flâneur

Learning Latent Permutations with Gumbel-Sinkhorn Networks – arXiv Vanity

arxiv-vanity.com · 17,217 words · saved by 1 readers

Permutations and matchings are core building blocks in a variety of latent variable models, as they allow us to align, canonicalize, and sort data. Learning in such models is difficult, however, because exact marginalization over these combinatorial objects is intractable. In response, this paper introduces a collection of new methods for end-to-end learning in such models that approximate discrete maximum-weight matching using the continuous Sinkhorn operator. Sinkhorn operator is attractive because it functions as a simple, easy-to-implement analog of the softmax operator. With this, we can define the Gumbel-Sinkhorn method, an extension of the Gumbel-Softmax method (Jang et al., 2016; Maddison et al., 2016) to distributions over latent matchings. We demonstrate the effectiveness of our method by outperforming competitive baselines on a range of qualitatively different tasks: sorting numbers, solving jigsaw puzzles, and identifying neural signals in worms.

Learning Latent Permutations with Gumbel-Sinkhorn Networks Gonzalo E. Mena Department of Statistics, Columbia University gem2131@columbia.edu &David Belanger Google Brain &Scott Linderman Department of Statistics, Columbia University &Jasper Snoek Google Brain Work done while the author was at Google Brain. Abstract Permutations and matchings are core building blocks in a variety of latent variable models, as they allow us to align, canonicalize, and sort data. Learning in such models is difficult, however, because exact marginalization over these combinatorial objects is intractable. In respo

related reading