Gradland：用梯度与矩阵刻画多维现象体验
原标题：Gradland: On Phenomenal Experience, Differentiated Across Many Dimensions
AI 摘要
这篇 arXiv 论文提出「Gradland」框架，主张用微分、矩阵乘法与对角化等应用数学工具来研究现象体验（phenomenal experience）如何「联结」在一起。作者提出「假设 G」：物质的一阶结构刻画了心智的结构，并引入「透明性」与「内聚性」两个概念来量化体验。该方法源自作者与 Tononi 共同开发的整合信息论 v2，但改用梯度与矩阵而非信息论，理由是梯度工具更廉价、更灵活、可扩展性更强。
正文节选
tcboxmath \tl_set:Ne\tcbhighmathtcbhighmath Gradland: On Phenomenal Experience, Differentiated Across Many Dimensions 1 Introduction How do things hang together [1]? To tackle the question for physical phenomena, Leibniz and Newton invented calculus and initiated the most successful research program in history: studying nature via differential equations. Centuries later the branch of that research program called applied mathematics has three main workhorses: linearization (differentiation), agg