서브메뉴
검색
Algorithms for Adversarially Robust Deep Learning
Algorithms for Adversarially Robust Deep Learning
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211152722
- ISBN
- 9798384044369
- DDC
- 004
- 서명/저자
- Algorithms for Adversarially Robust Deep Learning
- 발행사항
- [Sl] : University of Pennsylvania, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 470 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-03, Section: B.
- 주기사항
- Advisor: Hassani, Hamed;Pappas, George J.
- 학위논문주기
- Thesis (Ph.D.)--University of Pennsylvania, 2024.
- 초록/해제
- 요약Given the widespread use of deep learning models in safety-critical applications, ensuring that the decisions of such models are robust against adversarial exploitation is of fundamental importance. In this thesis, we discuss recent progress toward designing algorithms that exhibit desirable robustness properties. First, we discuss the problem of adversarial examples in computer vision, for which we introduce new technical results, training paradigms, and certification algorithms. Next, we consider the problem of domain generalization, wherein the task is to train neural networks to generalize from a family of training distributions to unseen test distributions. We present new algorithms that achieve state-of-the-art generalization in medical imaging, molecular identification, and image classification. Finally, we study the setting of jailbreaking large language models (LLMs), wherein an adversarial user attempts to design prompts that elicit objectionable content from an LLM. We propose new attacks and defenses, which represent the frontier of progress toward designing robust language-based agents.
- 일반주제명
- Computer science
- 일반주제명
- Systems science
- 키워드
- AI safety
- 키워드
- Jailbreaking
- 기타저자
- University of Pennsylvania Electrical and Systems Engineering
- 기본자료저록
- Dissertations Abstracts International. 86-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017163544
■00520250211152722
■006m o d
■007cr#unu||||||||
■020 ▼a9798384044369
■035 ▼a(MiAaPQ)AAI31489830
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a004
■1001 ▼aRobey, Alexander Beck.
■24510▼aAlgorithms for Adversarially Robust Deep Learning
■260 ▼a[Sl]▼bUniversity of Pennsylvania▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a470 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-03, Section: B.
■500 ▼aAdvisor: Hassani, Hamed;Pappas, George J.
■5021 ▼aThesis (Ph.D.)--University of Pennsylvania, 2024.
■520 ▼aGiven the widespread use of deep learning models in safety-critical applications, ensuring that the decisions of such models are robust against adversarial exploitation is of fundamental importance. In this thesis, we discuss recent progress toward designing algorithms that exhibit desirable robustness properties. First, we discuss the problem of adversarial examples in computer vision, for which we introduce new technical results, training paradigms, and certification algorithms. Next, we consider the problem of domain generalization, wherein the task is to train neural networks to generalize from a family of training distributions to unseen test distributions. We present new algorithms that achieve state-of-the-art generalization in medical imaging, molecular identification, and image classification. Finally, we study the setting of jailbreaking large language models (LLMs), wherein an adversarial user attempts to design prompts that elicit objectionable content from an LLM. We propose new attacks and defenses, which represent the frontier of progress toward designing robust language-based agents.
■590 ▼aSchool code: 0175.
■650 4▼aComputer science
■650 4▼aSystems science
■653 ▼aAdversarial robustness
■653 ▼aAI safety
■653 ▼aGenerative models
■653 ▼aJailbreaking
■653 ▼aLarge language models
■690 ▼a0800
■690 ▼a0984
■690 ▼a0790
■71020▼aUniversity of Pennsylvania▼bElectrical and Systems Engineering.
■7730 ▼tDissertations Abstracts International▼g86-03B.
■790 ▼a0175
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17163544▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
