This Deep Learning MCQ test evaluates your understanding of core concepts like neural networks, training, and optimization.

Deep Learning MCQ 

It helps avoid the vanishing gradient problem

What is a major benefit of using ReLU as an activation function?

Which technique helps prevent overfitting in deep neural networks?

Gradients become too small for deep layers to update weights effectively

What happens in the vanishing gradient problem?

Captures long-range dependencies using attention mechanisms

What is the main advantage of Transformer architecture?

They suffer from vanishing/exploding gradients on long sequences

What is a key limitation of vanilla RNNs?

What is the main goal of the backpropagation algorithm?

It randomly deactivates neurons during training

What does the dropout technique do in deep learning models?

It speeds up training and stabilizes learning

Why is batch normalization used in neural networks?

Handles long-term dependencies in sequences

What is a key benefit of LSTM networks in deep learning?

Gradients become too small to update weights in earlier layers

What is the vanishing gradient problem in deep learning?

One complete pass through the training dataset

What does one epoch in deep learning training mean?

Captures spatial hierarchies using convolution layers

What is the main characteristic of a Convolutional Neural Network (CNN)?

To introduce non-linearity into the network

What is the primary purpose of an activation function in a neural network?

It allows only positive values, improving convergence

Why is ReLU preferred over sigmoid in deep networks?

Image classification and feature extraction