Abstract:Binary code similarity analysis is widely used in the field of vulnerability search where source code may not be available to detect whether two binary functions are similar or not. Based on deep learning and natural processing techniques, several approaches have been proposed to perform cross-platform binary code similarity analysis using control flow graphs. However, existing schemes suffer from the shortcomings of large differences in instruction syntaxes across different target platforms, inability to align control flow graph nodes, and less introduction of high-level semantics of stability, which pose challenges for identifying similar computations between binary functions of different platforms generated from the same source code. We argue that extracting stable, platform-independent semantics can improve model accuracy, and a cross-platform binary function similarity comparison model N_Match is proposed. The model elevates different platform instructions to the same semantic space to shield their underlying platform instruction differences, uses graph embedding technology to learn the stability semantics of neighbors, extracts high-level knowledge of naming function to alleviate the differences brought about by cross-platform and cross-optimization levels, and combines the stable graph structure as well as the stable, platform-independent API knowledge of naming function to represent the final semantics of functions. The experimental results show that the model accuracy of N_Match outperforms the baseline model in terms of cross-platform, cross-optimization level, and industrial scenarios. In the vulnerability search experiment, N_Match significantly improves hit@N, the mAP exceeds the current graph embedding model by 66%. In addition, we also give several interesting observations from the experiments. The code and model are publicly available at https://www.github.com/CSecurityZhongYuan/Binary-Name_Match.

Deep Learning-Based Program-Wide Binary Code Similarity for Smart Contracts

Cross-Modality Mutual Learning for Enhancing Smart Contract Vulnerability Detection on Bytecode

Ethereum Smart Contract Representation Learning for Robust Bytecode-Level Similarity Detection

Smart Contract Bytecode Similarity Detection Based on Self-supervised Learning

Towards Automatic Smart-contract Codes Classification by Means of Word Embedding Model and Transaction Information

BinDeep: A Deep Learning Approach to Binary Code Similarity Detection.

Bytecode Similarity Detection of Smart Contract across Optimization Options and Compiler Versions Based on Triplet Network

Binary Code Similarity Detection via Graph Contrastive Learning on Intermediate Representations

Hierarchical Attention Graph Embedding Networks for Binary Code Similarity against Compilation Diversity

A Review of Deep Learning-Based Binary Code Similarity Analysis

Research and implementation of obfuscation binary code similarity detection

Neural Network-based Graph Embedding for Cross-Platform Binary Code Similarity Detection.

SimCGE: Simple Contrastive Learning of Graph Embeddings for Cross-Version Binary Code Similarity Detection

Binary code similarity analysis based on naming function and common vector space

Codeformer: A GNN-Nested Transformer Model for Binary Code Similarity Detection

A Lightweight Cross-Version Binary Code Similarity Detection Based on Similarity and Correlation Coefficient Features

Multi-feature based Function Embedding Network for Binary Code Similarity

Neural Machine Translation Inspired Binary Code Similarity Comparison beyond Function Pairs

Multi-Level Cross-Architecture Binary Code Similarity Metric

FUSION: Measuring Binary Function Similarity with Code-Specific Embedding and Order-Sensitive GNN

Hunting Vulnerable Smart Contracts via Graph Embedding Based Bytecode Matching