CaltechAUTHORS
  A Caltech Library Service

A Fresh Perspective: Learning to Sparsify for Detection in Massive Noisy Sensor Networks

Faulkner, Matthew and Liu, Annie H. and Krause, Andreas (2013) A Fresh Perspective: Learning to Sparsify for Detection in Massive Noisy Sensor Networks. In: IPSN '13 Proceedings of the 12th international conference on Information processing in sensor networks. ACM , New York, NY, pp. 7-18. ISBN 978-1-4503-1959-1. https://resolver.caltech.edu/CaltechAUTHORS:20130422-142151892

Full text is not posted in this repository. Consult Related URLs below.

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20130422-142151892

Abstract

Can one trade sensor quality for quantity? While larger networks with greater sensor density promise to allow us to use noisier sensors yet measure subtler phenomena, aggregating data and designing decision rules is challenging. Motivated by dense, participatory seismic networks, we seek efficient aggregation methods for event detection. We propose to perform aggregation by sparsification: roughly, a sparsifying basis is a linear transformation that aggregates measurements from groups of sensors that tend to co-activate, and each event is observed by only a few groups of sensors. We show how a simple class of sparsifying bases provably improves detection with noisy binary sensors, even when only qualitative information about the network is available. We then describe how detection can be further improved by learning a better sparsifying basis from network observations or simulations. Learning can be done offline, and makes use of powerful off-the-shelf optimization packages. Our approach outperforms state of the art detectors on real measurements from seismic networks with hundreds of sensors, and on simulated epidemics in the Gnutella P2P communication network.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
http://dx.doi.org/10.1145/2461381.2461387DOIUNSPECIFIED
http://dl.acm.org/citation.cfm?doid=2461381.2461387PublisherUNSPECIFIED
ORCID:
AuthorORCID
Krause, Andreas0000-0001-7260-9673
Additional Information:Copyright 2013 ACM. The authors would like to thank their Caltech collaborators working on the Community Seismic Network project: Prof. Robert Clayton and Dr. Richard Guy of Geophysics; Prof. Thomas Heaton, Dr. Monica Kohler, and Ming-Hei Cheng from Earthquake Engineering; Prof. Mani Chandy and Michael Olson from Computer Science; Dr. Julian Bunn, Dr. Michael Aivazis, and Leif Strand from the Center for Advanced Computing Research. Special thanks to Prof. Robert Clayton and NodalSeismic Inc. for the Long Beach array data set and Prof. Masumi Yamada and NIED for the Japan data set. This research is supported in part by a grant from the Betty and Gordon Moore Foundation, by NSF award CNS0932392 and ERC StG 307036.
Funders:
Funding AgencyGrant Number
Betty and Gordon Moore FoundationUNSPECIFIED
NSFCNS0932392
European Research Council (ERC)StG 307036
Subject Keywords:Algorithms, Experimentation, Theory, Sparsifying transformation, basis learning, sensor networks, community sensing, event detection, ICA, SLSA
Classification Code:C.2.1 [Computer-Communication Networks]: Network Architecture and Design; G.3 [Probability and Statistics]: Experimental Design; I.2.6 [AI]: Learning
DOI:10.1145/2461381.2461387
Record Number:CaltechAUTHORS:20130422-142151892
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20130422-142151892
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:38067
Collection:CaltechAUTHORS
Deposited By: Kristin Buxton
Deposited On:22 Apr 2013 21:34
Last Modified:09 Nov 2021 23:33

Repository Staff Only: item control page