BriefGPT.xyz
Feb, 2021
随机微分方程的无限深贝叶斯神经网络
Infinitely Deep Bayesian Neural Networks with Stochastic Differential Equations
HTML
PDF
Winnie Xu, Ricky T. Q. Chen, Xuechen Li, David Duvenaud
TL;DR
本文介绍了在连续深度贝叶斯神经网络中进行可扩展的近似推断的方法，借助随机微分方程给出了隐藏单元，并使用基于梯度的随机变分推断和新的梯度估计器，这种方法使连续深度贝叶斯神经网络具有和离散深度替代方法同样的竞争力，并继承了神经ODE的内存高效训练和可调节精度。
Abstract
We perform scalable approximate inference in a recently-proposed family of continuous-depth
bayesian neural networks
. In this model class, uncertainty about separate weights in each layer produces dynamics that follow a
→