Home // ICONS 2020, The Fifteenth International Conference on Systems // View article

Teaching Machines to Understand Urban Networks

Authors:
Maria Coelho
Mark Austin

Keywords: Systems Engineering; Machine Learning; Graph Embeddings.

Abstract:
Next-generation urban systems will be enabled by technological (cyber) advances deeply embedded within the physical domain.Thevolumeandvarietyofcollecteddatainyearstoc me is only going to grow and diversify, making the task of urban system design and management much more difficult than in the past. We believe these challenges can be addressed by teaching machines to understand urban networks. This paper explores opportunities for using recently developed graph embedding procedures to encode the structure and associated network attributes as low-dimensional vectors. These embeddings can be later used to advance various learning tasks. We exercise the proposed approach on a problem involving identification of leaks in an urban water distribution system. The Dynamic Attributed Network Embedding (DANE) framework is used to generate low-dimensional vectors for a water distribution network, whose pressure attributes are simulated with EPANET. The embeddings are then fed to a Random Forest algorithm trained to identify water leaks.

Pages: 37 to 42

Copyright: Copyright (c) IARIA, 2020

Publication date: February 23, 2020

Published in: conference

ISSN: 2308-4243

ISBN: 978-1-61208-771-9

Location: Lisbon, Portugal

Dates: from February 23, 2020 to February 27, 2020