2022
DOI: 10.1016/j.cose.2022.102607
|View full text |Cite
|
Sign up to set email alerts
|

Binary code traceability of multigranularity information fusion from the perspective of software genes

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
4
0

Year Published

2022
2022
2024
2024

Publication Types

Select...
5
1

Relationship

2
4

Authors

Journals

citations
Cited by 6 publications
(4 citation statements)
references
References 20 publications
0
4
0
Order By: Relevance
“…These methods often rely on extensive labeled data, a requirement our approach reduces by applying the K-Nearest Neighbors (KNN) algorithm, allowing for effective classification with fewer labeled samples. Adding to the graph-based analysis landscape, Huang et al [11] proposed a multi-granularity fusion feature based on biological gene concepts for binary code traceability. Similarly, Zhao et al [12] introduced a malware homology identification method using subgraphs of the Function Dependence Graph (FDG) as genes.…”
Section: Related Workmentioning
confidence: 99%
“…These methods often rely on extensive labeled data, a requirement our approach reduces by applying the K-Nearest Neighbors (KNN) algorithm, allowing for effective classification with fewer labeled samples. Adding to the graph-based analysis landscape, Huang et al [11] proposed a multi-granularity fusion feature based on biological gene concepts for binary code traceability. Similarly, Zhao et al [12] introduced a malware homology identification method using subgraphs of the Function Dependence Graph (FDG) as genes.…”
Section: Related Workmentioning
confidence: 99%
“…The graph convolutional network (GCN) [25][26][27] is a model that performs convolution operations on graphs. Marcheggiani et al [28] and Huang et al [29] demonstrated that sequence models and GCNs have complementary modeling capabilities; therefore, based on the instruction sequence vector obtained earlier, the GCN is used to fuse the edge information between basic blocks into block-level information. Based on this basic block intermediate representation vector, the main discussion is how to extract the jump relationship information between the CFG basic block nodes and generate basic block embeddings.…”
Section: Gcn-based Basic Block Embeddingmentioning
confidence: 99%
“…Graph neural networks mainly include graph convolution networks (GCN) [20,21], graph attention networks (GAT) [22], graph autoencoders (GAE), [23] etc. In the field of semantic representation of binary codes, Qiao et al [24] and Massarelli et al [25] use GCN for semantic embedding of functions, but considering that the CFG of functions is a directed graph, which will cause a certain loss of the structural information.…”
Section: Gatmentioning
confidence: 99%