摘要
人工智能生成内容(AIGC)正推动计算艺术从早期以规则驱动为特征的范式转向更具创造力的自动生成机制。当前,AIGC在文本、图像、音频等多模态领域已展现出接近人类的内容生成能力,其在内容理解、风格迁移与语义联想等方面的优异表现,使其在创意实践中的应用成为可能。在此基础上,本文构建了一个以概率建模为核心的AIGC理论框架,将数据、模型与生成机制三大关键要素整合为协同交互的系统。进一步地,本文以中国传统文化的当代表达为实践范式,探索包括多模态文化数据集构建、多模态模型训练与文化风格建构、人机交互式创作平台开发在内的系统路径,旨在探讨AIGC赋能下计算艺术的技术逻辑、文化转译机制与美学重构的可能性。
Artificial Intelligence Generated Content(AIGC)is driving a paradigm shift in computational art,transitioning from early rule-driven models to more creative,autonomous generative mechanisms.Currently,AIGC demonstrates human-like generative capabilities across multimodal domains such as text,image,and audio.Its outstanding performance in content understanding,style transfer,and semantic association has enabled its application in creative practices.This paper proposes a theoretical framework of AIGC centered on probabilistic modeling,integrating data semantics,generative models,and control mechanisms into a collaborative interactive system.Furthermore,taking the contemporary representation of Chinese traditional culture as a practical paradigm,this paper explores a systematic approach that includes the construction of multimodal cultural datasets,the training of multimodal models for cultural style formation,and the development of human-computer interactive creative platforms.The aim is to investigate the technical logic,cultural translation mechanisms,and aesthetic reconstruction possibilities of computational art empowered by AIGC.
出处
《中国艺术》
2025年第4期45-58,共14页
Chinese Art
基金
2021年度国家自然科学基金面上项目“基于人工智能的绘画艺术关键技术研究”(项目编号:62176006)
科大讯飞股份有限公司校企合作项目“计算艺术应用研究”的阶段性成果。
关键词
AIGC
计算艺术
理论框架
中国传统文化
AIGC
Computational Art
Theoretical Framework
Chinese Traditional Culture
