{{{credits}}}
| L | T | P | C |
| 3 | 0 | 0 | 3 |
- To understand the basics of deep neural networks
- To understand CNN and RNN architectures of deep neural networks
- To comprehend the advanced deep learning models
- To learn deep learning algorithms and their applications to solve real world problems
{{{unit}}}
| Unit I | Deep Networks Basics | 9 |
Linear Algebra: Scalars – Vectors – Matrices and tensors; Probability Distributions – Gradient-based Optimization – Machine Learning Basics: Capacity – Overfitting and underfitting – Hyperparameters and validation sets – Estimators – Bias and variance – Stochastic gradient descent – Challenges motivating deep learning; Deep Networks: Deep feedforward networks; Regularization – Optimization
{{{unit}}}
| Unit II | Convolutional Neural Networks | 9 |
Convolution Operation – Sparse Interactions – Parameter Sharing – Equivariance – Pooling – Convolution Variants: Strided – Tiled – Transposed and dilated convolutions; CNN Learning: Nonlinearity Functions – Loss Functions – Regularization – Optimizers – Gradient Computation – CNN through Visualization; CNN Architectures: LeNet – AlexNet – VGGnet – ResNet – ResNeXt
{{{unit}}}
| Unit III | Recurrent Neural Networks | 9 |
Unfolding Graphs – RNN Design Patterns: Acceptor – Encoder – Transducer; Gradient Computation – Sequence Modeling Conditioned on Contexts – Bidirectional RNN – Sequence to Sequence RNN – Deep Recurrent Networks – Recursive Neural Networks – Long Term Dependencies; Leaky Units: Skip connections and dropouts; Gated Architecture: LSTM – Gated RNN
{{{unit}}}
| Unit IV | Autoencoders and Generative Models | 9 |
Autoencoders: Undercomplete autoencoders – Regularized autoencoders – Stochastic encoders and decoders – Learning with autoencoders; Representation Learning: Unsupervised pretraining – Transfer learning and domain adaptation; Deep Generative Models: Variational autoencoders – Generative adversial networks
{{{unit}}}
| Unit V | Deep Learning with TensorFlow | 9 |
TensorFlow: Basics – Optimizers – XOR implemetation – Multi-class classification; CNN: Components of CNN – Backpropagation – Dropout layer – Digit recognition – Solving real-world problems; NLP Using RNN: Word2Vec – Next-word prediction and Sentence completion.
\hfill Total: 45
After the completion of this course, students will be able to:
- Understand basics in deep neural networks (K2)
- Apply Convolution Neural Network for real-world problems in image processing (K3)
- Apply Recurrent Neural Network and its variants for text analysis (K3)
- Understand the concepts in autoencoders and generative models (K2)
After the completion of this course, students will be able to:
- Explain the basics in deep neural networks (K2)
- Apply Convolution Neural Network for real-world problems in image processing (K3)
- Apply Recurrent Neural Network and its variants for text analysis (K3)
- Explain the concepts of autoencoders and generative models (K2)
- Ian Goodfellow, Yoshua Bengio, Aaron Courville, “Deep Learning”, MIT Press, 2016.
- Salman Khan, Hossein Rahmani, Syed Afaq Ali Shah, Mohammed Bennamoun, “A Guide to Convolutional Neural Networks for Computer Vision”, Synthesis Lectures on Computer Vision, Morgan & Claypool publishers, 2018.
- Yoav Goldberg, “Neural Network Methods for Natural Language Processing”, Synthesis Lectures on Human Language Technologies, Morgan & Claypool publishers, 2017.
- Santanu Pattanayak, “Pro Deep Learning with TensorFlow: A Mathematical Approach to Advanced Artificial Intelligence in Python”, Apress, 2017.
| PO1 | PO2 | PO3 | PO4 | PO5 | PO6 | PO7 | PO8 | PO9 | PO10 | PO11 | ||
| K3 | K6 | K6 | K6 | K6 | ||||||||
| CO1 | K2 | 2 | 1 | 1 | 1 | 1 | ||||||
| CO2 | K3 | 3 | 2 | 2 | 2 | 2 | ||||||
| CO3 | K3 | 3 | 2 | 2 | 2 | 2 | ||||||
| CO4 | K2 | 2 | 1 | 1 | 1 | 1 | ||||||
| Total | 10 | 6 | 6 | 6 | 6 | |||||||
| Course Mapping | 3 | 2 | 2 | 2 | 2 |