1 Hit in 2.4 sec

SIReN-VAE: Leveraging Flows and Amortized Inference for Bayesian Networks [article]

Jacobie Mouton, Steve Kroon
2022 arXiv   pre-print
This is achieved by extending both the prior and inference network with graphical residual flows - residual flows that encode conditional independence by masking the weight matrices of the flow's residual  ...  This work explores incorporating arbitrary dependency structures, as specified by Bayesian networks, into VAEs.  ...  A SIREN-VAE ILLUSTRATION Figure 1 : 1 Figure 1: SIReN-VAE encodes the BN's graphical structure into the decoder via masking of the flow and decoder neural network weights.  ... 
arXiv:2204.11847v1 fatcat:53jbdpa7ovaolc555pehqgzdmu