Collaboration of Digital Literacy and Artificial Intelligence in Preserving Traditional Malay House Culture

¹ Universitas Ahmad Dahlan, Yogyakarta, Indonesia
² Sekolah Tinggi Arsitektur YKPN, Yogyakarta, Indonesia

^* Corresponding author: sri.winiarti@tif.uad.ac.id

Abstract

The community needs digital literacy as part of learning. Digital literacy is a concept and practice that discusses expertise in mastering technology and using digital media to carry out mediation processes on digital media. Digital literacy can be used in learning and to preserve culture. This study tries to create a model for maintaining culture by utilizing Artificial Intelligence (AI) technology as part of digital literacy. Digital literacy has created a digital mediation process to identify similarities in traditional Malay buildings. The digital mediation process to identify traditional Malay buildings uses four components: digital skills, culture, ethics, and safety. In making this AI-based digital literacy, data is used as photo data of traditional Malay buildings, with as many as 750 images taken directly in Pekanbaru, Riau, Indonesia, using smartphones and digital cameras. The data taken is the dominant characteristics of traditional Malay buildings, as many as four features: ornaments, roof shape, window shape, and pillars. The data has been processed using data processing techniques in AI: scale, cropping, resizing, and brightness to improve image quality. Next, an architectural model with a convolutional neural network will be created as one of the architectural models used in deep learning. Deep learning is a scientific part of Artificial Intelligence. The data has been validated by a cultural architect and a Riau Malay community figure, both from the perspective of Malay cultural philosophy and building characteristics. Experiments have been conducted in training and data validation to produce this AI-based digital literacy machine. The experiments produced an AI machine that can be used for digital literacy and can identify Malay building objects with an accuracy of 96.92% using the confusion matrix accuracy method.