Deep Learning Unit 3

How many layers does the original LeNet architecture consist of?

• 2

• 3

• 5

• 7

What is the purpose of max-pooling layers in LeNet?

• Increase spatial dimensionality

• Reduce computational complexity

• Enhance feature maps

• Add non-linearity

Which activation function is typically used in AlexNet?

• ReLU

• Sigmoid

• Tanh

• Linear

What is the purpose of dropout layers in AlexNet?

• To increase spatial dimensionality

• To reduce overfitting

• To enhance feature maps

• To add non-linearity

Which type of convolutional layers are predominantly used in VGG?

• 1x1 convolutions

• 3x3 convolutions

• 5x5 convolutions

• 7x7 convolutions

What was the primary application domain of VGG?

• Natural Language Processing

• Image Classification

• Speech Recognition

• Reinforcement Learning

What is the primary advantage of using multiple branches in the Inception module?

• It reduces computational cost.

• It increases model interpretability.

• It improves feature extraction at different scales.

• It enhances gradient flow during backpropagation.

GoogleNet introduced which type of module to perform convolutions at different scales?

• VGG module

• ResNet module

• Inception module

• AlexNet module

What is transfer learning?

• Training a model from scratch using a large dataset

• Utilizing pre-trained models and fine-tuning them for a specific task

• Generating synthetic data to improve model performance

• Using ensemble learning techniques for model training

What is the primary benefit of using transfer learning in image classification tasks?

• Faster convergence during training

• Improved generalization to unseen data

• Reduced risk of overfitting

• All the options