# De Shanshui Canonical URL: https://baoyangchen.com/works/de-shanshui/ Interactive page: https://baoyangchen.com/#work/de-shanshui Year: 2014-2025 Category: AI + Art Works Medium: Artist-trained machine learning model, image generation, video generation, silk prints ## Proposition An artist-trained machine learning model learns shanshui as image generation in 2014, then returns as video generation in 2025. ## Description De Shanshui begins in 2014 with one of the oldest image forms in Chinese art: mountain and water painting. Chen approaches this tradition by training a machine learning model himself, treating the model not as a style filter but as a way to absorb, fracture, and recompose inherited pictorial language. The work links machine learning to the practice of copying. In Chinese painting, copying is a path of study, memory, and transformation. Here, the trained model becomes a contemporary site of that process. The 2014 images are generated through this trained model and printed on Chinese silk. In 2025, the project returns as video generation, extending the same question from still image into moving image, memory, and synthetic pictorial time. ## Viewer Entry The viewer encounters images that seem to remember landscape without returning to a single historical picture. Forms collide, perspectives blur, seals and fragments drift from their original contexts. The image carries the trace of a tradition that has been learned by another kind of system. ## Apparatus Chen feeds images of classical shanshui into an artist-trained machine learning model. Through training, the model identifies patterns, structures, and visual rules, then generates new compositions. The process extends the logic of copying into algorithmic form. The 2014 images are presented on silk, where the woven surface echoes the grid and texture of digital construction; the 2025 return moves the learned landscape into video generation. ## Sections ### Shanshui as Source The work starts from landscape painting as a cultural structure, not simply a visual style. Mountain, water, atmosphere, seal, composition, and empty space become material for algorithmic transformation. ### Copying and Training Traditional copying teaches through repetition and internalization. The artist-trained machine learning model extends that relation into computation, where inherited forms become learned structures. ### From Image to Video The 2025 return does not replace the 2014 work. It reopens the trained landscape as moving image, allowing generated shanshui to unfold through duration rather than remain fixed as a single composition. ### Image After Landscape The generated images do not restore the past. They loosen it. Familiar elements separate from their original order and return as new pictorial space. ### Silk Surface Printing on silk keeps the work connected to the material history of Chinese painting. The image is digital, but its surface carries another memory. ### Authorship and Model De Shanshui shifts attention from output to process. The trained model becomes part of the artwork's authorship, a compressed field of the artist's choices, rules, and inherited images. ## Images 1. https://baoyangchen.com/original-site-media/de-shanshui/02-6e6d50de-DeShanShui_WuHan_Web.jpg - Shanshui image generated through an artist-trained machine learning model and printed on Chinese silk. 2. https://baoyangchen.com/original-site-media/de-shanshui/03-5ed086ac-BCHEN_10.jpg - Detail of silk texture, where digital image and traditional support meet. 3. https://baoyangchen.com/original-site-media/de-shanshui/04-06778dcb-BCHEN_11.jpg - Multiple generated landscapes show how the model recomposes inherited pictorial rules. 4. https://baoyangchen.com/original-site-media/de-shanshui/01-c70f0505-baoyang_chen_deshanshui_001_01_75.jpg - Seals, landscape forms, and tonal fragments detach from their historical settings and enter a new composition. 5. https://baoyangchen.com/original-site-media/de-shanshui/05-a0139261-baoyang_chen_deshanshui_002_02_32.jpg - Portfolio spread from the De Shanshui section, documenting the project's long-running development. 6. https://baoyangchen.com/original-site-media/de-shanshui/06-7191c70f-baoyang_chen_deshanshui_003_04_69.jpg 7. https://baoyangchen.com/original-site-media/de-shanshui/07-e0f1aa89-baoyang_chen_deshanshui_004_05_12.jpg 8. https://baoyangchen.com/original-site-media/de-shanshui/08-882f5a57-baoyang_chen_deshanshui_005_05_66.jpg 9. https://baoyangchen.com/original-site-media/de-shanshui/09-8a05a84a-baoyang_chen_deshanshui_006_05_106.jpg 10. https://baoyangchen.com/original-site-media/de-shanshui/10-2fdb32eb-baoyang_chen_deshanshui_007_02_77.jpg 11. https://baoyangchen.com/original-site-media/de-shanshui/11-f2d145fa-baoyang_chen_deshanshui_008_05_52.jpg 12. https://baoyangchen.com/original-site-media/de-shanshui/12-3f0b9f55-baoyang_chen_deshanshui_009_02_08.jpg 13. https://baoyangchen.com/original-site-media/de-shanshui/13-d6644056-baoyang_chen_deshanshui_010_05_22.jpg 14. https://baoyangchen.com/original-site-media/de-shanshui/14-85028033-baoyang_chen_deshanshui_011_04_66.jpg 15. https://baoyangchen.com/original-site-media/de-shanshui/15-c7cddaf0-baoyang_chen_deshanshui_012_04_65.jpg 16. https://baoyangchen.com/original-site-media/de-shanshui/16-9578973a-baoyang_chen_deshanshui_013_02_34.jpg 17. https://baoyangchen.com/original-site-media/de-shanshui/17-4f306b76-baoyang_chen_deshanshui_014_02_50.jpg 18. https://baoyangchen.com/original-site-media/de-shanshui/18-c7c2cc3f-baoyang_chen_deshanshui_015_03_15.jpg 19. https://baoyangchen.com/original-site-media/de-shanshui/19-5efc917b-baoyang_chen_deshanshui_016_02_74.jpg 20. https://baoyangchen.com/original-site-media/de-shanshui/20-414da257-baoyang_chen_deshanshui_017_05_60.jpg 21. https://baoyangchen.com/original-site-media/de-shanshui/21-bd868092-baoyang_chen_deshanshui_018_03_62.jpg ## Videos No video listed. ## Source Pages - https://baoyangchen.com/De-Shan-Shui-Works