1
From Layers to Networks: Comparing Neural Representations via Diffusion Geometry
利用扩散几何从多视角学习比较神经表征，揭示网络与层间相似性，关键创新在于将多视图工具引入表征分析。
arXiv:2605.15901v1 Announce Type: new Abstract: Diffusion geometry is a manifold learning framework that uses random walks defined by Markov transitio…