서브메뉴
검색
Democratizing Human-Centered AI with Visual Explanation and Interactive Guidance
Democratizing Human-Centered AI with Visual Explanation and Interactive Guidance
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202105556
- ISBN
- 9798263399412
- DDC
- 005
- 저자명
- Wang, Zijie J.
- 서명/저자
- Democratizing Human-Centered AI with Visual Explanation and Interactive Guidance
- 발행사항
- [Sl] : Georgia Institute of Technology, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 198 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
- 주기사항
- Advisor: Chau, Duen Horng.
- 학위논문주기
- Thesis (Ph.D.)--Georgia Institute of Technology, 2024.
- 초록/해제
- 요약While artificial intelligence (AI) systems have been increasingly integrated into our everyday lives, how they make predictions often remains obscure to both their developers and the people they impact. The opacity of AI models contributes to their perception as "mysterious"-rendering both developers and those impacted by these models powerless when it comes to aligning AI models with their values.My dissertation aims to address these challenges with a human-centered approach, by designing and developing novel techniques and easy-to-adopt interactive tools that explain and guide AI models. Specifically, this thesis focuses on three complementary thrusts:1. Explain AI to Everyone. We pioneer easy-to-access interactive visualization systems that help AI novices and experts understand AI models (e.g., WIZMAP and CNN EXPLAINER used by 360k+ novices worldwide). We also present first-of-its-kind resources (e.g., 6.5TB DIFFUSIONDB with 14 million prompt-image pairs) to help AI developers and policymakers understand the impacts of large generative AI models.2. Guide AI with Human Values. To harness the potential of AI, gaining a better understanding of it is not enough. We empower AI developers to vet and fix problematic model behaviors (e.g., GAM CHANGER deployed by Microsoft) and those impacted by AI to receive customizable suggestions to alter unfavorable AI decisions (e.g., GAM COACH).3. Democratize Human-Centered AI. Human-centered AI practices are maximally valuable when they find practical adoption. To lower the barrier to applying these practices, we introduce in situ tools (e.g., FARSIGHT) to foster responsible AI awareness among practitioners during the prototyping stage within their current workflows.Our work is making significant impacts on academia, industry, and society: CNN EXPLAINER has helped 360k+ novices learn about CNNs worldwide, and it has been integrated into deep learning courses (Carnegie Mellon, Georgia Tech, Duke University, University of Tokyo and more). It has also been highlighted as a top visualization publication (top 1%) invited to SIGGRAPH. FARSIGHT has received a CHI Best Paper Honorable Mention award. DIFFUSIONDB has received an ACL Best Paper Honorable Mention award. GAM CHANGER has received the Best Paper Award at NeurIPS Workshop on Bridging the Gap: From ML Research to Clinical Practice, and the tool is now deployed by Microsoft and integrated into their inheritability library. Our work has been recognized by an Apple Scholars in AI/ML PhD fellowship and a J.P. Morgan AI PhD Fellowship.
- 일반주제명
- User interface
- 일반주제명
- Linear programming
- 일반주제명
- Large language models
- 일반주제명
- Neural networks
- 일반주제명
- Empowerment
- 기본자료저록
- Dissertations Abstracts International. 87-05B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2024 us c eng d■001000017360619
■00520260202105556
■006m o d
■007cr#unu||||||||
■020 ▼a9798263399412
■035 ▼a(MiAaPQ)AAI32315890
■035 ▼a(MiAaPQ)GeorgiaTech76860
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a005
■1001 ▼aWang, Zijie J.
■24510▼aDemocratizing Human-Centered AI with Visual Explanation and Interactive Guidance
■260 ▼a[Sl]▼bGeorgia Institute of Technology▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a198 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-05, Section: B.
■500 ▼aAdvisor: Chau, Duen Horng.
■5021 ▼aThesis (Ph.D.)--Georgia Institute of Technology, 2024.
■520 ▼aWhile artificial intelligence (AI) systems have been increasingly integrated into our everyday lives, how they make predictions often remains obscure to both their developers and the people they impact. The opacity of AI models contributes to their perception as "mysterious"-rendering both developers and those impacted by these models powerless when it comes to aligning AI models with their values.My dissertation aims to address these challenges with a human-centered approach, by designing and developing novel techniques and easy-to-adopt interactive tools that explain and guide AI models. Specifically, this thesis focuses on three complementary thrusts:1. Explain AI to Everyone. We pioneer easy-to-access interactive visualization systems that help AI novices and experts understand AI models (e.g., WIZMAP and CNN EXPLAINER used by 360k+ novices worldwide). We also present first-of-its-kind resources (e.g., 6.5TB DIFFUSIONDB with 14 million prompt-image pairs) to help AI developers and policymakers understand the impacts of large generative AI models.2. Guide AI with Human Values. To harness the potential of AI, gaining a better understanding of it is not enough. We empower AI developers to vet and fix problematic model behaviors (e.g., GAM CHANGER deployed by Microsoft) and those impacted by AI to receive customizable suggestions to alter unfavorable AI decisions (e.g., GAM COACH).3. Democratize Human-Centered AI. Human-centered AI practices are maximally valuable when they find practical adoption. To lower the barrier to applying these practices, we introduce in situ tools (e.g., FARSIGHT) to foster responsible AI awareness among practitioners during the prototyping stage within their current workflows.Our work is making significant impacts on academia, industry, and society: CNN EXPLAINER has helped 360k+ novices learn about CNNs worldwide, and it has been integrated into deep learning courses (Carnegie Mellon, Georgia Tech, Duke University, University of Tokyo and more). It has also been highlighted as a top visualization publication (top 1%) invited to SIGGRAPH. FARSIGHT has received a CHI Best Paper Honorable Mention award. DIFFUSIONDB has received an ACL Best Paper Honorable Mention award. GAM CHANGER has received the Best Paper Award at NeurIPS Workshop on Bridging the Gap: From ML Research to Clinical Practice, and the tool is now deployed by Microsoft and integrated into their inheritability library. Our work has been recognized by an Apple Scholars in AI/ML PhD fellowship and a J.P. Morgan AI PhD Fellowship.
■590 ▼aSchool code: 0078.
■650 4▼aUser interface
■650 4▼aLinear programming
■650 4▼aLarge language models
■650 4▼aNeural networks
■650 4▼aEmpowerment
■690 ▼a0800
■71020▼aGeorgia Institute of Technology.
■7730 ▼tDissertations Abstracts International▼g87-05B.
■790 ▼a0078
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360619▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
