高逸菲创作发展年表

高逸菲从 2019 年开始对艺术感兴趣。在此之前，她是一位对艺术完全不感兴趣的青少年。这几年的作品被视为高逸菲创作中的前史。

Gao Yifei became interested in art in 2019. Before then, she had been a teenager with virtually no interest in art. The works from these years are regarded as the prehistory of her artistic practice.

这一时期，她跟随具有印象派背景的老师学习，系统临摹过约十件威廉·透纳与克劳德·莫奈的作品。这些作品为她留下了两个长期存在的视觉基因。透纳是高逸菲早期最喜欢的艺术家之一。

During this period, she studied with a teacher trained in an Impressionist tradition and systematically copied around ten works by J. M. W. Turner and Claude Monet. These studies left two visual genes that would remain present throughout her later work. Turner was one of Gao Yifei's favorite artists in her early years.

她喜欢透纳作品中大气、雄伟而宏大的感觉。高逸菲后来即使创作纯抽象图像，画面依然经常具有风景画的观看方式：旋涡、气流、地平线、中心光源与远近层次。她并不热衷于细小、私密、日记式的情绪，而更偏爱具有规模、力量和压倒性的视觉经验。

She was drawn to the atmospheric, majestic, and expansive quality of his paintings. Even when Gao Yifei later created purely abstract images, they often retained the viewing logic of landscape painting: vortices, currents of air, horizons, central sources of light, and layered depth. She is less interested in small, private, diary-like emotions than in visual experiences defined by scale, force, and overwhelming intensity.

透纳为她的艺术生涯带来了宏大构图与场域意识。莫奈与印象派留下的则是颜色的密集并置。她喜欢让大量细碎而缤纷的色彩彼此堆积、碰撞，并在观看中形成整体。

Turner introduced into her practice a sense of monumental composition and immersive spatial fields. Monet and Impressionism, meanwhile, left her with a taste for dense juxtapositions of color. She likes to let numerous fragmented and vivid colors accumulate and collide, forming a unified whole through the act of viewing.

2024 年初，高逸菲进入芝加哥艺术学院。进入学院后，她的创作迅速从传统再现逐渐转向表现性变形与战后具象绘画的语言。那段时间，她最喜欢的艺术家是弗朗西斯·培根与意大利艺术家古列尔莫·卡斯泰利。

At the beginning of 2024, Gao Yifei entered the School of the Art Institute of Chicago. After entering the school, her practice quickly shifted from traditional representation toward expressive distortion and the language of postwar figurative painting. At that time, her favorite artists were Francis Bacon and the Italian artist Guglielmo Castelli.

上半年间，高逸菲创作了约三件受到这两位艺术家影响的作品。这个阶段仍然可以明显看到外部艺术语言的痕迹，但它的意义并不在于建立最终风格，而在于让她从传统再现进入一种更加现代、扭曲而具有压迫感的图像结构。

During the first half of the year, Gao Yifei created around three works influenced by these two artists. Traces of external artistic languages remained clearly visible in this phase, but its significance did not lie in establishing a final style. Rather, it allowed her to move from traditional representation toward a more modern, distorted, and oppressive image structure.

与很多长期停留在一种成熟风格中的艺术家不同，高逸菲并不喜欢依赖重复来确认已经掌握的语言。她往往在一套方法刚刚成立后，便开始继续向外探索。在吸收外部影响后，高逸菲进入了一段较为开放的探索期。

Unlike many artists who remain within one mature style for an extended period, Gao Yifei does not like to rely on repetition to confirm a language she has already mastered. As soon as a method begins to work, she tends to continue exploring beyond it. After absorbing these external influences, Gao Yifei entered a more open-ended period of exploration.

