# Multimodal Network Alignment

## TL;DR

This paper introduces a novel method for aligning multimodal networks by directly computing a low-rank matrix factorization, enabling efficient approximate maximum weight matchings, demonstrated on synthetic and real transportation networks.

## Contribution

It presents a new low-rank factorization approach for multimodal network alignment and methods for approximate maximum weight matchings, advancing network analysis techniques.

## Key findings

- Effective on synthetic networks
- Successfully de-anonymized transportation network
- Outperforms existing alignment methods

## Abstract

A multimodal network encodes relationships between the same set of nodes in multiple settings, and network alignment is a powerful tool for transferring information and insight between a pair of networks. We propose a method for multimodal network alignment that computes a matrix which indicates the alignment, but produces the result as a low-rank factorization directly. We then propose new methods to compute approximate maximum weight matchings of low-rank matrices to produce an alignment. We evaluate our approach by applying it on synthetic networks and use it to de-anonymize a multimodal transportation network.

---
Source: https://tomesphere.com/paper/1703.10511