Training Data
-
AI Ethics & Policy
You Cannot Audit a Model Whose Training Data You Cannot Describe
Questions about bias, licensing and consent all reduce to the same prerequisite — knowing what went into the model in…
Read More » -
Synthetic Training Data Amplifies the Gaps It Was Meant to Fill
Generating examples from a model that already underperforms on rare cases produces more of what it already knows, not what…
Read More » -
Computer Vision
Your Model Is Learning Your Annotators’ Disagreements
When two labellers disagree on a fifth of examples, no architecture recovers the signal, because the ceiling is set by…
Read More »