BriefGPT.xyz
Aug, 2020
深度神经网络中冷后验的统计理论
A statistical theory of cold posteriors in deep neural networks
HTML
PDF
Laurence Aitchison
TL;DR
本文讨论贝叶斯神经网络在图片分类上的应用，发现此类应用使用的是错误的似然度。作者开发了一个描述“筛选”过程的生成模型，并与之前使用的调整后的后验概率似然度进行了基于贝叶斯思想的对比。
Abstract
To get
bayesian neural networks
to perform comparably to
standard neural networks
it is usually necessary to artificially reduce uncertainty using a "tempered" or "cold" posterior. This is extremely concerning: i
→