I am Wei-Zhuo Zhang , an independent AI researcher based in China, currently associated with Tianjin College of Commerce and the Geometric-Dynamic AI Fundamental Theory Laboratory. My research explores the intersection of differential geometry, cognitive science, and AI alignment.
My current work focuses on Kansei Geometric Dynamics—a framework that treats affect and intentionality not as statistical noise, but as an endogenous dynamical driver that steers semantic inference. In my recent preprint, “Affect as a Formal Description of Intelligence Dynamics” , I introduced the UGF model to explain how large language models can dynamically bypass semantic traps through geometric curvature.
I am highly interested in discussing how gauge theory and continuous dynamical models can provide a more rigorous, mechanistic understanding of LLM steerability, hallucination dialectics, and in-context learning.
Connect with me:
X (Twitter): @WeiZhuo_GeoDyn
ORCID: 0009-0006-1856-6514
Email: zhizhuplayer@gmail.com