作品选集
2021—2026Selected Works

作品 / Work

作者不在中心

The author is not at the center

Kant's "Copernican revolution" placed the subject at the center of knowledge: human beings do not merely receive the world; they also organize experience. Artificial intelligence shifts the subject's position again: images and forms no longer arise from human intention alone. Algorithms, data, materials, and systems in operation can also generate unforeseeable forms.

艺术家 / ARTISTbuilds, chooses, assumes responsibility
系统 / SYSTEMgenerates forms
观者 / VIEWERimages, emotions, meanings arise

系列 / Series

镜中递归Recursion in the Mirror

03/082026 · 2024

A.I.O.O.N.

Generative · Self-replicating neural network · Curl noise

Weights from a self-replicating neural network enter a particle system. No flower images enter the system, yet flower-like forms emerge. In 花, 艹 (grass) meets 化 (change): through change, grass regenerates itself. The work explores the autonomy of artificial intelligence and its art.

2026

Video (Alternative )
A.I.O.O.N., 2026: six rotating red and white particle forms
Video (Alternative )
Self-generating words: a complete 39-character relationship diagram

Language does more than describe the world; it can also encode how things exist, change, and come into being. This work traces self-generation through the structures and relationships of existing and newly constructed Chinese characters.

2024

A.I.O.O.N. exhibition view showing two red generative works and research panels
A.I.O.O.N. exhibition view with the artist and visitors
Exhibition view

04/082026 · 2024

S.E.D.I.A.

Generative · Self-replicating neural network
Digital image

In self-replication, a neural network’s own weights are both the material and the object of calculation. Finding a self-replicating state is equivalent to solving a set of equations. Each update changes the input, the network itself, and the target. The stool’s dimensions and structure form another set of equations, through which the network’s internal weights become surface and structure.

05/082026 · 2023

Generation, Operation, Destruction

Generative · Self-replicating neural network · Particle system

Parameters from a self-replicating neural network enter a Perlin-noise flow field, where particles emerge, operate, decay, and disappear before the cycle begins again. Form exists only briefly between generation and destruction.

This work explores different ways for self-replicating neural networks to shape form: using changing weights during training, or using the final weights after training.

2026

Fine lines emerge from a single point, grow and overlap, then gradually fade.

2023

A sequence of a white particle form emerging, operating, decaying, and disappearing

The training process is visualized through the motion of particles.

In collaboration with: Lan Di, Xu Muhan, Li Bing

06/082026 · 2023

M.O.N.A.S.

Generative · Permutation algorithm
Digital image · Dimensions variable

Chinese characters are constructed from squares.

2026

Characters drift, swirl, gather, and disperse in space–time.

2023

The characters form a quotation adapted from Zhuangzi: “Heaven and earth arise together with me; the myriad things and I are one.”

Paper

Method source: Kostas Terzidis, Permutation Design (2015)

07/082023研究 / Research

无题Untitled

Generative · Particle system · Geometric primitives
Real-time digital visual · Dimensions variable

Geometric particles connect, attract, or repel according to distance and random thresholds, generating images that resemble constellations. Adapted from Kostas Terzidis's particle-generation research.

A constellation-like field generated by the particle system

08/082021

人工情绪Artificial Emotions

Robotic arm · Neural network · Acrylic on canvas
60 × 50 cm (each)

The machine paints fields of colour that can stir human emotion, yet it feels none itself. It responds not to human emotion, but to its own previous output.

Artificial Emotions exhibition view with three paintings and a video display
展览现场 / Exhibition view
Robotic arm selecting colour and leaving traces of paint
系统 · 色彩 · 痕迹 / System · Color · Trace

Words → neural network → color and direction → robotic arm → traces of paint

I'm a PhD candidate in Design at Tongji University. My doctoral research treats AI autonomy as a design material, connecting generative systems with the philosophy of technology. From 2025 to 2026, I was a visiting researcher at the University of Turin's Center for Ontology.

代表展览 / Selected exhibitions

2024

A.I.O.O.N.
Tongji University · B4B Art Space

2023

Shijian (The World)
With artist Yu Xingze · 14th Gwangju Biennale Global Metaverse Exhibition · Cheng Shifa Art Museum · Shanglin Art Museum

2022

Artificial Emotions
Tongji University · Zhangjiang Artificial Intelligence Island

研究与联系 / Research & Contact

共同研究 / Collaborative research

My research has grown through long-term discussions and collaboration with Kostas Terzidis and Filippo Fabrocini on artificial intelligence, object-oriented ontology, and authorship.