Timezone: »
Simplex-valued data appear throughout statistics and machine learning, for example in the context of transfer learning and compression of deep networks. Existing models for this class of data rely on the Dirichlet distribution or other related loss functions; here we show these standard choices suffer systematically from a number of limitations, including bias and numerical issues that frustrate the use of flexible network models upstream of these distributions. We resolve these limitations by introducing a novel exponential family of distributions for modeling simplex-valued data – the continuous categorical, which arises as a nontrivial multivariate generalization of the recently discovered continuous Bernoulli. Unlike the Dirichlet and other typical choices, the continuous categorical results in a well-behaved probabilistic loss function that produces unbiased estimators, while preserving the mathematical simplicity of the Dirichlet. As well as exploring its theoretical properties, we introduce sampling methods for this distribution that are amenable to the reparameterization trick, and evaluate their performance. Lastly, we demonstrate that the continuous categorical outperforms standard choices empirically, across a simulation study, an applied example on multi-party elections, and a neural network compression task.
Author Information
Elliott Gordon-Rodriguez (Columbia University)
Gabriel Loaiza-Ganem (Layer 6 AI)
John Cunningham (Columbia)
More from the Same Authors
-
2023 : Practical and Asymptotically Exact Conditional Sampling in Diffusion Models »
Brian Trippe · Luhuan Wu · Christian Naesseth · David Blei · John Cunningham -
2023 Poster: TR0N: Translator Networks for 0-Shot Plug-and-Play Conditional Generation »
Zhaoyan Liu · Noël Vouitsis · Satya Krishna Gorti · Jimmy Ba · Gabriel Loaiza-Ganem -
2022 : Neural Implicit Manifold Learning for Topology-Aware Generative Modelling »
Brendan Ross · Gabriel Loaiza-Ganem · Anthony Caterini · Jesse Cresswell -
2022 Poster: Scaling Structured Inference with Randomization »
Yao Fu · John Cunningham · Mirella Lapata -
2022 Spotlight: Scaling Structured Inference with Randomization »
Yao Fu · John Cunningham · Mirella Lapata -
2022 Poster: Bayesian Nonparametrics for Offline Skill Discovery »
Valentin Villecroze · Harry Braviner · Panteha Naderian · Chris Maddison · Gabriel Loaiza-Ganem -
2022 Spotlight: Bayesian Nonparametrics for Offline Skill Discovery »
Valentin Villecroze · Harry Braviner · Panteha Naderian · Chris Maddison · Gabriel Loaiza-Ganem -
2022 Poster: Preconditioning for Scalable Gaussian Process Hyperparameter Optimization »
Jonathan Wenger · Geoff Pleiss · Philipp Hennig · John Cunningham · Jacob Gardner -
2022 Oral: Preconditioning for Scalable Gaussian Process Hyperparameter Optimization »
Jonathan Wenger · Geoff Pleiss · Philipp Hennig · John Cunningham · Jacob Gardner -
2019 Poster: Discriminative Regularization for Latent Variable Models with Applications to Electrocardiography »
Andrew Miller · Ziad Obermeyer · John Cunningham · Sendhil Mullainathan -
2019 Oral: Discriminative Regularization for Latent Variable Models with Applications to Electrocardiography »
Andrew Miller · Ziad Obermeyer · John Cunningham · Sendhil Mullainathan