Authors: Jian Zhang
Addresses: School of Computer Science and Technology, China University of Mining and Technology, Xuzhou, 221116, China
Abstract: This paper mainly introduces the theory and the abilities of generating images of restricted Boltzmann machines (RBMs) and variational autoencoders (VAEs). First, we introduce these models, which can be treated as the basic blocks of deep neural nets. Second, this paper introduces the ability of generating images based on these models. Lastly, this paper introduces the hybrid model based on RBMs and VAEs and another model called a Boltzmann encoded adversarial machine (BEAM). The experiments show the effectiveness of the hybrid models.
Keywords: restricted Boltzmann machines; RBMs; variational autoencoders; VAEs; generative adversarial nets; deep neural nets.
International Journal of Collaborative Intelligence, 2019 Vol.2 No.1, pp.66 - 74
Received: 23 Aug 2018
Accepted: 29 Sep 2018
Published online: 14 Mar 2019 *