Method and device for learning stochastic inference models between several random variables with unpaired data

摘要

A system and method for operating a neural network. In some embodiments, the neural network includes a variance auto-encoder, and training the neural network includes training the variance auto-encoder with a plurality of samples of a first random variable; and a plurality of samples of a second random variable, the plurality of samples of the first random variable and the plurality of samples of the second random variable being unpaired, training the neural network including updating weights in the neural network based on a first loss function, the first Loss function is based on a measurement of a deviation from consistency between: a conditional generation path from the first random variable to the second random variable, and a conditional generation path from the second random variable to the first random variable.