Hanna, Sean;
(2022)
Urban Complexity.
In: Carta, Silvio, (ed.)
Machine Learning and the City: Applications in Architecture and Urban Design.
(pp. 1-13).
Wiley
Text
Hanna_Chapter 1 images_s.pdf - Submitted Version Access restricted to UCL open access staff until 22 May 2024. Download (1MB) |
Abstract
Cities are arguably the most complex things we have ever built. Machine learning (ML) approaches instead begin with the data, and attempt to discern patterns from within them, with the intent that these patterns will reliably inform decisions we make about the city. For some phenomena, patterns are predictable because they converge with increasing scale or time. Agent-based modelling often owes its effectiveness to the fact that a large population is used. Turner and Penn's exosomatic visual architecture (EVA) agents, for example, are extremely simplified models of pedestrians in space, which make random navigation decisions based on a probability weighted by how far they can see in any given direction. This trait suggests the reason why ML is useful in the context of the complex city, a reason too easily overlooked in the day-to-day training of learning models, which are judged on their success in prediction.
Type: | Book chapter |
---|---|
Title: | Urban Complexity |
ISBN-13: | 9781119815075 |
DOI: | 10.1002/9781119815075.ch1 |
Publisher version: | https://onlinelibrary.wiley.com/doi/10.1002/978111... |
Language: | English |
Additional information: | This version is the author submitted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. |
Keywords: | agent-based modelling, exosomatic visual architecture, learning models, machine learning navigation decisions |
UCL classification: | UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of the Built Environment UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of the Built Environment > The Bartlett School of Architecture UCL > Provost and Vice Provost Offices > UCL BEAMS UCL |
URI: | https://discovery.ucl.ac.uk/id/eprint/10150191 |
Archive Staff Only
View Item |