====== Gumbel-max Trick ======

See [[https://timvieira.github.io/blog/post/2014/07/31/gumbel-max-trick/|Gumbel-max trick]] at Tim Vieira's blog, and this follow-up post on [[https://timvieira.github.io/blog/post/2014/08/01/gumbel-max-trick-and-weighted-reservoir-sampling/|Weighted reservoir sampling]].

The Gumbel-max trick is widely used in deep learning from this paper: [[https://arxiv.org/pdf/1611.01144.pdf|Categorical Reparameterization with Gumbel-Softmax]].  See also the Reparameterization Trick section 6.4 in [[https://arxiv.org/pdf/1812.06834.pdf|Kim et al 2018 - A Tutorial on Deep Latent Variable Models of Natural Language]]