这一时期，她广泛关注并吸收了约五至八位不同艺术家的视觉语言，其中既有知名艺术家，也有尚未进入主流视野的创作者。这一阶段的六件重要作品在外观上并不统一，风格显得混乱。但这段混乱承担了重要作用：她开始有意识地摆脱已有艺术家的语言，寻找自己的视觉偏好究竟由什么组成，并尝试向更加现代和当代的语言跃进。

During this period, she closely followed and absorbed the visual languages of roughly five to eight different artists, including both established figures and creators who had not yet entered the mainstream. The six major works from this phase were visually inconsistent and appeared stylistically chaotic. Yet this chaos played an important role: she began consciously separating herself from the languages of existing artists, identifying the components of her own visual preferences, and attempting to leap toward a more modern and contemporary vocabulary.

第一次重大突破：高逸菲 1.0

2024 年末，高逸菲完成了一件可以被视为个人风格真正开端的作品。这件作品开启了“高逸菲 1.0”。高逸菲 1.0 的核心特征包括：

At the end of 2024, Gao Yifei completed a work that can be regarded as the true beginning of her personal style. This work marked the beginning of “Gao Yifei 1.0.” The defining features of Gao Yifei 1.0 include:

在 1.0 出现之后，高逸菲没有立即连续制作大量相似图像。她再次进入探索期，并开始将注意力转向材料。她在学校修读了一门时装设计系课程，接触了织物塑形，并学习通过折叠、缝合、拉伸与塑形改变织物结构。

After the emergence of 1.0, Gao Yifei did not immediately produce a large number of similar images. She entered another period of exploration and began turning her attention toward materials. She took a course in the Fashion Design Department, where she encountered fabric manipulation and learned to alter textile structures through folding, sewing, stretching, and shaping.

这一年除去少量平面创作，高逸菲主要进行各种材料试验，其中数量最多的是与织物塑形有关的实验。布料不再只是图像的承载面，而被折叠、操控和变形，成为作品结构的一部分。随后，她又使用了亮片、透明材料、表面加工及其他综合媒介。

Apart from a small number of two-dimensional works, Gao Yifei spent most of the year conducting material experiments, especially experiments involving manipulated fabric. Fabric was no longer merely a surface carrying an image; it was folded, manipulated, and transformed into part of the work's structure. She later incorporated sequins, transparent materials, surface treatments, and other mixed media.

这一阶段的视觉仍建立在高逸菲 1.0 的基础上，但材料与工序变得高度实验性。她开始研究：图像如何成为物体，表面如何产生视觉欺骗，透明、反射和覆盖如何制造空间，材料能否像数字图层一样被组合，以及平面图像如何进入真实空间。高逸菲在这一年完成了多件综合材料与技术实验作品，其中也夹杂着一至两件延续 1.0 语言的纯图像作品。

The visual language of this phase remained grounded in Gao Yifei 1.0, while its materials and processes became highly experimental. She began investigating how images could become objects, how surfaces could produce visual deception, how transparency, reflection, and layering could create space, whether materials could be combined like digital layers, and how two-dimensional images could enter physical space. During this year, Gao Yifei completed multiple mixed-media and technical experiments, alongside one or two purely image-based works that continued the language of 1.0.

从后来的发展看，材料时期并非支线。它为 2.0 中的金属感、透明折面、层叠空间和混种材质提供了重要基础。2025 年下半年，在持续进行材料试验的同时，她还初步接触了艺术与科技，学习了 p5.js 和 Arduino，并开始自行探索人工智能。

In retrospect, the material period was not a side branch. It provided an important foundation for the metallic qualities, transparent facets, layered spaces, and hybrid materials of 2.0. In the second half of 2025, while continuing her material experiments, she also began engaging with art and technology, studied p5.js and Arduino, and started independently exploring artificial intelligence.

她完成了一系列以技术学习和方法测试为主的实验，其中多数尚未被她视为正式作品。

She completed a series of experiments focused primarily on technical learning and methodological testing, most of which she did not yet regard as finished works.

