Learning is never secure: Poison learning by Intrusion Detection System based on Self-Organizing Map

Rupam Kumar Sharma; Hemanta Kumar Kalita; S Das; Biju Issac

Learning is never secure: Poison learning by Intrusion Detection System based on Self-Organizing Map

Rupam Kumar Sharma, Hemanta Kumar Kalita, S Das, Biju Issac

Research output: Contribution to conference › Paper › peer-review

Abstract

Machine learning has proven to enhance the detection rate by any Intrusion Detection Engine. However, if learning is tampered than there is a likeliness that the entire detection engine might fail. This paper explores the adversary learning by SOM (Self Organizing Map) trained on NSL-KDD data set. Experimental results demonstrate the deviation of the learning behaviour after being trained with crafted poison packets to mislead the learning.

Original language	English
Publication status	Published - 24 Dec 2016
Event	5th IEEE-EDS International Conference on Computing, Communication and Sensor Network - Kolkata, India Duration: 24 Dec 2016 → 25 Dec 2016

Conference

Conference	5th IEEE-EDS International Conference on Computing, Communication and Sensor Network
Abbreviated title	CCSN 2016
Country/Territory	India
City	Kolkata
Period	24/12/16 → 25/12/16

Cite this

@conference{3067fcc1b4f0419aa588a67f720eee0e,

title = "Learning is never secure: Poison learning by Intrusion Detection System based on Self-Organizing Map",

abstract = "Machine learning has proven to enhance the detection rate by any Intrusion Detection Engine. However, if learning is tampered than there is a likeliness that the entire detection engine might fail. This paper explores the adversary learning by SOM (Self Organizing Map) trained on NSL-KDD data set. Experimental results demonstrate the deviation of the learning behaviour after being trained with crafted poison packets to mislead the learning.",

author = "Sharma, {Rupam Kumar} and Kalita, {Hemanta Kumar} and S Das and Biju Issac",

year = "2016",

month = dec,

day = "24",

language = "English",

note = "5th IEEE-EDS International Conference on Computing, Communication and Sensor Network, CCSN 2016 ; Conference date: 24-12-2016 Through 25-12-2016",

}

TY - CONF

T1 - Learning is never secure: Poison learning by Intrusion Detection System based on Self-Organizing Map

AU - Sharma, Rupam Kumar

AU - Kalita, Hemanta Kumar

AU - Das, S

AU - Issac, Biju

PY - 2016/12/24

Y1 - 2016/12/24

N2 - Machine learning has proven to enhance the detection rate by any Intrusion Detection Engine. However, if learning is tampered than there is a likeliness that the entire detection engine might fail. This paper explores the adversary learning by SOM (Self Organizing Map) trained on NSL-KDD data set. Experimental results demonstrate the deviation of the learning behaviour after being trained with crafted poison packets to mislead the learning.

AB - Machine learning has proven to enhance the detection rate by any Intrusion Detection Engine. However, if learning is tampered than there is a likeliness that the entire detection engine might fail. This paper explores the adversary learning by SOM (Self Organizing Map) trained on NSL-KDD data set. Experimental results demonstrate the deviation of the learning behaviour after being trained with crafted poison packets to mislead the learning.

M3 - Paper

T2 - 5th IEEE-EDS International Conference on Computing, Communication and Sensor Network

Y2 - 24 December 2016 through 25 December 2016

ER -

Learning is never secure: Poison learning by Intrusion Detection System based on Self-Organizing Map

Abstract

Conference

Fingerprint

Cite this