Learning to Sample with Adversarially Learned Likelihood-Ratio

By Chunyuan Li, Jianqiao Li, Guoyin Wang, Lawrence Carin. 2018

14 citations recorded by the source corpus.

Citations in the corpus

• Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems (2019, methodology): “ly from data and without the need to assume any specic parametrization for the generative model distribution p (x;y). 2.2.2. Entropic regularization bound Here we follow the derivation of Li et. al [55] to construct a computable lower bound for the entropy H(p (x;y)). To this end, we start by consid”
• Adversarial Uncertainty Quantification in Physics-Informed Neural Networks (2018, methodology): “Here we follow the derivation of Li et. al [34] to construct a computable lower bound for the entropy h ( p θ ( x , t, u )).”
• Physics-informed deep generative models (2018, unknown): “As the entropy term h(pθ(x, t,u))) is intractable, we can obtain a computable training objective by deriving the following lower bound (see proof in Appendix A) h(pθ(x, t,u)) ≥ h(p(z)) + Epθ(x,t,u,z)[log(qφ(z|x, t,u))], (5) where the inference model qφ(z|x, t,u) plays the role of a variational appro”