Asynchronous distributed principal component analysis using stochastic approximation

Gemma Morral (1, 2) , Pascal Bianchi (1, 2) , Jérémie Jakubowicz (3, 4, 5)

1 LTCI - Laboratoire Traitement et Communication de l'Information
2 TSI - Département Traitement du Signal et des Images
3 TSP - RST - Département Réseaux et Services de Télécommunications
4 SAMOVAR - Services répartis, Architectures, MOdélisation, Validation, Administration des Réseaux
5 CNRS - Centre National de la Recherche Scientifique

In this paper we address the problem of asynchronous distributed principal component analysis. We provide several algorithms coping with different situations according to the underlying graph structure. A general enough framework allows us to analyze all these algorithms at the same time. Convergence is proved with probability 1 under suitable assumptions, and numerical experiments illustrate their good behavior.

Mots clés: Asynchronous algorithm; Distributed algorithm; Oja's algorithm; Principal component analysis; Stochastic approximation; Wireless sensor network

Domaines: Réseaux et télécommunications [cs.NI]

Gemma Morral, Pascal Bianchi, Jérémie Jakubowicz. Asynchronous distributed principal component analysis using stochastic approximation. CDC '12 : 51st IEEE Conference on Decision and Control, Dec 2012, Maui, Hi, United States. pp.1398-1403, ⟨10.1109/CDC.2012.6425898⟩. ⟨hal-00788714⟩