Model Deployment and Monitoring

Model deployment makes a trained model available for inference in a real application. Model monitoring then tracks behavior and performance, including changes such as data drift.

Japanese terms

1. Model deployment — モデル展開てんかい: Making a trained model available in a production system or application.
2. Inference — 推論すいろん: Using a trained model to produce an output from new input.
3. Model monitoring — モデル監視かんし: Tracking a deployed model’s behavior, performance, inputs, and operational health.
4. Data drift — データドリフト: A change over time in the statistical properties of data received by a deployed system.

Related topics

Training, validation, and testing establishes pre-deployment evidence, while deployment reveals how a model behaves under real operating conditions.