在高逸菲 1.0 风格与此前材料试验的基础上，她继续创作具有相近视觉语言、采用不同材料的作品。这些作品越来越接近装置。随着材料实验继续发展，她开始接触代码、生成图像、交互、工作流与人工智能图像生产工具。

Building on the language of Gao Yifei 1.0 and her earlier material experiments, she continued creating works that shared a related visual vocabulary while using different materials. These works moved increasingly closer to installation. As her material experiments developed, she began working with code, generative images, interaction, workflows, and AI-based image-production tools.

此时，她的兴趣已经不仅在于完成一件作品，也开始转向一个更底层的问题：

At this point, her interest no longer lay only in completing an individual work. It also turned toward a more fundamental question:

于是，《制作图像的最短路径》，简称 SWI，诞生了。项目全名为《制作图像的最短路径：人工智能时代从生成到生产的图像生产研究》。项目以她自己的约 40 件过往作品作为训练材料，通过多轮 LoRA 训练，尝试将个人视觉经验转化为可被模型识别和调用的风格结构。

This question led to The Shortest Way to Make Images, abbreviated as SWI. The project's full title is The Shortest Way to Make Images: A Study of Image Production in the Age of AI, from Generation to Production. Using around forty of her previous works as training material, the project employed multiple rounds of LoRA training in an attempt to transform her personal visual experience into a stylistic structure that a model could recognize and invoke.

由于这些作品横跨不同阶段，风格一致性有限，模型最初无法准确识别她的图像语言。经过五次训练后，生成结果逐渐接近她此前作品中的部分特征，尤其捕捉到了布料折叠、表面变形和密集结构等视觉元素。随后，她通过 ComfyUI 搭建节点式生成工作流，将个人 LoRA、基础模型、提示词和图像处理步骤连接起来，形成一套可重复运行的图像生成流程。

Because these works spanned different phases and lacked strong stylistic consistency, the model was initially unable to identify her visual language accurately. After five training iterations, the generated results gradually began to resemble certain features of her earlier work, especially folded fabric, surface deformation, and dense structures. She then used ComfyUI to construct a node-based generative workflow, connecting her personal LoRA, a base model, prompts, and image-processing steps into a repeatable image-generation process.

她从系统产出的图像中进行筛选，并将其中一幅生成图像委托给一位中国画家，以手工绘画的方式完成实体转译。图像由模型生成进入人工制作，最终转化为实体绘画。项目最终被整合进她的个人网站，并以开源形式公开。

She selected images from the system's outputs and commissioned a Chinese painter to translate one generated image into a physical, hand-painted work. The image thus moved from model generation into human production and ultimately became a physical painting. The project was ultimately integrated into her personal website and released as an open-source system.

由此，《制作图像的最短路径》不再只是一组人工智能生成图像，而成为一套从个人作品训练、模型生成、作者筛选、人工绘制到公开传播的完整生产系统。这一项目也是高逸菲创作中的重要转折。她第一次将自己的视觉语言视为一种可以被拆解、训练、调用和重新生产的结构，也第一次把“制作作品”推进到“制作能够生产作品的工作流”。

As a result, The Shortest Way to Make Images was no longer merely a collection of AI-generated images. It became a complete production system encompassing personal-work training, model generation, authorial selection, manual painting, and public distribution. The project also marked an important turning point in Gao Yifei's practice. For the first time, she treated her visual language as a structure that could be dismantled, trained, invoked, and reproduced. She also shifted from “making works” to “making a workflow capable of producing works.”

这一工作流在当时仍高度实验性，只能勉强运行。LoRA 训练的不稳定，也暴露了她的作品难以被压缩为单一风格标签的问题。从这里开始，高逸菲的创作开始同时拥有两条互相连接的路径：

At the time, the workflow remained highly experimental and could only barely function. The instability of the LoRA training also revealed how difficult it was to compress her work into a single stylistic label. From this point onward, Gao Yifei's practice began to follow two interconnected paths:

在经历材料、故障艺术、综合媒介和艺术与科技的扩张后，高逸菲重新回到了纯粹的图像制作。这次回归并非回到原点。她带回了此前所有阶段积累的能力：印象派式的色彩密度、透纳式的宏大场域、培根式的扭曲、材料阶段的表面意识、故障艺术的图像故障、3D 元素、重复结构，以及技术艺术中的系统思维。

After expanding into materials, glitch, mixed media, and art and technology, Gao Yifei returned to pure image-making. This return was not a return to the starting point. She brought back the capacities accumulated across every earlier phase: Impressionist color density, Turner-like monumental fields, Bacon-like distortion, the material period's attention to surface, the visual disruptions of glitch, 3D elements, repetitive structures, and the systems thinking of technological art.

