Smooth Modeling of Flows on Graphs
Abstract
The high-level goal of this project is to bridge the gap between theoretical developments in the field of optimal transport---designed to understand flows along smooth domains---and the analysis of signals over graphs. In contrast to existing combinatorial algorithms and models for flows on graphs, our approach is inspired by continuous ideas from differential equations, functional analysis, and geometry. In the end, we propose fundamentally different constructions from well-known graph algorithms that more directly link smooth (continuously-varying in time and space) and discrete interpretations of flows.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 15, 2020
- Accession Number
- AD1204561
Entities
People
- Justin Solomon
Organizations
- Massachusetts Institute of Technology