NeuroDeX: Unlocking Diverse Support in Decompiling Deep Neural Network Executables
This addresses the threat of reverse engineering for on-device deep learning models, providing a more comprehensive solution for security researchers and developers, though it appears incremental by building on prior decompilation efforts.
The paper tackles the problem of decompiling deep neural network executables, which are vulnerable to reverse engineering, by introducing NeuroDeX, a method that uses LLMs and dynamic analysis to recover high-level models from compiled executables, achieving nearly identical models for non-quantized cases and 72% top-1 accuracy for quantized ones.
On-device deep learning models have extensive real world demands. Deep learning compilers efficiently compile models into executables for deployment on edge devices, but these executables may face the threat of reverse engineering. Previous studies have attempted to decompile DNN executables, but they face challenges in handling compilation optimizations and analyzing quantized compiled models. In this paper, we present NeuroDeX to unlock diverse support in decompiling DNN executables. NeuroDeX leverages the semantic understanding capabilities of LLMs along with dynamic analysis to accurately and efficiently perform operator type recognition, operator attribute recovery and model reconstruction. NeuroDeX can recover DNN executables into high-level models towards compilation optimizations, different architectures and quantized compiled models. We conduct experiments on 96 DNN executables across 12 common DNN models. Extensive experimental results demonstrate that NeuroDeX can decompile non-quantized executables into nearly identical high-level models. NeuroDeX can recover functionally similar high-level models for quantized executables, achieving an average top-1 accuracy of 72%. NeuroDeX offers a more comprehensive and effective solution compared to previous DNN executables decompilers.