2026 AI 训练行旅
2026 AI Training Journey
2026
人工智能影像、双屏影像
4 分 57 秒
世界模型
强化学习
人类动作
人工智能影像、双屏影像
4 分 57 秒
世界模型
强化学习
人类动作
AI的训练之旅,实则是对人类现实的巡览;而我们的每一次日常行为,即是我们亲自为AI装好的行李。发生在AI背后的对人类日常行为进行解码与数据剥离的事件,也将以“现实↔数据集”的结构不断循环。
作品通过对无人机宏观测绘、仿真环境强化学习以及缝纫工第一视角等现实切片的并置视觉实验,讨论人与AI中间“中转”地带。
2026
AI Video, Dual-channel Vide
4 mins 57 s
World Models
Reinforcement Learning
Human Movements
AI Video, Dual-channel Vide
4 mins 57 s
World Models
Reinforcement Learning
Human Movements
The training process of AI is actually a tour through human reality, and our everyday actions are the very luggage we pack for the machines ourselves. Behind the screen, the decoding and data stripping of human behavior will continuously loop within the structure of "Reality ↔ Dataset."
By juxtaposing visual slices of reality—drone mapping, simulation reinforcement learning, and the first-person view of sewing workers—this experiment explores the transitional "transit" zone between humans and AI.