Feedback Loop

noun

say it: “FEED-bak LOOP” /ˈfidˌbæk ˈlup/

A situation in which the results of a process are fed back into it, so its effects grow stronger over time.

What is a feedback loop in machine learning?

A feedback loop is when a model's decisions shape the data its next version learns from. If a screen invites few people from one district, few of them get the chance to show they qualify, so the next model, trained on those results, invites even fewer. An early bias grows instead of fading.

In a sentence

  • “A feedback loop meant the neighborhoods patrolled most produced the most reports, which sent still more patrols.”
  • “Retraining on its own decisions created a feedback loop in the hiring model.”

Level: Journeyman · More AI Basics terms at this level

Updated