Idioms: Neural Decompilation With Joint Code and Type Prediction

2025-02-06

AI Summary

Existing neural decompilers struggle with user-defined composite types (e.g., structs and classes), leading to semantic loss, poor readability, and heavy reliance on test cases. Method: We propose the first joint code-and-type prediction paradigm, introducing a cross-procedural control- and data-flow-aware contextual modeling mechanism, and construct Realtype—the first benchmark dataset featuring real-world composite types. Leveraging a large language model fine-tuning framework, we design a multi-task joint decoding architecture. Contribution/Results: Our approach achieves state-of-the-art performance on Realtype. We open-source the lightweight and efficient Idioms model series, significantly improving semantic completeness and readability of decompiled output. This work advances both open scientific research and industrial-grade reverse engineering.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Deep Generative Models & Autoencoders
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Decompilers are important tools for reverse engineers that help them analyze software at a higher level of abstraction than assembly. Unfortunately, because compilation is lossy, deterministic decompilers produce code that is missing many of the details that make source code readable in the first place, like variable names and types. Neural decompilers, on the other hand, offer the ability to statistically fill in these details. Existing work in neural decompilation, however, suffers from substantial drawbacks that limits its ability to handle real code: it is unable to handle user-defined composite types, which are essential to fully specifying many functions' semantics, or require test cases. In this work, we introduce a new training process to finetune any LLM into a neural decompiler capable of generating the appropriate user-defined types alongside the decompilation. We introduce a new dataset, Realtype, that includes substantially more complicated and realistic types than existing neural decompilation benchmarks. Motivated by the intuition that different parts of data structures can be operated upon by different parts of the program, we show that interprocedural context can help improve neural decompilers' ability to handle user-defined types. We show that our training process yields state-of-the-art results in neural decompilation. We also publicly release the Idioms series of finetuned neural decompilation models in support of open science. In summary, we identify the need for joint code and type prediction, show that it is a hard problem, and take the first steps towards solving it.

Problem

• Neural decompilers miss user-defined types
• Existing methods require test cases
• Interprocedural context improves type handling

Innovation

• Neural decompilation with type prediction
• Interprocedural context enhances decompilation
• Realtype dataset for realistic type handling