Deep Learning (AL-503 (B)) - Important Questions

Explain feed-forward neural networks and sigmoid neurons with suitable diagram.

Explain the backpropagation algorithm and how it is used to update and optimize weights in a neural network during training.

Explain PCA and SVD, how decomposition is performed, and their role in representation learning.

Discuss regularization in autoencoders, including techniques and its role and impact on training.

Explain the Gradient Descent method in deep learning.

Explain denoising, sparse, contractive and variational autoencoders.

Explain the architecture and concept of Convolutional Neural Networks (CNNs) and their role in image recognition, with neat diagram.

Define overfitting and underfitting and explain regularization techniques to prevent overfitting such as dropout, weight decay and early stopping.

Define the ReLU activation function and explain its properties and significance in neural networks and CNNs.

Explain convolution operation, kernels, stride, padding and pooling in CNNs.

Explain the fundamental concepts and architectures of Deep Recurrent Neural Networks with a neat diagram.

Explain backpropagation through time (BPTT), including vanishing and exploding gradient problems and truncated BPTT.

Explain the working and architecture of Long Short-Term Memory (LSTM) networks and their advantages.

Explain directed graphical models.

Explain Restricted Boltzmann Machines with an example.

Explain GANs and applications of deep learning in vision, speech, and medical domains.