Understanding Failures in Out-of-Distribution Detection with
Deep Generative Models
Lily H. Zhang 1 Mark Goldstein 1 Rajesh Ranganath 1
Abstract distribution, estimated via a model (Bishop, 1994).
Deep generative models (DGMs) seem a natu- Deep generative models (DGMs) estimate complex distri-
ral fit for detecting out-of-distribution (OOD) in- butions from often high-dimensional inputs and produce
...


雷达卡




京公网安备 11010802022788号







