3df80af53dce8435cf9ad6c3e7a403fd-Paper.pdf

The Gumbel-Max trick is the basis of many relaxed gradient estimators. These estimators areeasy toimplement and lowvariance, butthegoal ofscaling them comprehensively to large combinatorial distributions is still outstanding. Working within the perturbation model framework, we introduce stochastic softmax tricks, which generalizetheGumbel-Softmax tricktocombinatorial spaces.