東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Understanding and Mitigating Neural ...

Pang, Ren,

Linked to FindBook

Google Book

Amazon

博客來

Understanding and Mitigating Neural Backdoors /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Understanding and Mitigating Neural Backdoors // Ren Pang.
Author:	Pang, Ren,
Description:	1 electronic resource (138 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 86-03, Section: A.
Contained By:	Dissertations Abstracts International86-03A.
Subject:	Poisoning. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31353855
ISBN:	9798384211402

Understanding and Mitigating Neural Backdoors /
Pang, Ren,

Understanding and Mitigating Neural Backdoors /Ren Pang. - 1 electronic resource (138 pages)

Source: Dissertations Abstracts International, Volume: 86-03, Section: A.

The rapid progress in deep learning has led to significant breakthroughs in various machine learning tasks. Despite the remarkable success of deep learning models across domains, the intensive research has produced a plethora of backdoor attacks/defenses, resulting in a constant arms race. However, previous studies have highlighted the intricate trade-offs and complexities involved, yet a fundamental understanding of the connections between different attack vectors remains elusive. Furthermore, the lack of standardized evaluation benchmarks has hindered comprehensive research from multiple critical research questions.To address these limitations, we present three significant contributions in this dissertation. (i) We propose IMC, which enhances conventional backdoor attacks by jointly optimizing triggers and trojaned models, uncovering intriguing mutual reinforcement effects between the two attack vectors. (ii) We introduce TrojanZoo, an open-source platform designed to evaluate neural backdoor attacks and defenses holistically. Through systematic analysis, TrojanZoo reveals key insights into the design spectrum of existing attacks and defenses. (iii) We extend the scope of backdoor attacks to AutoML by introducing EVAS, a novel attack leveraging neural architecture search to discover architectures with inherent vulnerabilities. According to extensive evaluation, EVAS demonstrates high evasiveness, transferability, and robustness, raising important considerations for future defense strategies.

English

ISBN: 9798384211402Subjects--Topical Terms:

770903
Poisoning.

Understanding and Mitigating Neural Backdoors /
LDR:02772nmm a22003613i 4500 001 2432413
005 20250707081458.5
006 m o d
007 cr|nu||||||||
008 261011s2024 miu||||||m |||||||eng d
020 $a 9798384211402
035 $a (MiAaPQD)AAI31353855
035 $a (MiAaPQD)PennState30354rbp5354
035 $a AAI31353855
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Pang, Ren, $e author. $3 3821623
245 1 0 $a Understanding and Mitigating Neural Backdoors / $c Ren Pang.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (138 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-03, Section: A.
500 $a Advisors: Wang, Ting Committee members: Zhang, Xiang; Ma, Fenglong; Zhu, Sencun; Lee, Dongwon.
502 $b Ph.D. $c The Pennsylvania State University $d 2024.
520 $a The rapid progress in deep learning has led to significant breakthroughs in various machine learning tasks. Despite the remarkable success of deep learning models across domains, the intensive research has produced a plethora of backdoor attacks/defenses, resulting in a constant arms race. However, previous studies have highlighted the intricate trade-offs and complexities involved, yet a fundamental understanding of the connections between different attack vectors remains elusive. Furthermore, the lack of standardized evaluation benchmarks has hindered comprehensive research from multiple critical research questions.To address these limitations, we present three significant contributions in this dissertation. (i) We propose IMC, which enhances conventional backdoor attacks by jointly optimizing triggers and trojaned models, uncovering intriguing mutual reinforcement effects between the two attack vectors. (ii) We introduce TrojanZoo, an open-source platform designed to evaluate neural backdoor attacks and defenses holistically. Through systematic analysis, TrojanZoo reveals key insights into the design spectrum of existing attacks and defenses. (iii) We extend the scope of backdoor attacks to AutoML by introducing EVAS, a novel attack leveraging neural architecture search to discover architectures with inherent vulnerabilities. According to extensive evaluation, EVAS demonstrates high evasiveness, transferability, and robustness, raising important considerations for future defense strategies.
546 $a English
590 $a School code: 0176
650 4 $a Poisoning. $3 770903
650 4 $a Defense. $3 3681633
690 $a 0729
710 2 $a The Pennsylvania State University. $e degree granting institution. $3 3765934
720 1 $a Wang, Ting $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-03A.
790 $a 0176
791 $a Ph.D.
792 $a 2024
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31353855