Câu 2: CT-AI: ISTQB Certified Tester - AI Testing

Which ONE of the below is MOST likely to indicate a problem with underfitting in an ML model?

1. A. The model is vulnerable to adversarial attacks
2. B. The model fails to generalize on new data
3. C. The model uses a large amount of resources to make a prediction
4. D. The model is inaccurate on data similar to the training data — đáp án hiện tại