Concept drift
When the relationship between inputs and the correct answer changes, so the same-looking input now means something different. The classic case is fraud, where adversaries deliberately change behavior to evade detection. Harder to catch because inputs can look normal while the model's learned logic has gone stale.
Defined in 2 GAGE programs, which carry 2 distinct definitions of it. The wording above is taught in Certified AI Transformation Professional (CATP).
How each discipline defines it
The same term does different work depending on who is using it. These are the definitions as each program teaches them, unedited.
When the relationship between inputs and the correct answer changes, so the same-looking input now means something different. The classic case is fraud, where adversaries deliberately change behavior to evade detection. Harder to catch because inputs can look normal while the model's learned logic has gone stale.
A change in the relationship between inputs and the correct output, written as a change in P(Y given X). The inputs can look identical to training while the correct answers change, which makes concept drift the hardest kind to detect, because input monitoring alone will miss it.
Where it is taught
The exact lessons this term appears in. The first module of every program is free with a free account.
- Model drift: detecting the quiet degradation nobody reports · Evaluation and Trust, Certified AI Governance Professional (CAIGP)
- Process Governance and Continuous Improvement · Process and Operations Redesign, Certified AI Transformation Professional (CATP)
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.