Functional performance testing · Question 40 of 60
Why does functional performance testing of probabilistic ML systems go beyond a simple "passed/failed"?
1 answer is correct. Medium
Show answer
Answer
- D To address non-determinism by evaluating behavior statistically across a lot of data
Explanation
This testing statistically measures a model's performance against its acceptance criteria instead of just determining passed or failed. This is necessary to address the non-determinism of ML systems by evaluating behavior across a large, representative dataset.