* Tags:: #📜Papers
* Authors: [[janis-klaise|Janis Klaise]], [[arnaud-van-looveren|Arnaud Van Looveren]], [[clive-cox|Clive Cox]], [[giovanni-vacanti|Giovanni Vacanti]], [[alexandru-coca|Alexandru Coca]]
* Link:: http://arxiv.org/abs/2007.06299
* Zotero Link:: [klaiseMonitoringExplainabilityModels2020](zotero://select/items/@klaiseMonitoringExplainabilityModels2020)
* Source date:: 2020
* Finished date:: [[2021-07-13]]
There are three areas of interest:
- The usual [[model-performance|Model Performance]], for example latency.
- [[model-outlier-detection|Model Outlier Detection]]: that is, for a single instance, whether it is out-of-distribution of training data.
- And [[Model drifting|Model Drifting]], which can be further distinguished in two:
- [[Covariate shift|Covariate shift]]: the input data distribution changes, but p(y|x) remains the same.
- [[Label shift|Label Shift]]: the label distribution changes (that is, reality) but the conditional distribution stays the same p(x|y)