Proceedings of the 2021 ACM Asia Conference on Computer and Communications Security 2021
DOI: 10.1145/3433210.3437533
|View full text |Cite
|
Sign up to set email alerts
|

BugGraph: Differentiating Source-Binary Code Similarity with Graph Triplet-Loss Network

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
14
0

Year Published

2021
2021
2023
2023

Publication Types

Select...
3
3
2

Relationship

1
7

Authors

Journals

citations
Cited by 24 publications
(14 citation statements)
references
References 42 publications
0
14
0
Order By: Relevance
“…This method achieves a classification accuracy of 80%. Ji et al 49 proposed a hybrid approach of Attributed Control Flow Graph (ACFG), and Graph Triplet Loss Network (GTN) to extract code similarity. The program flow is collected using ACFG, and a Tensorflow‐based GTN model is used to rank related functions across many software projects.…”
Section: Resultsmentioning
confidence: 99%
“…This method achieves a classification accuracy of 80%. Ji et al 49 proposed a hybrid approach of Attributed Control Flow Graph (ACFG), and Graph Triplet Loss Network (GTN) to extract code similarity. The program flow is collected using ACFG, and a Tensorflow‐based GTN model is used to rank related functions across many software projects.…”
Section: Resultsmentioning
confidence: 99%
“…Finally, because our focus is on comparing binaries without source code, we intentionally exclude similarity comparison techniques that require source code. Nevertheless, it is noteworthy that there has been plentiful literature on comparing two source code snippets [75], [154], [155], [156], [157], [158], [159], [160], [161], [162] or comparing source snippets with binary snippets [163], [164], [165], [166].…”
Section: Discussionmentioning
confidence: 99%
“…The works that consider the CFG from a graph perspective and apply deep learning methods to represent the CFG are described in the following. The attributed Control Flow Graph (ACFG) is a common preprocessing step in some of the following works, especially in binary code similarity detection, such as Genius [49], Gemini [178] and BugGraph [75]. There are also some typical modifications for CFG, for example, lazy-binding CFG [121], inter-procedural CFGs(ICFG) [47].…”
Section: Flow-graph-based Structuresmentioning
confidence: 99%
“…GAT extracts the context information for each statement, which is modeled by the events that are embedded to capture execution semantics. BugGraph [75] compares the source-binary code similarity in two steps: source binary canonicalization and code similarity computation. In the code similarity computation step, BugGraph computes the similarity between the target and the comparing binary code.…”
Section: Flow-graph-based Structuresmentioning
confidence: 99%