Paper Abstract and Keywords |
Presentation |
2023-10-11 14:30
[Poster Presentation]
Detecting Poisoning Attacks Using Adversarial Examples in Deep Phishing Detection Koko Nishiura, Tomotaka Kimura, Jun Cheng (Doshisha Univ.) |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In recent years, the convenience of online services has greatly improved, but the number of phishing scams has skyrocketed. To prevent phishing scams, detection systems based on deep learning have been proposed and shown to be effective. However, deep learning has a vulnerability of intentionally causing misclassification. In this presentation, we examine countermeasures against backdoor attacks that induce misclassification for input data containing specific triggers. Specifically, we generate noise using the AE (Adversarial Example) method from a detection system undergoing a backdoor attack, and compare the perturbed data with noise and the judgment labels of the perturbed data. When generating perturbation data, the area where noise is generated is restricted and the data is evaluated by shifting the position of the restricted area. If the trigger area contains noise, the trigger is broken and will not be misclassified, but if it does not contain noise, it will be misclassified. Comparing the judgment labels of the perturbation data with the perturbation data, the judgment label agreement differs depending on the trigger position, so the point where the value becomes abnormal can be determined to be the trigger. Furthermore, detection of a trigger indicates that the deep learning model used is under backdoor attack. In this presentation, we will evaluate the feasibility of detecting backdoor attacks and triggers using a dataset containing data with embedded triggers, assuming a real environment. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Adversarial Examples / Phishing / Deep Learning / Poisoning Attack / / / / |
Reference Info. |
IEICE Tech. Rep. |
Paper # |
|
Date of Issue |
|
ISSN |
|
Download PDF |
|
Conference Information |
Committee |
MIKA |
Conference Date |
2023-10-10 - 2023-10-12 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Okinawa Jichikaikan |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
|
Paper Information |
Registration To |
MIKA |
Conference Code |
2023-10-MIKA |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Detecting Poisoning Attacks Using Adversarial Examples in Deep Phishing Detection |
Sub Title (in English) |
|
Keyword(1) |
Adversarial Examples |
Keyword(2) |
Phishing |
Keyword(3) |
Deep Learning |
Keyword(4) |
Poisoning Attack |
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Koko Nishiura |
1st Author's Affiliation |
Doshisha University (Doshisha Univ.) |
2nd Author's Name |
Tomotaka Kimura |
2nd Author's Affiliation |
Doshisha University (Doshisha Univ.) |
3rd Author's Name |
Jun Cheng |
3rd Author's Affiliation |
Doshisha University (Doshisha Univ.) |
4th Author's Name |
|
4th Author's Affiliation |
() |
5th Author's Name |
|
5th Author's Affiliation |
() |
6th Author's Name |
|
6th Author's Affiliation |
() |
7th Author's Name |
|
7th Author's Affiliation |
() |
8th Author's Name |
|
8th Author's Affiliation |
() |
9th Author's Name |
|
9th Author's Affiliation |
() |
10th Author's Name |
|
10th Author's Affiliation |
() |
11th Author's Name |
|
11th Author's Affiliation |
() |
12th Author's Name |
|
12th Author's Affiliation |
() |
13th Author's Name |
|
13th Author's Affiliation |
() |
14th Author's Name |
|
14th Author's Affiliation |
() |
15th Author's Name |
|
15th Author's Affiliation |
() |
16th Author's Name |
|
16th Author's Affiliation |
() |
17th Author's Name |
|
17th Author's Affiliation |
() |
18th Author's Name |
|
18th Author's Affiliation |
() |
19th Author's Name |
|
19th Author's Affiliation |
() |
20th Author's Name |
|
20th Author's Affiliation |
() |
Speaker |
Author-1 |
Date Time |
2023-10-11 14:30:00 |
Presentation Time |
90 minutes |
Registration for |
MIKA |
Paper # |
|
Volume (vol) |
vol. |
Number (no) |
|
Page |
|
#Pages |
|
Date of Issue |
|
|