Skip to content

CS Professional · Artificial Intelligence, Data Analytics and Cyber Security - Laws and Practice · Artificial Intelligence - Introduction and Basics

A bank's credit-scoring model was accurate at launch, but two years later its accuracy falls because customer income patterns and spending behaviour have changed. Which practice in the AI lifecycle is designed to detect and address this?

Continuous monitoring with periodic retraining addresses this. Falling accuracy because customer behaviour has changed is model or data drift, which is caught by tracking performance after deployment and fixed by retraining on recent data. A one-time validation, encryption or a bigger server cannot correct drift.

  1. AContinuous monitoring with periodic retrainingCorrect
  2. BOne-time validation before launch
  3. CEncrypting the model files
  4. DIncreasing the size of the server

Explanation

A fall in accuracy because real-world data has shifted is called model or data drift. It is detected by continuous monitoring of performance and addressed by retraining on recent data. A one-time pre-launch check cannot catch later change, and encryption or hardware affect security or speed, not accuracy.

Did you get it right without looking?

One question tells you little. A timed set on Artificial Intelligence - Introduction and Basics shows your real accuracy, how long you take and where you lose marks.

More Artificial Intelligence - Introduction and Basics questions