Dept. of Comput Sci., & Inf. Eng., National Taiwan Univ.Liou, Cheng-YuaCheng-YuaLiouLin, shiao-Lishiao-LiLin2007-04-192018-07-052007-04-192018-07-051989-06http://ntur.lib.ntu.edu.tw//handle/246246/2007041910032166A novel theory for studying the learning behavior of a neural network which is formed by interconnecting neurons is presented. This learning theory constitutes a new approach to the Boltzmann machine. The central idea is to minimize one of the two cross-entropy-like criteria, the cross-entropy and the reversed cross-entropy; the latter is used by Ackley et al. (Cognitive Sci., vol.9, p.147-59, 1985) in deriving the Boltzmann machine. The results derived by the present approach are closely related to those obtained by Ackley et al., with several significant modifications in the algorithm. A detailed discussion of the new algorithm, which is shown to be a probability-weighted version of the algorithm by Ackley et al., is presented.>application/pdf400727 bytesapplication/pdfen-USjournal article10.1109/IJCNN.1989.118618http://ntur.lib.ntu.edu.tw/bitstream/246246/2007041910032166/1/00118618.pdf