Abstract | ||
---|---|---|
We present an example of a joint spatial and temporal task learning algorithm that results in a generative model that has applications for on-line visual control. We review work on learning transformed mixture of gaussians (due to Frey and Jojic) and Variable Length Markov Models (VLMMS due to Ron, Singer and Tishby). We show how a temporal model, learned through an extension of VLMMs to deal with multinomially distributed input symbol vectors, can be used as an improvement on Maximum Likelihood (ML) for prior parameter estimation for the Expectation Maximisation (EM) process. |
Year | DOI | Venue |
---|---|---|
2004 | 10.1109/ICPR.2004.516 | ICPR (2) |
Keywords | Field | DocType |
prior parameter estimation,temporal task,maximum likelihood,joint spatial,expectation maximisation,temporal model,temporal structure learning,variable length markov models,generative model,input symbol vector,on-line visual control,mixture of gaussians,learning artificial intelligence,parameter estimation,multinomial distribution,maximum likelihood estimation,markov model,computer vision,markov processes | Markov process,Pattern recognition,Computer science,Markov model,Symbol,Maximum likelihood,Artificial intelligence,Estimation theory,Visual control,Mixture model,Generative model | Conference |
ISSN | ISBN | Citations |
1051-4651 | 0-7695-2128-2 | 4 |
PageRank | References | Authors |
0.44 | 10 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kingsley Sage | 1 | 8 | 1.56 |
Hilary Buxton | 2 | 491 | 135.93 |