Abstract:

The aim of this work is to generate synthetic images with an artistic style based on comic culture and using artificial intelligence tools (deep learning) to bring Science, Technology, Engineering, and Mathematics (STEM) subjects accessible to the young students.

Nowadays, one of the social challenges is to connect people, especially the youngest, with science. For this purpose, there is a real need to create tools that allow citizens to be more directly involved in scientific topics. One of the current trends in STEM education is the integration of art and aesthetic aspects in the learning process [1]. In this sense, comics represent a visually attractive educational tool, which can help to explain complex scientific concepts in a more accessible way through storytelling. The informal structure of the comic (panels, speech balloons, special effects…) also help to hold the attention and arouse the interest of public, especially students [2]. This kind of educational material was, until now, difficult to transpose to classrooms and scientific communications, as it requires artistic skills, its production can be time-consuming, and there may be permission or licensing issues for the use of copyrighted images. It is here where artificial intelligence can be very valuable, offering an easy and fast access to the creation and development of attractive comic-style images as learning resources, with the additional value of not having associated any kind copyright issues.

Artificial intelligence (AI) can be defined as “the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages” [3]. Although AI has been around for many years it has seen a great explosion in popularity in recent years due to the advent of most of the computing technology employed in deep learning. Deep learning is part of a broader family of machine learning methods based on artificial neural networks with representation learning. Recently, the emergence of new and robust models such as DALL-E (Open AI, San Francisco, CA), in 2021, or Midjourney (Inc. San Francisco, CA), and Stable Diffusion (CompVis group, LMU Munich, Germany) in 2022, is revolutionizing the world of AI-art by allowing the creation of synthetic images based on text-to-image models. These novel deep learning models provide professionals with additional plastic skills and give the possibility to generate their own royalty-free images. Thus, AI-art could contribute to enlighten and make more comprehensible notions that are difficult to explain for scientists and teachers and help citizens to better understand them.

Through this cooperative and transversal work between researchers, artists and artificial intelligences, the main objective of this work is to provide a didactic tool based on the creation of comic-style synthetic, colorful, attractive, and appealing copyright-free images using text-to-image AI models in order to increase the curiosity, enthusiasm, and interest of citizens, especially young people, in science and make it more attractive and accessible.

References:
[1] Burnard, P. et al. The Art of co-creating arts-based possibility spaces for fostering STE(A)M practices in primary education, 2017. [2] Post, M. Scientific dissemination via comic strip: a case study with SacreBLEU, 2021.
[3] Oxford English Dictionary, Oxford University Press, 2022.

Keywords:

STEM, deep learning, text-to-image model, DALL-E, Midjourney, Stable Diffusion, comic, young students.