Name
Title | ||
|---|---|---|
A machine learning-based scheme for the security analysis of authentication and key agreement protocols |
Abstract | ||
|---|---|---|
This paper proposes a novel machine learning-based scheme for the automatic analysis of authentication and key agreement protocols. Considering the traditional formal protocol analysis schemes, their analysis accuracies depend heavily on the prior knowledge possessed by the analyst and the subjective understanding of the protocol. The rapid development of artificial intelligence in security field shows that the ideal way to get rid of the dependency is to use machine learning. Hence, we elaborately compare more than 2000 protocol analysis results and select 500 most representative ones of them to build a protocol dataset. Combining the protocol representation method of traditional schemes, these selected protocols are expressed as weight matrixes based on security components. Furthermore, a machine learning-based security analysis model is proposed to automatically find the attacks of the protocol. For now, three types of attacks against authentication and key agreement protocols can be identified based on our model. And experiment results show that it can reach almost 72% upper-bound performance. From the derivative of the accuracy curves, it can be inferred that the performance of our scheme will definitely get better as the dataset expands. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1007/s00521-018-3929-8 | Neural Computing and Applications |
Keywords | DocType | Volume |
Authentication protocols, Machine learning, Formal analysis of protocol security, Protocol dataset | Journal | 32 |
Issue | ISSN | Citations |
22 | 1433-3058 | 0 |
PageRank | References | Authors |
0.34 | 18 | 5 |