GPT类AI技术支持下的艺术创新

图书馆杂志 ›› 2023, Vol. 42 ›› Issue (385): 14-19.

GPT类AI技术支持下的艺术创新

张家榕1, 2, 3 洪贇3, 4 叶鹰3, 4, 5 （1 南京艺术学院文化产业学院 2 南京艺术学院紫金文创研究院 3 南京大学江苏省国际联合信息学实验室 4 南京大学信息管理学院 5 复旦大学国家智能评价与治理实验基地）

出版日期:2023-05-15 发布日期:2023-05-24
作者简介:张家榕南京艺术学院文化产业学院，讲师。研究方向：人工智能艺术。作者贡献：确定选题、论文撰写及修改、资料搜集。E-mail：jszhangjiarong@foxmail.com 江苏南京 210013 洪贇南京大学信息管理学院，硕士研究生。研究方向：定量信息分析。作者贡献：论文撰写及修改、资料搜集。江苏南京 210023 叶鹰复旦大学国家智能评价与治理实验基地，特聘研究员。研究方向：智能信息处理、定量信息分析。作者贡献：确定论文框架与思路、撰写及修改。上海 200433

A Probe into the Innovation in Arts Support by GPT-Type AI

Zhang Jiarong1, 2, 3, Hong Yun3, 4, Ye Ying3, 4, 5 (1 School of Cultural Industries, Nanjing University of the Arts; 2 Purple Academy of Culture & Creativity, Nanjing University of the Arts; 3 International Joint Informatics Laboratory, Nanjing University; 4 School of Information Management, Nanjing University; 5 National Institute of Intelligent Evaluation and Governance, Fudan University)

Online:2023-05-15 Published:2023-05-24
About author:Zhang Jiarong1, 2, 3, Hong Yun3, 4, Ye Ying3, 4, 5 (1 School of Cultural Industries, Nanjing University of the Arts; 2 Purple Academy of Culture & Creativity, Nanjing University of the Arts; 3 International Joint Informatics Laboratory, Nanjing University; 4 School of Information Management, Nanjing University; 5 National Institute of Intelligent Evaluation and Governance, Fudan University)

摘要/Abstract

摘要： GPT类人工智能技术，特别是GPT-4等大型语言模型，正在深刻地改变艺术创作领域。该技术透过大量数据训练，能适应多种任务场景，实现跨模态信息理解和任务迁移，可望在文学、音乐、绘画和戏剧等艺术创新领域发挥不可估量的作用。GPT类技术的应用场景包括生成和激发创造力、交互与协作，以及跨模态实现等。GPT类技术将推动跨学科合作、个性化体验、艺术工程和创新实验等方面的发展，掀起艺术创作领域的革命，值得学术界和工业界高度关注。

Abstract: GPT-type AI technologies, especially large-scale language models such as GPT-4, are profoundly changing the field of artistic creation. Through a large amount of data training, this technology can adapt to a variety of task scenarios, realize cross-modal information understanding and task transfer, and is expected to play an inestimable role in artistic innovation fields such as literature, music, painting and drama. Application scenarios for GPT-type technologies include generating and stimulating creativity, interaction and collaboration, and cross-modal implementation. GPT-type technologies will promote the development of interdisciplinary cooperation, personalized experience, art engineering and innovative experimentation, and set off a revolution in the field of artistic creation, which deserves great attention from academia and industry.

张家榕洪贇叶鹰（南京艺术学院文化产业学院南京艺术学院紫金文创研究院南京大学江苏省国际联合信息学实验室南京大学信息管理学院复旦大学国家智能评价与治理实验基地）. GPT类AI技术支持下的艺术创新[J]. 图书馆杂志, 2023, 42(385): 14-19.

Zhang Jiarong, Hong Yun, Ye Ying( School of Cultural Industries, Nanjing University of the Arts, Purple Academy of Culture & Creativity, Nanjing University of the Arts, International Joint Informatics Laboratory, Nanjing University, School of Information Management, Nanjing University, National Institute of Intelligent Evaluation and Governance, Fudan University). A Probe into the Innovation in Arts Support by GPT-Type AI[J]. Libraly Journal, 2023, 42(385): 14-19.

参考文献

[ 1 ] Floridi L，Chiriatti M. GPT-3： Its nature，scope，limits，and consequences[J]. Minds and Machines，2020，30（4）： 681–694.
[ 2 ] OpenAI. GPT-4 technical report [J]. arXiv preprint arXiv：2303.08774，2023.
[ 3 ] Bongini P，Becattini F，Del Bimbo A. Is GPT-3 All You Need for Visual Question Answering in Cultural Heritage?[C]//Computer Vision–ECCV 2022 Workshops： Tel Aviv，Israel，October 23–27，2022，Proceedings，Part I. Cham： Springer Nature Switzerland，2023： 268–281.
[ 4 ] 吴敏. 关于哲学，艺术及其辩证关系[J]. 理论探索，2006 （3）： 13–15.
[ 5 ] 辞海[M]. 7版.上海：上海辞书出版社，2020.
[ 6 ] 程承坪. 人工智能对劳动的替代，极限及对策[J]. 上海师范大学学报（哲学社会科学版），2020，49（2）： 85–93.
[ 7 ] 马立新，涂少辉.AI艺术创作机理研究[J].美术研究，2022（6）：82–86.
[ 8 ] 叶鹰. 图书情报学的学术思想与技术方法及其开新[J]. 中国图书馆学报，2019，45（2）： 15–25.
[ 9 ] Vaswani A，Shazeer N，Parmar N，et al. Attention is all you need[J]. Advances in Neural Information Processing Systems，2017，30： 5998–6008.
[10] Su H，Shen X，Zhao S，et al. Diversifying dialogue generation with non-conversational text[J]. arXiv preprint arXiv：2005.04346，2020.
[11] Oord A，Dieleman S，Zen H，et al. Wavenet： A generative model for raw audio[J]. arXiv preprint arXiv：1609.03499，2016.
[12] Saharia C，Chan W，Saxena S，et al. Photorealistic text-to-image diffusion models with deep language understanding[J]. Advances in Neural Information Processing Systems，2022，35： 36479–36494.
[13] Casella P，Paiva A. Magenta： An architecture for real time automatic composition of background music[C]//Intelligent Virtual Agents： Third International Workshop，IVA 2001 Madrid，Spain，September 10–11，2001 Proceedings 3. Springer Berlin Heidelberg，2001： 224–232.
[14] Sun Q，Chen Y，Tao W，et al. A GAN-based approach toward architectural line drawing colorization prototyping[J]. The Visual Computer，2022： 1–18.
[15] Zhao Y，Fu G，Wang H，et al. Image Style Migration
Based on Improved CycleGAN[J]. Scientific Journal of Intelligent Systems Research Volume，2021，3（2）： 36–42.
[16] Kalakonda S S，Maheshwari S，Sarvadevabhatla R K. Action-GPT： Leveraging Large-scale Language Models for Improved and Generalized Zero Shot Action Generation[J]. arXiv preprint arXiv：2211.15603，2022.