通用创意智能实验室 Creative General Intelligent Lab
» 探索机器创造力的边界：从“机械生成”到“设计思维” Exploring the Boundaries of Machine Creativity: From Mechanical Generation to Design Thinking
从“理解人类创作”出发，我们旨在打破传统 AI 的局限，融合顶层设计思维、类脑启发与科学赋能，研究人类创作意图与认知机制，赋予 AI 类似导演或设计师的创造性智能。
Starting from an understanding of human creation, we aim to move beyond the limits of traditional AI. We combine high-level design thinking, brain-inspired insights, and scientific advances to study creative intent and cognition, giving AI creative intelligence akin to that of a director or designer.
- 跨模态美学对齐深度对齐文字、音乐、色彩、情感与动态效果，赋予 AI 美学感知、创作推理与全景构思能力。
- 可解释性与轻量化从微观、中观到宏观多尺度剖析大模型“黑盒”机理，以机理解析指导架构优化，实现“以小博大”的高效轻量计算。
- 类脑启发与双向赋能内化探索：将创造力神经机制转化为可计算模型，研发下一代创意智能架构。外哺科研：利用生成式 AI 反哺神经科学，辅助攻克神经连接、疾病机理、创造力神经机制等重大科学难题。
- Cross-modal Aesthetic AlignmentAlign text, music, color, emotion, and motion to equip AI with aesthetic perception, creative reasoning, and the ability to conceive works as a whole.
- Interpretability and EfficiencyAnalyze the black-box mechanisms of large models across micro, meso, and macro scales. Use these insights to optimize architectures and achieve efficient, lightweight computation that does more with less.
- Brain-inspired Insights and Two-way EmpowermentInward exploration: Turn the neural mechanisms of creativity into computable models to develop next-generation creative intelligence architectures.Outward scientific impact: Use generative AI to advance neuroscience and help tackle major questions about neural connectivity, disease mechanisms, and the neural basis of creativity.
- 跨模态美学 Cross-modal Aesthetics
- 设计思维 Design Thinking
- 可解释性 Interpretability
- 轻量化 Efficiency
- 类脑启发 Brain-inspired Research
- 神经科学 Neuroscience
