東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Side Channel Analysis for AI Infrast...

Yuan, Yuanyuan,

FindBook

Google Book

Amazon

博客來

Side Channel Analysis for AI Infrastructures /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Side Channel Analysis for AI Infrastructures // Yuanyuan Yuan.
作者:	Yuan, Yuanyuan,
面頁冊數:	1 electronic resource (244 pages)
附註:	Source: Dissertations Abstracts International, Volume: 86-06, Section: A.
Contained By:	Dissertations Abstracts International86-06A.
標題:	Software. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31810430
ISBN:	9798346824343

Side Channel Analysis for AI Infrastructures /
Yuan, Yuanyuan,

Side Channel Analysis for AI Infrastructures /Yuanyuan Yuan. - 1 electronic resource (244 pages)

Source: Dissertations Abstracts International, Volume: 86-06, Section: A.

Side channel analysis (SCA) investigates the unintended secret leakage in a system's nonfunctional characteristics such as execution time or memory access patterns. Given the growing adoption of AI systems in security-critical and privacy-preserving applications, this thesis comprehensively studies side channel leakages in infrastructures that underpin the entire life cycle of modern AI systems, including data processing libraries, trusted execution environments (TEEs), runtime interpreters, executables on edge devices, etc. It also proposes highly practical solutions for SCA.On the offensive side, this thesis demonstrates end-to-end attacks that recover user's inputs (i.e., the user's privacy) and the underlying neural networks (i.e., the intellectual property) of AI systems from various side channels. We first consider a non-privileged co-process as the attacker and exploit cache side channels. Modern AI systems adopt data processing libraries (e.g., Libjpeg, FFmpeg) to handle various formats of user inputs like images and audios. Our work for the first time recovers such complex inputs from these libraries' cache side channels. Then, we consider untrusted hosts as adversaries. While TEEs are widely employed to shield AI systems from malicious host platforms,our works exploit the ciphertext side channels in TEEs and reconstruct both inputs and neural networks from TEE-shielded AI systems, breaking the security belief of TEEs. We also propose the first unified recovery scheme for complex input data like images, audios, text, videos, etc., and the first practical reconstruction technique for model weights of real-world neural networks.On the defensive side, this thesis localizes hundreds of new leakage sources in the exploited systems. Prior works often adopt program analysis techniques to model leakage patterns, whose localization is leakage-specific and suffers from the scalability issue. We recast the information leakage as a cooperative game among all leakage sources and reduce the cost from exponential to nearly constant. Our work presents a generic localization pipeline and supports analyzing production-size programs and AI infrastructures for the first time. We systematically examine the leakage in AI runtime interpreters including TensorFlow and PyTorch, and study how their different computation paradigms affect the leakage. Our analysis also reveals that optimizations in AI compilers (e.g., TVM, Glow) enlarge the leakage in compiled neural network executables.

English

ISBN: 9798346824343Subjects--Topical Terms:

619355
Software.

Side Channel Analysis for AI Infrastructures /
LDR:03763nmm a22003733i 4500 001 2433054
005 20250818063918.5
006 m o d
007 cr|nu||||||||
008 261011s2024 miu||||||m |||||||eng d
020 $a 9798346824343
035 $a (MiAaPQD)AAI31810430
035 $a AAI31810430
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Yuan, Yuanyuan, $e author. $3 3822507
245 1 0 $a Side Channel Analysis for AI Infrastructures / $c Yuanyuan Yuan.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (244 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: 86-06, Section: A.
500 $a Advisors: Wang, Shua; Zhou, Xiaofang.
502 $b Ph.D. $c Hong Kong University of Science and Technology (Hong Kong) $d 2024.
520 $a Side channel analysis (SCA) investigates the unintended secret leakage in a system's nonfunctional characteristics such as execution time or memory access patterns. Given the growing adoption of AI systems in security-critical and privacy-preserving applications, this thesis comprehensively studies side channel leakages in infrastructures that underpin the entire life cycle of modern AI systems, including data processing libraries, trusted execution environments (TEEs), runtime interpreters, executables on edge devices, etc. It also proposes highly practical solutions for SCA.On the offensive side, this thesis demonstrates end-to-end attacks that recover user's inputs (i.e., the user's privacy) and the underlying neural networks (i.e., the intellectual property) of AI systems from various side channels. We first consider a non-privileged co-process as the attacker and exploit cache side channels. Modern AI systems adopt data processing libraries (e.g., Libjpeg, FFmpeg) to handle various formats of user inputs like images and audios. Our work for the first time recovers such complex inputs from these libraries' cache side channels. Then, we consider untrusted hosts as adversaries. While TEEs are widely employed to shield AI systems from malicious host platforms,our works exploit the ciphertext side channels in TEEs and reconstruct both inputs and neural networks from TEE-shielded AI systems, breaking the security belief of TEEs. We also propose the first unified recovery scheme for complex input data like images, audios, text, videos, etc., and the first practical reconstruction technique for model weights of real-world neural networks.On the defensive side, this thesis localizes hundreds of new leakage sources in the exploited systems. Prior works often adopt program analysis techniques to model leakage patterns, whose localization is leakage-specific and suffers from the scalability issue. We recast the information leakage as a cooperative game among all leakage sources and reduce the cost from exponential to nearly constant. Our work presents a generic localization pipeline and supports analyzing production-size programs and AI infrastructures for the first time. We systematically examine the leakage in AI runtime interpreters including TensorFlow and PyTorch, and study how their different computation paradigms affect the leakage. Our analysis also reveals that optimizations in AI compilers (e.g., TVM, Glow) enlarge the leakage in compiled neural network executables.
546 $a English
590 $a School code: 1223
650 4 $a Software. $2 gtt. $3 619355
650 4 $a Privacy. $3 528582
650 4 $a Interpreters. $3 3563334
650 4 $a Neural networks. $3 677449
650 4 $a Translation studies. $3 3310485
690 $a 0800
690 $a 0213
710 2 $a Hong Kong University of Science and Technology (Hong Kong). $e degree granting institution. $3 3814653
720 1 $a Wang, Shua $e degree supervisor.
720 1 $a Zhou, Xiaofang $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-06A.
790 $a 1223
791 $a Ph.D.
792 $a 2024
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31810430