Deep Learning | Page 25 | testinflow.com

Semester Exam Deep Learning InternationalMedium

Which option correctly explains Gradient descent in a neural-network workflow?

A. Gradient descent updates weights randomly without gradients B. Gradient descent is a file compression algorithm C. Gradient descent only applies to databases D. Gradient descent updates parameters in the direction that reduces loss

Correct Answer: D. Gradient descent updates parameters in the direction that reduces loss

Semester Exam Deep Learning InternationalMedium

Which option correctly explains Data augmentation in a neural-network workflow?

A. Data augmentation means deleting all training images B. Data augmentation is used only after deployment C. Data augmentation makes validation unnecessary D. Data augmentation creates modified training examples to improve robustness

Correct Answer: D. Data augmentation creates modified training examples to improve robustness

Semester Exam Deep Learning InternationalMedium

Which option correctly explains Test set in a neural-network workflow? Focus on the most precise definition.

A. A test set is used to update weights every epoch B. A test set should be identical to the training set C. A test set is used for data normalization only D. A test set estimates performance on unseen data

Correct Answer: D. A test set estimates performance on unseen data

Semester Exam Deep Learning InternationalMedium

For an exam question on Learning rate, which answer is most correct?

A. The learning rate controls the step size of parameter updates B. The learning rate is the number of output classes C. The learning rate is the dataset size D. The learning rate disables optimization

Correct Answer: A. The learning rate controls the step size of parameter updates

Semester Exam Deep Learning InternationalMedium

For an exam question on Data normalization, which answer is most correct?

A. Data normalization scales or centers data to support stable training B. Data normalization changes labels into random values C. Data normalization always increases image size D. Data normalization is the same as model pruning

Correct Answer: A. Data normalization scales or centers data to support stable training

Semester Exam Deep Learning InternationalMedium

For an exam question on Validation set, which answer is most correct? Focus on the most precise definition.

A. A validation set supports model selection and hyperparameter tuning B. A validation set is always larger than the training set C. A validation set replaces the loss function D. A validation set stores GPU kernels

Correct Answer: A. A validation set supports model selection and hyperparameter tuning

Semester Exam Deep Learning InternationalHard

Which misconception about CNN convolution is avoided by the correct answer?

A. Convolution removes all spatial structure B. Convolution applies learnable filters to local regions of input data C. Convolution always sorts neurons by value D. Convolution is used only for text tokenization

Correct Answer: B. Convolution applies learnable filters to local regions of input data

Semester Exam Deep Learning InternationalHard

Which misconception about Dropout is avoided by the correct answer?

A. Dropout is a file handling operation B. Dropout randomly disables units during training to reduce co-adaptation C. Dropout permanently deletes neurons during inference D. Dropout always increases overfitting

Correct Answer: B. Dropout randomly disables units during training to reduce co-adaptation

Semester Exam Deep Learning InternationalHard

Which misconception about Generalization is avoided by the correct answer? Focus on the most precise definition.

A. Generalization can only be measured on training loss B. Generalization is the ability to perform well on unseen examples C. Generalization means memorizing every training sample D. Generalization is unrelated to overfitting

Correct Answer: B. Generalization is the ability to perform well on unseen examples

Semester Exam Deep Learning InternationalHard

When applying deep learning to real-world data, how should CNN pooling be understood?

A. Pooling is the same as dropout B. Pooling changes class labels C. Pooling reduces spatial dimensions while keeping important features D. Pooling increases image resolution in every layer

Correct Answer: C. Pooling reduces spatial dimensions while keeping important features

Semester Exam Deep Learning InternationalHard

When applying deep learning to real-world data, how should Batch normalization be understood?

A. Batch normalization replaces the optimizer entirely B. Batch normalization is only used for text spelling correction C. Batch normalization normalizes intermediate activations to stabilize and speed up training D. Batch normalization stores batches in a database

Correct Answer: C. Batch normalization normalizes intermediate activations to stabilize and speed up training

Semester Exam Deep Learning InternationalHard

When applying deep learning to real-world data, how should Overfitting be understood? Focus on the most precise definition.

A. Overfitting always improves test accuracy B. Overfitting occurs only in linear regression C. Overfitting occurs when a model learns noise or specific training patterns too strongly D. Overfitting means the model is too simple for the data

Correct Answer: C. Overfitting occurs when a model learns noise or specific training patterns too strongly

Deep Learning MCQs