ДОСЛІДЖЕННЯ МОЖЛИВОСТЕЙ ВИКОРИСТАННЯ ШТУЧНОГО ІНТЕЛЕКТУ У ВІЗУАЛЬНОМУ ТА КОМУНІКАТИВНОМУ ВИДАХ ГЕНЕРАТИВНОГО ДИЗАЙНУ
RESEARCH OF THE POSSIBILITIES OF USING ARTIFICIAL INTELLIGENCE IN VISUAL AND COMMUNICATIVE TYPES OF GENERATIVE DESIGN
Сторінки: 307-309. Номер: №4, 2023 (323) 
Автори:
СТРИЖОВА ОКСАНА
Хмельницький національний університет
ORCID ID: 0000-0003-2751-1807
e-mail: stryzhovao@khnu.edu.ua
БАЗИЛЮК Ельвіра
Хмельницький національний університет
ORCID ID: 0000-0003-3123-8372
e-mail: bazyliukel@khmnu.edu.ua
STRYZHOVA Oksana.
Khmelnytskyi National University
ORCID ID: 0000-0003-2751-1807
e-mail: stryzhovao@khnu.edu.ua
BAZYLIUK Elvira.
Khmelnytskyi National University
ORCID ID: 0000-0003-3123-8372
e-mail: bazyliukel@khmnu.edu.ua
DOI: https://www.doi.org/10.31891/2307-5732-2023-323-4-307-309
Анотація мовою оригіналу
У статті висвітлено основні результати аналізу сервісів нейромереж зі штучним інтелектом, які були апробовані при створення графічних зображень за допомогою нейромережевих принципів генеративного дизайну.
Ключові слова: візуальний і комунікативний дизайн, генеративний дизайн, нейромережі в дизайні.
Розширена анотація англійською мовою
The principles of generative design are based on parametric modeling and optimization of initial data, which is an effective means of solving design problems. But there is a complication due to the need for preliminary processing of the raw data by the designer manually. This leads to a decrease in the productivity of generative design. Therefore, to work out a large number of parameters and determine their limitations, the designer should use the capabilities of neural networks with artificial intelligence. AI-services are capable of processing large volumes of information, as well as finding their own errors and self-improvement. This also has a positive effect on the development and improvement of options and combinations created by neural networks. The deep machine self-learning of the neural network helps to free up the working time of the designer, which can be used to produce other creative solutions. The article highlights the main results of the analysis of the main services of neural networks with artificial intelligence, which were tested when creating graphic images with the help of neural network services. The work was carried out with AI-services: Midjourney, Leonardo AI, Jasper Art, Waifu2x, Zyro AI Image Upscaler, AI Image Enlarger, Scribble Diffusion 1.5. The peculiarities of the generation of graphic objects and the time with which various services learned to create images within the given data were studied. The involvement of generative services of neural networks with artificial intelligence allows obtaining multi-vector and parallel ways of making pre-project decisions, instead of the traditional linear trajectory of movement when solving tasks in graphic design. The use of AI-services allows you to very quickly create and assemble design prototypes and check and evaluate their visual or communicative hypothesis. The advantage of these services is their possibility of deep learning and self-learning without human supervision on data that may be unstructured at first.
Keywords: visual and communicative design, generative design, neural networks in design.
