罗霄作品封面_960×1200

CyberFish – 正面扩散

CyberFish – Front Diffusion

视频, 9:16, 1440×2560
Video, 9:16, 1440×2560
创作年份 Year of creation:2024-2025
罗霄照片_850×1280
罗霄
Luo Xiao
罗霄,算法工程师,数字艺术家。实验编程创始人,从事交互媒体、人工智能、生成艺术的创作与教学。毕业于北京航空航天大学,电子信息工程与软件工程专业背景。20...
Luo Xiao, an creative technologist and digital artist, is the founder of Experimental Programming and is engaged in the creation and teaching of inter...
作品介绍 Works Introduction
“CyberFish – 正面系列” 作于2024年,在初代基础上进化而来,使用生成对抗网络(StyleGAN)进行模型训练和视觉生成,并加入了着色器像素排序算法。这一代CyberFish模型在训练时,使用了基于扩散模型的AI绘画工具Midjourney来创建数据集,这与初代采集自人工插画的数据集完全不同,是对CyberFish向着完整AI智能体进化的探索,也是对当前信息来源和电子垃圾问题的反思。
"CyberFish - Front Series" was created in 2024 and evolved from the first generation. It uses a generative adversarial network (StyleGAN) for model training and visual generation, and adds a shader pixel sorting algorithm. This generation of CyberFish model uses a diffusion model to create a data set during training, which is completely different from the first generation of data sets collected from artificial illustrations. It is an exploration of the evolution of CyberFish towards a complete intelligent agent, and it is also a reflection on the current sources of information and electronic waste.
创作论述 Creative Discourse
CyberFish是始于2021年的机器学习系列作品,作者罗霄将实验艺术家和算法工程师的两套思维模式,以数据集和算法的形式交给神经网络进行学习训练,所产出的AI模型承载了作者本人以及 AI自我的双重意识和技术特性。
CyberFish is a series of machine learning works that began in 2021. The author Luo Xiao gave the two sets of thinking modes of experimental artists and algorithm engineers to the neural network in the form of data sets and algorithms for learning and training. The resulting AI model carries the dual consciousness and technical characteristics of the author and the AI self.