktool-mcp-server
WPeace-HcH/WPeChatGPT
zyings/ReverseLLM
jtang613/GhidrAssistMCP
APiTJLillo/GhAIdra
JusticeRage/Gepetto
atredispartners/aidapal
likvidera/GhidraChatGPT
lr-m/GhidrOllama
cyberark/ByteCodeLLM
evyatar9/GptHidra
moyix/gpt-wpre
symgraph/GhidrAssist