Antoine Harary, Director at Strategy One, discusses the preliminary outcome of the GE Innovation Bar...
Complete video at: fora.tv Author and former World Bank executive Antoine van Agtmael forecasts what...
Keine Beschreibung verfügbar
EGS, professor emeritus (Sorbonne), is an internationally recognized authority on Western philosophy...
EGS, professor emeritus (Sorbonne), is an internationally recognized authority on Western philosophy...
citeseer |
(0) (0 Votes)
|
Views: (1016) Date: (08-04-09) Pages: () |
Abstract: Abstract Guided by the goal of obtaining an optimization algorithm that is both fast andyields good generalization, we study the descent direction maximizing the decrease in generalization error or the probability of not increasing generalizationerror. The surprising result is that from both the Bayesian and frequentist perspectives this can yield the natural gradient direction. Although that direction can bevery expensive to compute we develop an efficient, general, online approximation to the natural gradient descent which is suited to large scale problems. We re-port experimental results showing much faster convergence in computation time and in number of iterations with TONGA (Topmoumoute Online natural GradientAlgorithm) than with stochastic gradient descent, even on very large datasets.