Domain Shift

Intermediate

A mismatch between training and deployment data distributions that can degrade model performance.

Full Definition

A mismatch between training and deployment data distributions that can degrade model performance.

Keywords

covariate shift drift

Domains

MLOps & Infrastructure

Related Terms

MLOps related to

Practices for operationalizing ML: versioning, CI/CD, monitoring, retraining, and reliable production management.

 CI/CD for ML related to

Automated testing and deployment processes for models and data workflows, extending DevOps to ML artifacts.

 Monitoring related to

Observing model inputs/outputs, latency, cost, and quality over time to catch regressions and drift.

 Training Pipeline related to

End-to-end process for model training.

Concept Map

See how Domain Shift connects to other concepts.

Open Knowledge Graph