ES 413 Deep Learning

Instructor: Mayank Singh: (office: AB-4/304, email: singh.mayank@iitgn.ac.in)

Class Timings: M and F slot (AB7/207, Thursday  Friday 1530-1650) 

Tutorial: J1 slot (AB7/206, Tuesday 1400-1520)

QA sessions: Email me to book an appointment for discussing any doubts, clarifications, and concepts. 

TAs: 

1. Shruti Singh (singh_shruti@iitgn.ac.in)

2. V P Shivasankaran (vp.shivasan@iitgn.ac.in)

3. Dwip Dalal (dwip.dalal@iitgn.ac.in)

Prerequisite

Data Structures and Algorithms-I [ES 242 or equivalent]

Course Contents 

1. Introduction to Deep learning 

2. Basics of Linear Algebra

3. Probability and Information Theory

4. Numerical Computation

5. Machine Learning Basics

6. Deep Feedforward Networks

7. Regularization

8. Optimization

9. Convolutional Networks

10. Recurrent and Recursive Networks

11. Generative Adversarial Networks

12. Representation Learning

13. Applications

14. Tips and Tricks to enhance the Performance of Deep Neural Network implementations.

Lecture Slides and Additional Material

 0. Introduction and logistics of the course [Slides]

1. Introduction to Deep Learning [Slides]

2. Basics of Linear Algebra [Slides, Coding practice]

3. Probability and Information Theory [Slides]

4. Numerical Computation [Slides] [1,2]

5. Basics of ML [Slides]

6. Feedforward networks and MLP [Slides] [Notes]

7. Regularization for DL [Notes]

8. Convolution Networks [Notes]

9. RNNs [Notes]

10. Attention Mechanism and Transformers [Slides]

11. BERT and GPT [Notes]

12. Tokenization Algorithms [Notes]

Tutorials:

1: Numpy, Pandas, How to enable GPU environment [Colab]

2. PyTorch tensors, Linear algebra [Colab]

3. Probability [Colab]

4. PyTorch Datasets and Dataloaders [Colab]

5. Feed Forward Networks [Colab]

6. Optimization, Gradients, and Weight Visualization [Colab]

7. Convolutional neural networks [Colab]

8. Image Generation using GANs [Colab]

13. Recurrent Neural Networks [Colab]

14. LSTMs, GRUs, Attention [Colab]

15. Transformers [Colab] 

16. Transformer Based Models (BERT, GPT), HuggingFace API [Colab]

Grading Policy and Schedule

4 Assignments: 50% [3x10% + 1x20%] (Note: You shall get a week's time to complete the assignment. )

A0: Feedforward NNs and MLPs [Link]

A1: Computer Vision Architectures and Benchmarks [Link]

A2: NLP Architectures [Link]

A3: Prompt Engineering [Link]

1 Exam: 30% (This includes everything covered before Feb 04, 2023)

2 Surprise quizzes: 20% [10% each]

Texts/References

• Goodfellow, Ian, Yoshua Bengio, Aaron Courville, and Yoshua Bengio. Deep learning. Cambridge: MIT press, 2016. https://www.deeplearningbook.org/

• Glassner, Andrew. Deep Learning: A Visual Approach. United States: No Starch Press, 2021.

• Glassner, Andrew. Deep Learning: From Basics to Practice. Volume 1 and 2. 2021.

• Zhang, Aston, Zachary C. Lipton, Mu Li, and Alexander J. Smola. Dive into Deep Learning. 2020. https://d2l.ai/

• Krohn, Jon., Bassens, Aglaé., Beyleveld, Grant. Deep Learning Illustrated: A Visual, Interactive Guide to Artificial Intelligence. N.p.: Pearson Education, 2019.

• Nielsen, Michael A. Neural networks and deep learning. Vol. 25. San Francisco, CA: Determination press, 2015. http://neuralnetworksanddeeplearning.com/