|
Published Articles >> Table of Contents >> Abstract
14th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP'06)
pp. 373-379
Fast Decentralized Learning of a Gaussian Mixture Model for Large-Scale Multimedia Retrieval
A. Nikseresht, Ecole polytechnique de l’universite de Nantes, Nantes cedex, France.
M. Gelgon, Ecole polytechnique de l’universite de Nantes, Nantes cedex, France.
Full Article Text:
 
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/PDP.2006.37
Send link to a friend
| Abstract |
|
We address herein the distributed computation of a probability
density estimate. Class-conditional probability density
estimation is a central need in multimedia pattern
recognition, but has classically be conducted in a centralized
fashion. In contrast, the present work is motivated by
the perspective of a multimedia indexing and retrieval peerto-
peer system over the internet. In a decentralized fashion,
algorithms and data from various contributors would
cooperate towards a collective statistical learning. A typical
need is aggregation of probabilistic Gaussian mixture
models describing the same class, but estimated on several
nodes on different data sets. We tackle this goal through
an approach requiring only moderate computation at each
node and little data to transit between nodes. Both properties
are obtained by fusion models via their (few) parameters,
rather than via multimedia data itself. Estimation of
the aggregated model is provided by an iterative scheme,
derived from a modification on Kullback divergence. We
provide experimental results on a speaker recognition task
with real data, in a gossip propagation setting.
|
 Additional Information
|
Citation:
A. Nikseresht, M. Gelgon,
"Fast Decentralized Learning of a Gaussian Mixture Model for Large-Scale Multimedia Retrieval,"
pdp,
pp. 373-379,
14th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP'06),
2006
|
|
