BAMU · Third Year B.E./B.Tech Artificial Intelligence and Data Science · NEP 2020
Neural Network and Deep Learning
PCCT-3 · Semester VI · 3 credits · CIA 40 + SEE 60
Module 1: Foundations of Deep Learning
15 hoursMachine and deep learning, supervised and unsupervised learning, the bias–variance trade-off, hyperparameters, under- and overfitting and regularisation; the history, advantages and challenges of deep learning, architectural principles, applications, and tools such as TensorFlow, Keras, PyTorch, Caffe and Shogun.
Examined in the 60-mark end-semester paper
- Hyperparameters in deep learning
- Regularisation and dropout
- Deep learning frameworks: TensorFlow, Keras and PyTorch
Module 2: Deep Neural Networks
15 hoursThe biological neuron, the perceptron and multilayer feed-forward networks; forward and backward propagation; linear, sigmoid, tanh, hard tanh, softmax and ReLU activations; loss functions for regression, classification and reconstruction; learning rate, momentum and sparsity; XOR, gradient descent, vanishing and exploding gradients, and sentiment analysis with PyTorch.
Examined in the 60-mark end-semester paper
- Biological and artificial neurons
- Perceptron learning rule
- Multilayer perceptron
- Activation functions
- Loss functions
- Backpropagation
- Gradient descent variants
- Vanishing and exploding gradients
Module 3: Convolutional Neural Networks; Recurrent and Generative Models
15 hoursCNN structure — padding, strides, ReLU, pooling, fully connected layers, local response normalisation — and training; unfolded graphs, RNNs, bidirectional and deep RNNs, encoder–decoder models and recursive networks; Boltzmann machines, deep belief networks and GANs with their types and applications.
Examined in the 60-mark end-semester paper
- Convolution and feature maps
- Padding, stride and pooling
- Training a CNN
- Recurrent neural networks
- Bidirectional and deep RNNs
- Encoder-decoder sequence models
- Boltzmann machines and deep belief networks
- Types of GANs
Module contents are summarised from the official BAMU syllabus . AiCanCode.org is not affiliated with Dr. Babasaheb Ambedkar Marathwada University (BAMU).