Monitoring and explainability of models in production

Firstly, it is critical to ensure model performance does not degrade in a production setting. Inability to detect model The machine learning lifecycle extends beyond performance degradation can lead to stale models and increased the deployment stage. Monitoring deployed models technical debt (Breck et al., 2017; Sculley et al., is crucial for continued provision of high quality 2015). Whilst trained models usually come with performance machine learning enabled services. Key areas metrics on offline test sets, this does not guarantee include model performance and data monitoring, similar performance in live systems.