Additive versus exponentiated gradient updates for linear prediction
STOC '95 Proceedings of the twenty-seventh annual ACM symposium on Theory of computing
A comparison of new and old algorithms for a mixture estimation problem
COLT '95 Proceedings of the eighth annual conference on Computational learning theory
Hi-index | 0.01 |
Hidden units play an important role in neural networks, althoughtheir activation values are unknown in many learning situations.The EM algorithm (statistical algorithm) and the emalgorithm (information-geometric one) have been proposed so far inthis connection, and the effectiveness of such algorithms isrecognized in many areas of research. The present note points outthat these two algorithms are equivalent under a certain condition,although they are different in general.