Failures

Drift

1 min read

Definition

Gradual degradation of agent performance over time due to changes in data, environment, or the agent itself.

Drift refers to the phenomenon where an agent's performance degrades gradually, often unnoticed until significant damage occurs.

Types

  • Data drift: Input distribution changes from training
  • Concept drift: Relationship between inputs and outputs changes
  • Model drift: Agent behavior changes (through fine-tuning, updates)

Detection

Continuous monitoring of performance metrics with statistical tests for significant changes.

failuresmonitoringdegradation