2026 年 7 月，高逸菲 2.0 出现。这一阶段由一件作品开启。她将这段时间掌握的视觉方法几乎全部投入同一张图像：重复结构、摩尔纹、金属、透明折面、网格、尖刺、虹彩物体与具象异物共同进入画面。

In July 2026, Gao Yifei 2.0 emerged. The phase began with a single work. She poured nearly every visual method she had mastered during this period into one image: repetitive structures, moiré patterns, metal, transparent facets, grids, spikes, iridescent objects, and figurative anomalies all entered the same pictorial field.

高逸菲 2.0 的核心并不是突然更换风格，而是第一次真正实现了 1.0 时期已经存在的科技欲望。1.0 依靠喷溅、色彩、拖影和材质效果逼近科技感。2.0 则让科技感进入画面的骨架、空间和运行规则。

The core of Gao Yifei 2.0 is not a sudden change of style, but the first full realization of a technological desire already present in 1.0. 1.0 approached technological aesthetics through splashes, color, motion trails, and material effects. 2.0 allows technological aesthetics to enter the image's skeleton, space, and operational logic.

它的主要特征包括：

回看她的发展，高逸菲的作品并没有真正经历彼此无关的风格切换。她一直在追求同一种感受： 画面必须足够大、足够满、足够强，信息几乎超出眼睛能够立即处理的范围，但内部仍然保持秩序。

Looking back across her development, Gao Yifei's work has not truly undergone a series of unrelated stylistic shifts. She has continually pursued the same sensation: The image must be large enough, full enough, and forceful enough that its information nearly exceeds what the eye can process immediately, while still retaining an internal order.

她的混乱是经过组织的混乱。重复、密度、颜色、材质、空间和视觉重心都受到理性判断的控制。她从透纳那里继承了规模，从莫奈那里继承了色彩颗粒，从培根那里继承了扭曲，从材料实验中获得表面和物体性，又从艺术与科技中获得规则、程序与系统。

Her chaos is organized chaos. Repetition, density, color, material, space, and visual emphasis are all governed by rational judgment. She inherited scale from Turner, particles of color from Monet, distortion from Bacon, surface and objecthood from material experimentation, and rules, programs, and systems from art and technology.

这条顺序并非巧合。从写实的观看训练，到印象派的色彩感知，到战后表现主义的形体扭曲，到材料与综合媒介的物性探索，再到技术与系统的介入——这恰好是艺术史本身走过的路径，只是被压缩进了一个人六年的创作生命里。

This sequence is not a coincidence. From the observational training of realism, to the color perception of Impressionism, to the bodily distortion of postwar expressionism, to the material inquiry of mixed media, to the intervention of technology and systems — this is, in essence, the path art history itself has traveled, only compressed into a single life across six years.

这些影响都没有被保留为原样，而是在之后的作品中不断变形、叠加和吞并。随着她继续学习 TouchDesigner，她的静态视觉可能进一步获得时间、运动、声音、实时反馈和交互。未来的变化未必意味着 2.0 视觉迅速消失。

None of these influences remained intact. They were continually transformed, layered, and absorbed into later works. As she continues learning TouchDesigner, her static visual language may further acquire time, movement, sound, real-time feedback, and interaction. Future changes will not necessarily mean the rapid disappearance of the 2.0 visual language.

更可能发生的情况是，这套语言将从静态图像进入动态系统。

More likely, this language will move from static images into dynamic systems.