4 ms·
And for Bayesian GAN, you have a model that generates those two models that are pitted against each other.
by xcodevn 9y ago
And for Bayesian GAN, you have a model that generates those two models that are pitted against each other.
- pas 9y agoWhat does it do with the models? So it samples a probability distribution, by training GANs, but how does it evaluate the points? (Which GAN is better?) Even after looking at the readme, I don't understand how it works. Could you explain it a bit, please?
- mendeza 9y agoI believe it models a distribution of the weights the network after training. Once you can generate samples, you can average the values to get expected mean, which is the improved prediction.
- xcodevn 9y agoIf you read the paper https://arxiv.org/abs/1705.09558 https://arxiv.org/abs/1705.09558, in section 2.1, it defines two conditional posteriors for the parameters of the generator and discriminator. Then, classical GAN is just a maximum likelihood estimate (or, a MAP estimate with uniform prior) of the parameters. Next, the paper said: how about sampling the whole posterior distribution instead of finding only one point of maximum likelihood as classical GANs do. This is the point where the famous Markov chain Monte Carlo (MCMC) algorithms become useful. They use something called stochastic gradient Hamiltonian Monte Carlo, basically, it is a random walk algorithm, at each step, you follow a noisy gradient, as a result, you converge to the posterior distribution instead of a local minima as gradient descent does. The paper claims that sampling the whole posterior helps to resolve problems with classical GANs. IMHO, this isn't a surprise claim, this is exactly what is good about Bayesian statistics.
- carbocation 9y agoIs it fair to presume that we'll next see a variational inference approach to this, with faster but slightly less optimal results?