True or false: The neural network is a classification model.

• A. Not specified
• B. Sometimes
• C. True
• D. False

Answer: (B)

Neural networks are versatile models that can be used for many tasks, not just one. They can function as classifiers when their outputs represent class probabilities and they’re trained with a classification loss like cross-entropy. They can also function as regressors when the outputs are continuous values and trained with a regression loss like mean squared error. Because of this flexibility, calling a neural network strictly a classification model isn’t accurate; it’s sometimes used for classification, but not exclusively. For example, a network designed to recognize digits uses a softmax output and cross-entropy loss, while a network predicting house prices uses a linear output with a regression loss. So the statement is true only in some contexts, making “Sometimes” the best answer.