Cheeger Inequalities for Directed Graphs and Hypergraphs using Reweighted Eigenvalues
Lap Chi Lau, Kam Chuen Tung, Robert Wang
Abstract
We derive Cheeger inequalities for directed graphs and hypergraphs using the reweighted eigenvalue approach that was recently developed for vertex expansion in undirected graphs [OZ22, KLT22, JPV22]. The goal is to develop a new spectral theory for directed graphs and an alternative spectral theory for hypergraphs. The first main result is a Cheeger inequality relating the vertex expansion ψ(G) of a directed graph G to the vertex-capacitated maximum reweighted second eigenvalue λ v * 2 that where ∆ is the maximum degree of G. This provides a combinatorial characterization of the fastest mixing time of a directed graph by vertex expansion, and builds a new connection between reweighted eigenvalued, vertex expansion, and fastest mixing time for directed graphs. The second main result is a stronger Cheeger inequality relating the edge conductance φ(G) of a directed graph G to the edge-capacitated maximum reweighted second eigenvalue λ e * 2 that λ e * 2 φ(G) λ e * 2 • log
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext f13ffd56-6023-40b4-8ee8-a7403a3b8369Cited by top-tier papers3
- Fast Algorithms for Directed Graph Partitioning Using Flows and Reweighted EigenvaluesLap Chi Lau, Kam Chuen Tung, Robert WangSODA 2024 · 2 citations
- Fast Algorithms for Hypergraph PageRank with Applications to Semi-Supervised LearningKonstantinos Ameranis, Adela Frances DePavia, Lorenzo Orecchia, Erasmo TaniICML 2024 · 1 citation
- New Approximation Bounds for Small-Set Vertex ExpansionSuprovat Ghoshal, Anand LouisSODA 2024
Builds on4
- Higher-Order Spectral Clustering of Directed GraphsSteinar Laenen, He SunNeurIPS 2020 · 33 citations
- Near-linear Size Hypergraph Cut SparsifiersYu Chen, Sanjeev Khanna, Ansh NagdaFOCS 2020 · 15 citations
- Spectral Hypergraph Sparsifiers of Nearly Linear SizeMichael Kapralov, Robert Krauthgamer, Jakab Tardos, Yuichi YoshidaFOCS 2021 · 14 citations
- Cheeger Inequalities for Vertex Expansion and Reweighted EigenvaluesTsz Chiu Kwok, Lap Chi Lau, Kam Chuen TungFOCS 2022 · 5 citations
Related papers
- Higher-Order Cheeger Inequality for Partitioning with BuffersKonstantin Makarychev, Yury Makarychev, Liren Shan, Aravindan VijayaraghavanSODA 2024
- Edge Expansion and Spectral Gap of Nonnegative MatricesJenish C. Mehta, Leonard J. SchulmanSODA 2020 · 3 citations
- Towards tight bounds for spectral sparsification of hypergraphsMichael Kapralov, Robert Krauthgamer, Jakab Tardos, Yuichi YoshidaSTOC 2021 · 18 citations
- Faster Algorithms for Global Minimum Vertex-Cut in Directed GraphsJulia Chuzhoy, Ron Mosenzon, Ohad TrabelsiSODA 2026
- Higher-order Clustering in Complex Heterogeneous NetworksAldo G. Carranza, Ryan A. Rossi, Anup Rao, Eunyee KohKDD 2020 · 27 citations
