A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning

Iris A.M. Huijben (Corresponding author), Wouter Kool, Max Benedikt Paulus, Ruud J.G. van Sloun

Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

31 Citaten (Scopus)
706 Downloads (Pure)

Samenvatting

The Gumbel-max trick is a method to draw a sample from a categorical distribution, given by its unnormalized (log-)probabilities. Over the past years, the machine learning community has proposed several extensions of this trick to facilitate, e.g., drawing multiple samples, sampling from structured domains, or gradient estimation for error backpropagation in neural network optimization. The goal of this survey article is to present background about the Gumbel-max trick, and to provide a structured overview of its extensions to ease algorithm selection. Moreover, it presents a comprehensive outline of (machine learning) literature in which Gumbel-based algorithms have been leveraged, reviews commonly-made design choices, and sketches a future perspective.

Originele taal-2Engels
Artikelnummer9729603
Pagina's (van-tot)1353-1371
Aantal pagina's19
TijdschriftIEEE Transactions on Pattern Analysis and Machine Intelligence
Volume45
Nummer van het tijdschrift2
Vroegere onlinedatum7 mrt. 2022
DOI's
StatusGepubliceerd - 1 feb. 2023

Bibliografische nota

Publisher Copyright:
IEEE

Vingerafdruk

Duik in de onderzoeksthema's van 'A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning'. Samen vormen ze een unieke vingerafdruk.

Citeer dit