TY - GEN
T1 - Decentralized regression with asynchronous sub-Nyquist sampling
AU - Wai, Hoi To
AU - Scaglione, Anna
PY - 2015/4/24
Y1 - 2015/4/24
N2 - When capturing data on a sensor field to uncover its latent structure, there are often nuisance parameters in the observation model that turn even linear regression problems into non-convex optimizations. One common case is the lack of common timing source in ADCs, therefore samplings are done with time offsets. Motivated by the desire of estimating jointly the sensor field and nuisance parameters in a wide area deployment, this paper derives a new decentralized algorithm that combines alternating optimization and gossip-based learning. The proposed algorithm is shown to converge to the neighborhood of a local minimum, both analytically and empirically.
AB - When capturing data on a sensor field to uncover its latent structure, there are often nuisance parameters in the observation model that turn even linear regression problems into non-convex optimizations. One common case is the lack of common timing source in ADCs, therefore samplings are done with time offsets. Motivated by the desire of estimating jointly the sensor field and nuisance parameters in a wide area deployment, this paper derives a new decentralized algorithm that combines alternating optimization and gossip-based learning. The proposed algorithm is shown to converge to the neighborhood of a local minimum, both analytically and empirically.
KW - asynchronous sampling
KW - decentralized regression
KW - gossip algorithm
KW - sub-Nyquist sampling
UR - http://www.scopus.com/inward/record.url?scp=84940511185&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84940511185&partnerID=8YFLogxK
U2 - 10.1109/ACSSC.2014.7094559
DO - 10.1109/ACSSC.2014.7094559
M3 - Conference contribution
AN - SCOPUS:84940511185
T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers
SP - 798
EP - 802
BT - Conference Record of the 48th Asilomar Conference on Signals, Systems and Computers
A2 - Matthews, Michael B.
PB - IEEE Computer Society
T2 - 48th Asilomar Conference on Signals, Systems and Computers, ACSSC 2015
Y2 - 2 November 2014 through 5 November 2014
ER -