Machine Learning Based Malwae Code Classifier Attack Using Hostile Data 


Vol. 43,  No. 1, pp. 77-80, Jan.  2018
10.7840/kics.2018.43.1.77


PDF
  Abstract

In order to find vulnerabilities of malware classification method, this paper proposes a new attack scheme that can reduce the accuracy of malware classification. To realize the proposed concept, we use the following procedures; 1) After learning an imitation model by using black-box attack strategy, 2) Crafting adversarial samples from the imitation model, and 3) attacking target model. An implementation by using the Microsoft’s data and experimental analysis show that the previously proposed malware classification accuracy can be forced to decrease from 77.82% to 53.21%.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Cite this article

[IEEE Style]

S. Yoon and C. h. Kim, "Machine Learning Based Malwae Code Classifier Attack Using Hostile Data," The Journal of Korean Institute of Communications and Information Sciences, vol. 43, no. 1, pp. 77-80, 2018. DOI: 10.7840/kics.2018.43.1.77.

[ACM Style]

Sung-koock Yoon and Chang hoon Kim. 2018. Machine Learning Based Malwae Code Classifier Attack Using Hostile Data. The Journal of Korean Institute of Communications and Information Sciences, 43, 1, (2018), 77-80. DOI: 10.7840/kics.2018.43.1.77.

[KICS Style]

Sung-koock Yoon and Chang hoon Kim, "Machine Learning Based Malwae Code Classifier Attack Using Hostile Data," The Journal of Korean Institute of Communications and Information Sciences, vol. 43, no. 1, pp. 77-80, 1. 2018. (https://doi.org/10.7840/kics.2018.43.1.77)

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880