영상 디자인 전공을 위한 머신러닝 교육 방법 연구: 컴퓨팅 패러다임의 변화와 생성형 AI 모형 도구 활용을 중심으로

Abstract

The popularization of artificial intelligence progressed gradually with the development of machine learning technology, but with the advent of ChatGPT, artificial intelligence brought about revolutionary changes in all fields of society. To understand this process, the need to study it from the perspective of a change in the computing paradigm has emerged. This paper outlined the teaching methods of machine learning in the interactive media design curriculum based on the larger paradigm of the combination of design and programming and discussed paradigm changes, actual implementation methods, implications through introduction cases, and prospects. First, the change stages of the computing interface were analyzed by the Command Line Interface (CLI), Graphic User interface (GUI), and Natural Language Interface (NLI), and the change in the design environment according to the change in the interface was discussed. Next, to view the environmental status of generative machine learning, an analysis of Generative AI environments such as Mid-journey and Stable Diffusion was conducted, and the use cases of Generative AI were considered. As a real-world case study of Generative AI, a design class using NLI was constructed and conducted, and through in-depth interviews with students, the effectiveness of Generative AI tools for formative composition tasks, video production tasks, and two production tasks was discussed. In the case of configuration tasks, it has been reported that NLI has advantages in terms of image generation speed and ability to create various styles compared to CLI and GUI. In the case of video tasks, it has been reported that NLI shortens the pre-visualization time in the pre-production stage and has the greatest utility in exploring various ideas. Based on this result, implications for a design education model that can accommodate the change in the computing paradigm were additionally discussed.