BriefGPT.xyz
Jun, 2023
黑盒变分推断的可证收敛性保证
Provable convergence guarantees for black-box variational inference
HTML
PDF
Justin Domke, Guillaume Garrigos, Robert Gower
TL;DR
针对黑盒变分推理的随机优化的挑战，该论文提出了一种基于重参数化的梯度估计方法来保证其收敛性，并给出了针对密集高斯变分族收敛的新的收敛保证和独特噪声边界.
Abstract
While
black-box variational inference
is widely used, there is no proof that its
stochastic optimization
succeeds. We suggest this is due to a theoretical gap in existing
→