메뉴 건너뛰기
소속 기관 / 학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(Dongseo University) (Dongseo University) (Dongseo University)
저널정보
국제인공지능학회 The International Journal of Internet, Broadcasting and Communication The International Journal of Internet, Broadcasting and Communication Vol.18 No.5
오류 신고하기

피인용 0

검색

    초록·키워드

    We examine how probabilistic image generation reshapes the relations among visual reference, provenance, and authorship. Baudrillard's theory of simulacra remains valuable for understanding the autonomy of images but leaves unspecified how a learned model distributes training influence and how source attribution can be investigated. Combining a conceptual reconstruction of Baudrillard with a technical reading of diffusion models and identity-conditioning techniques such as IP-Adapter, we develop Generative Simulacra as a framework for images whose production cannot be traced to a single source artifact. Our analysis separates three distinctions: reference to a depicted subject versus evidence of an actual scene; documentation of a generation process versus attribution to training images; and distributed creative contribution versus legal co-authorship. We argue that learned probability distributions operationalize model-based simulation through statistical estimation and conditional sampling rather than abolishing code or causality, making the shift an operational distinction rather than a categorical break. We define No Singular Original and Distributed Authorship, propose a provenance evaluation protocol, and use the Zarya of the Dawn registration decision to clarify contribution-specific claims. The study contributes to media ontology, referentiality theory, and generative-AI research.

    최근 본 자료 전체보기