Semester –VI
Deep Learning
BTAIC601 Deep learning PCC7 3L- 1T - 0P 4 Credits
Teaching Scheme Examination Scheme
Lecture: 3 hrs./week Continuous Assessment : 20 Marks
Tutorial: 1 hr./week Mid Semester Exam: 20 Marks
End Semester Exam: 60 Marks (Duration 03 hrs.)
Pre-Requisites: Basic Knowledge of Machine learning, Soft Computing, Data Structures,
Python.
Course Objectives:
In this course, attendees will:
Understand the context of neural networks and deep learning
Have a working knowledge of neural networks and deep learning
Explore the parameters for neural networks
Use CNN and RNN for solving real world problem.
Course Outcomes:
On completion of the course, students will be able to:
CO1 Implement deep learning models in Python using the Keras/PyTorch library and train
them with real-world datasets.
CO2 Design convolution networks for image classification.
CO3 Perform regularization, training optimization, and hyperparameter selection on deep
models.
CO4 Design Recurrent Neural Networks for text and sequence classification.
CO5 Apply Generative Deep Learning for Generating images
Course Contents:
Unit 1: Introduction to Neural Network [8 Hours]
Working Of Simple Artificial Neural Network, Multilayer Perceptron, Forward Propagation
And Back Propagation Learning, Building Blocks of Deep Neural Networks, Optimization
Techniques, Gradient Descent and its variants, Derivatives, Batch Optimization, Momentum
Optimizer, RMSProp, Adam, Vectorization, Linear Regression and Logistic Regression with
Deep Neural Network.
Unit 2: Convolutional Neural Network [7 Hours]
Introduction Convolutional Neural Network, Fully Connected Network vs Convolutional
Neural Network , Building Blocks Of CNN: Filters, Convolution, Pooling, Activations Etc.
Training Procedure of CNN, Feeding Images And Videos to CNN, Different CNN
Architectures, Residual Networks, Skip Connections.
Unit 3: Transfer Learning and Effective training in Deep Net: [7 Hours]
Transfer Learning: Introduction To Transfer Learning, Need For Transfer Learning, Feature
Extraction Using Transfer Learning, Fine Tuning.
Effective Training: Bias Variance Tradeoff, Dealing With Overfitting and Underfitting, Data
Augmentation, Early Stopping, Dropout, Batch Normalization, Instance Normalization, Group
Normalization, Regularization, Hyperparameter Tuning.
Unit 4: Deep learning for text and Sequences [7 Hours]
Introduction To Sequential/Temporal Data, Sequential Models, Introduction to Recurrent
Neural Network ,Working of RNN, Representing Sequential Data using RNN, Working With
Text Data, Text Generation With LSTM, LSTM And GRU, Transformer Network.
Unit 5: Generative Deep Learning [7 Hours]
Neural Stryle Transfer ,Variational Autoencoder, Generative Adversarial Network , Classical
Supervised Tasks With Deep Learning, Image Denoising, Semantic Segmentation, Object
Detection Etc.
Text Books
1. Francois Challot, “Deep Learning with Python”, second edition.
2. Francois Challot, “Deep Learning with Pytorch”, second edition
Reference Books
1. 1 Michael Nielsen, Neural Networks and Deep Learning, 2016
2. Deep Learning- Ian Goodfelllow, Yoshua Benjio, Aaron Courville, The MIT Press
3. Pattern Classification- Richard O. Duda, Peter E. Hart, David G. Stork, John Wiley &
Sons Inc.
