Descripción
|
|
---|---|
Estimation of distribution algorithms (EDAs) are a recent optimization heuristic included in the class of stochastic population-based search methods. EDAs work by constructing an explicit probability model from a set of selected solutions, which is then conveniently used to generate new promising solutions in the next iteration of the evolutionary process. Regularization methods modify the likelihood function to get sensible procedures in these unstable situations. Parameter estimates are restricted maximum likelihood estimates since the new function to be optimized includes a penalty term where the size of the estimators is constrained. This size may be measured with different norms giving rise to different names: ridge, lasso, bridge, elastic net, etc. attracting the attention of many researchers from different fields in the last years. Some synergies between EDAs and regularization will be analyzed in the talk. On the one hand we will see that EDAs emerge as natural regularizers without having to be penalized. One the other hand we will introduce an EDA based on regularization. | |
Internacional
|
Si |
Nombre congreso
|
Forthcoming Seminar |
Entidad organizadora
|
University of Essex |
Nacionalidad Entidad
|
REINO UNIDO |
Lugar/Ciudad de impartición
|
Colchester |
Fecha inicio
|
30/01/2009 |
Fecha fin
|
30/01/2009 |