Neural networks · Question 39 of 60
Why is structural coverage alone not sufficient to assess a neural network?
2 answers are correct. Medium
Show answer
Answer
- A Networks can learn spurious correlations and thus activate correctly for incorrect reasons
- B It does not guarantee that the network generalizes well or handles real-world variations
Explanation
Structural coverage does not guarantee that a network generalizes well or handles real-world variations. Networks can learn spurious correlations and thus activate correctly for incorrect reasons. Testers should therefore also use techniques such as adversarial and metamorphic testing.