After an autoencoder (AE) has learnt to reconstruct one dataset, it might be expected that the likelihood on an out-of-distribution (OOD) input would be low. This has been studied as an approach to detect OOD inputs. Recent work showed this intuitive approach can fail for the dataset pairs FashionMNIST vs MNIST. This paper suggests this is due to the use of Bernoulli likelihood and analyses why this is the case, proposing two fixes: 1) Compute the uncertainty of likelihood estimate by using a Bayesian version of the AE. 2) Use alternative distributions to model the likelihood.
翻译:自动编码器(AE) 学会重建一个数据集后, 可能预计分配外( OOOD) 输入的可能性会很低。 已经作为检测 OOD 输入的一种方法对此进行了研究。 最近的工作表明, 对于数据集配对FashonMNIST 和 MNIST 来说,这种直观的方法可能会失败。 本文指出,这是因为使用Bernoulli 的可能性和分析了这种情况的原因, 提出了两种解决办法:(1) 通过使用Bayesian版本的 AE 2 来计算概率估计的不确定性。 (2) 使用替代的分布来模拟可能性。