Comparison of Maximum Likelihood and GAN-based training of Real NVPs — mentioned by ericjang

Sat, 28 Oct 2017 17:28:43 GMT

This is a general problem for generative models, regardless of whether they are explicit latent density (variational autoencoders) or implicit density (GANs) or fully-observed (PixelCNN). To date no consensus has been reached in the research community on what the "right" way to evaluate a generative model is: e.g. Parzen Windows estimators, log-probability, ELBO, etc. Yann Lecun himself (as of NIPS2016) was pretty critical of probability-based metrics, as those have strong dependence to the choice of model (e.g. if the model is poor the log-probability is meaningless). In GANs, the critic and…

Paper and discussion history. A mention is not an endorsement.

Comparison of Maximum Likelihood and GAN-based training of Real NVPs · HackerPapers discussions

Comparison of Maximum Likelihood and GAN-based training of Real NVPs — mentioned by ericjang