Papers
Topics
Authors
Recent
Search
2000 character limit reached

On the Diffusion Geometry of Graph Laplacians and Applications

Published 9 Nov 2016 in math.SP, math-ph, math.AP, math.MP, and stat.ML | (1611.03033v1)

Abstract: We study directed, weighted graphs G=(V,E)G=(V,E) and consider the (not necessarily symmetric) averaging operator (Lu)(i)=−∑j∼ipij(u(j)−u(i)), (\mathcal{L}u)(i) = -\sum_{j \sim_{} i}{p_{ij} (u(j) - u(i))}, where pijp_{ij} are normalized edge weights. Given a vertex i∈Vi \in V, we define the diffusion distance to a set B⊂VB \subset V as the smallest number of steps dB(i)∈Nd_{B}(i) \in \mathbb{N} required for half of all random walks started in ii and moving randomly with respect to the weights pijp_{ij} to visit BB within dB(i)d_{B}(i) steps. Our main result is that the eigenfunctions interact nicely with this notion of distance. In particular, if uu satisfies Lu=λu\mathcal{L}u = \lambda u on VV and B={i∈V:−ε≤u(i)≤ε}≠∅, B = \left\{ i \in V: - \varepsilon \leq u(i) \leq \varepsilon \right\} \neq \emptyset, then, for all i∈Vi \in V, dB(i)log⁡(1∣1−λ∣)≥log⁡(∣u(i)∣∥u∥L∞)−log⁡(12+ε). d_{B}(i) \log{\left( \frac{1}{|1-\lambda|} \right) } \geq \log{\left( \frac{ |u(i)| }{\|u\|_{L^{\infty}}} \right)} - \log{\left(\frac{1}{2} + \varepsilon\right)}. dB(i)d_B(i) is a remarkably good approximation of ∣u∣|u| in the sense of having very high correlation. The result implies that the classical one-dimensional spectral embedding preserves particular aspects of geometry in the presence of clustered data. We also give a continuous variant of the result which has a connection to the hot spots conjecture.

Citations (20)

Summary

No one has generated a summary of this paper yet.

Paper to Video (Beta)

No one has generated a video about this paper yet.

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Continue Learning

We haven't generated follow-up questions for this paper yet.