佛GAN
In traditional Buddhist statue carving, visuality and orality serve as heritage vehicles for the transmission of memory, deeply embedded in generations of artisans' empiricism and instrumental rationality. This has constrained the development of Buddhist sculptural art.
Therefore, we aim to break these empirical limitations by introducing machine learning technology.The recursive learning process of StyleGAN will be applied to the generation and reconstruction of traditional Buddhist statue-making techniques, capturing and reproducing the artisans' tacit knowledge.
Turing memory, as a cryptic medium within the heritage of craftsmanship, is fine-tuned into temporality, shadow, texture, and appearance, inscribed onto the facial structures of Buddha statues. Based on nonlinear Moiré patterns, its dynamic carving technique becomes a multi-layer perceiver. Similarly, the bodily techniques of artisans function as a physical discriminator for GANs (Generative Adversarial Networks) in the real world, breaking free from the constraints of traditional empiricism.
在传统佛像雕刻中，视觉、口传性作为传承记忆的载体，深深地根植于世代工匠的经验主义和工具理性之中，使得佛教雕刻艺术的发展。因此，我们希望通过引入机器学习技术来打破这些经验上的限制。StyleGAN 的递归学习过程将被应用于传统佛教造像技艺的生成和重构，捕捉并再现工匠们的隐性知识。图灵记忆作为手工艺传承中的隐性媒介，被微调为时间性、阴影、纹理和外观，并被刻在佛像的面部结构上。基于非线性摩尔纹，其动态雕刻技术成为多层次的感知器。同样，在现实世界中，工匠的身体技术也可以作为生成对抗网络（GANs）的物理判别器，摆脱传统经验主义的束缚。