arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2004.12211stat.MLcs.LGstat.APstat.COstat.ME
收藏

无折衷贝叶斯神经网络

Compromise-free Bayesian neural networks

Cavendish Laboratory(卡文迪许实验室);University of Cambridge(剑桥大学)

Kamran Javid, Will Handley, Mike Hobson, Anthony Lasenby

摘要:

我们对贝叶斯神经网络(BNN)的样本外性能和贝叶斯证据(边际似然)之间的关系进行了彻底的分析,并研究了BNN集合的性能,两者都使用波士顿住房数据集。使用最先进的嵌套抽样,我们对完整(非高斯和多模态)网络后验进行数字采样,并获得贝叶斯证据的数值估计,考虑具有多达156个可训练参数的网络模型。该网络具有零到四个隐藏层,$\tanh$或$ReLU$激活函数,并且具有和不具有分层先验。通过确定网络上的后验分布,从由相关贝叶斯证据值重新加权的单个BNN的后验样本中获得BNN的集成。样本外性能与证据之间存在良好的相关性,证据与模型大小之间以及样本外性能与模型大小平面之间存在显著的对称性。具有$ReLU$激活函数的网络始终比具有$\tanh$函数的网络具有更高的证据,这反映在它们的样本外性能中。集成架构的作用是进一步提高相对于单个BNN的性能。

英文摘要:

We conduct a thorough analysis of the relationship between the out-of-sample performance and the Bayesian evidence (marginal likelihood) of Bayesian neural networks (BNNs), as well as looking at the performance of ensembles of BNNs, both using the Boston housing dataset. Using the state-of-the-art in nested sampling, we numerically sample the full (non-Gaussian and multimodal) network posterior and obtain numerical estimates of the Bayesian evidence, considering network models with up to 156 trainable parameters. The networks have between zero and four hidden layers, either $\tanh$ or $ReLU$ activation functions, and with and without hierarchical priors. The ensembles of BNNs are obtained by determining the posterior distribution over networks, from the posterior samples of individual BNNs re-weighted by the associated Bayesian evidence values. There is good correlation between out-of-sample performance and evidence, as well as a remarkable symmetry between the evidence versus model size and out-of-sample performance versus model size planes. Networks with $ReLU$ activation functions have consistently higher evidences than those with $\tanh$ functions, and this is reflected in their out-of-sample performance. Ensembling over architectures acts to further improve performance relative to the individual BNNs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