Accession Number : AD1017792


Title :   Constrained Fisher Scoring for a Mixture of Factor Analyzers


Descriptive Note : Technical Report,01 Jan 2016,30 Sep 2016


Corporate Author : US Army Research Laboratory Adelphi United States


Personal Author(s) : Whipps,Gene T ; Ertin,Emre ; Moses,Randolph L


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/1017792.pdf


Report Date : 01 Sep 2016


Pagination or Media Count : 44


Abstract : This report considers the problem of learning an object appearance manifold using a spatially distributed network of sensors. Sensor nodes observe an object from different aspects and then learn a joint statistical model for the object manifold. We employ a mixture of factor analyzers model and derive a Fisher scoring method for maximum-likelihood estimation of the model parameters. We analyze convergence of the scoring method and derive stopping conditions for exiting the iterative algorithm. Simulation examples demonstrate that the proposed approach provides faster model learning over the popular expectation-maximization algorithm with similar computational requirements. Lastly, we demonstrate the efficacy of the proposed method for learning a global appearance model across the entire sensor network.


Descriptors :   maximum likelihood estimation , sensor networks , algorithms , machine learning , Factor analysis , algorithms


Subject Categories : Theoretical Mathematics
      Statistics and Probability
      Operations Research


Distribution Statement : APPROVED FOR PUBLIC RELEASE