BriefGPT.xyz
Feb, 2020
k-tied正态分布:贝叶斯神经网络中高斯均值场后验的紧凑参数化
The k-tied Normal Distribution: A Compact Parameterization of Gaussian Mean Field Posteriors in Bayesian Neural Networks
HTML
PDF
Jakub Swiatkowski, Kevin Roth, Bastiaan S. Veeling, Linh Tran, Joshua V. Dillon...
TL;DR
通过对高斯均值场变分推理方法训练的深层贝叶斯神经网络的后验标准差进行矩阵低秩分解,我们可以将变分推理方法更紧凑地参数化,并提高其信噪比,从而加速其收敛速度。
Abstract
variational bayesian inference
is a popular methodology for approximating posterior distributions over
bayesian neural network
weights. Recent work developing this class of methods has explored ever richer parame
→