TY  - RPRT
TI  - Bayesian model selection in the $\mathcal{M}$-open setting -- Approximate posterior inference and probability-proportional-to-size subsampling for efficient large-scale leave-one-out cross-validation
AU  - Riko Kelter
PY  - 2020
UR  - https://arxiv.org/abs/2005.13199
ID  - 2005.13199
ER  -