Mixing Time Matters: Accelerating Effective Resistance Estimation via Bidirectional Method
Abstract
We study the problem of efficiently approximating the \textit{effective resistance} (ER) on undirected graphs, where ER is a widely used node proximity measure with applications in graph spectral sparsification, multi-class graph clustering, network robustness analysis, graph machine learning, and more. Specifically, given any nodes and in an undirected graph , we aim to efficiently estimate the ER value between nodes and , ensuring a small absolute error . The previous best algorithm for this problem has a worst-case computational complexity of , where the value of depends on the mixing time of random walks on , , and , denote the degrees of nodes and , respectively. We improve this complexity to , achieving a theoretical improvement of over previous results. Here, denotes the number of edges. Given that is often very large in real-world networks (e.g., ), our improvement on is significant, especially for real-world networks. We also conduct extensive experiments on real-world and synthetic graph datasets to empirically demonstrate the superiority of our method. The experimental results show that our method achieves a to speedup in running time while maintaining the same absolute error compared to baseline methods.
Cite
@article{arxiv.2503.02513,
title = {Mixing Time Matters: Accelerating Effective Resistance Estimation via Bidirectional Method},
author = {Guanyu Cui and Hanzhi Wang and Zhewei Wei},
journal= {arXiv preprint arXiv:2503.02513},
year = {2025}
}
Comments
Technical Report. Full Paper Accepted by KDD 2025 (August Cycle)