Timezone: »
Poster
SVGD as a kernelized Wasserstein gradient flow of the chi-squared divergence
Sinho Chewi · Thibaut Le Gouic · Chen Lu · Tyler Maunu · Philippe Rigollet
Stein Variational Gradient Descent (SVGD), a popular sampling algorithm, is often described as the kernelized gradient flow for the Kullback-Leibler divergence in the geometry of optimal transport. We introduce a new perspective on SVGD that instead views SVGD as the kernelized gradient flow of the chi-squared divergence. Motivated by this perspective, we provide a convergence analysis of the chi-squared gradient flow. We also show that our new perspective provides better guidelines for choosing effective kernels for SVGD.
Author Information
Sinho Chewi (Massachusetts Institute of Technology)
Thibaut Le Gouic (Massachusetts Institute of Technology)
Chen Lu (Massachusetts Institute of Technology)
Tyler Maunu (Massachusetts Institute of Technology)
Philippe Rigollet (MIT)
More from the Same Authors
-
2021 Spotlight: Averaging on the Bures-Wasserstein manifold: dimension-free convergence of gradient descent »
Jason Altschuler · Sinho Chewi · Patrik R Gerber · Austin Stromme -
2022 : Sampling is as easy as learning the score: theory for diffusion models with minimal data assumptions »
Sitan Chen · Sinho Chewi · Jerry Li · Yuanzhi Li · Adil Salim · Anru Zhang -
2023 Poster: The emergence of clusters in self-attention dynamics »
Borjan Geshkovski · Cyril Letrouit · Yury Polyanskiy · Philippe Rigollet -
2023 Poster: The probability flow ODE is provably fast »
Sitan Chen · Sinho Chewi · Holden Lee · Yuanzhi Li · Jianfeng Lu · Adil Salim -
2023 Poster: Learning threshold neurons via edge of stability »
Kwangjun Ahn · Sebastien Bubeck · Sinho Chewi · Yin Tat Lee · Felipe Suarez · Yi Zhang -
2022 Poster: Variational inference via Wasserstein gradient flows »
Marc Lambert · Sinho Chewi · Francis Bach · Silvère Bonnabel · Philippe Rigollet -
2022 Poster: GULP: a prediction-based metric between representations »
Enric Boix-Adsera · Hannah Lawrence · George Stepaniants · Philippe Rigollet -
2021 Poster: Averaging on the Bures-Wasserstein manifold: dimension-free convergence of gradient descent »
Jason Altschuler · Sinho Chewi · Patrik R Gerber · Austin Stromme -
2021 Poster: Efficient constrained sampling via the mirror-Langevin algorithm »
Kwangjun Ahn · Sinho Chewi -
2020 Poster: Exponential ergodicity of mirror-Langevin diffusions »
Sinho Chewi · Thibaut Le Gouic · Chen Lu · Tyler Maunu · Philippe Rigollet · Austin Stromme -
2019 Poster: Power analysis of knockoff filters for correlated designs »
Jingbo Liu · Philippe Rigollet -
2017 Poster: Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration »
Jason Altschuler · Jonathan Niles-Weed · Philippe Rigollet -
2017 Spotlight: Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration »
Jason Altschuler · Jonathan Niles-Weed · Philippe Rigollet