Share Email Print
cover

Proceedings Paper

Diffusion source detection in a network using partial observations
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Diffusive phenomena are ubiquitous in nature and society, and have been extensively studied in various fields, such as natural sciences and engineering. Recently, however, the more challenging inverse problem of diffusion source detection in a network has started to receive a significant amount of attention. A lot of research has concentrated on finding origins in tree-like networks, however these approaches cannot be easily extended to generic networks. Furthermore, only some methods consider realistic temporal diffusion dynamics. We introduce a novel method to localise the source of multiple rumours in an arbitrary network of known topology, using partial observations of the network nodes. We first present two mathematical models of the discrete-time, susceptible infected propagation dynamics, which accurately capture the diffusion process and have low computational complexity. The first one is a simplified likelihood of infection at a node, at a certain time after the rumour is initiated. The second is a formulation of the infection likelihood of a node, as a function of its shortest distance to the source. We then design an efficient single source detection algorithm, which leverages these mathematical models of diffusion, and the assumption that the start time of the propagation is known. Finally, we show how these methods can be extended to the case when the start time of the rumour is unknown, by taking advantage of the dissimilarity in dynamics of infection, of different nodes in the network. Simulation results show that a high source estimation probability is achieved using a small number of observations.

Paper Details

Date Published: 9 September 2019
PDF: 15 pages
Proc. SPIE 11138, Wavelets and Sparsity XVIII, 111380L (9 September 2019);
Show Author Affiliations
Roxana Alexandru, Imperial College London (United Kingdom)
Pier Luigi Dragotti, Imperial College London (United Kingdom)


Published in SPIE Proceedings Vol. 11138:
Wavelets and Sparsity XVIII
Dimitri Van De Ville; Manos Papadakis; Yue M. Lu, Editor(s)

© SPIE. Terms of Use
Back to Top
PREMIUM CONTENT
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?
close_icon_gray