Coursework

Learn

A three-part syllabus the group teaches internally, from first principles to applied models — reference reading, lecture slides, and sample code for every week.

Neural Networks with TensorFlow · 4 weeks

Instructor: Sanasam Ranbir Singh (TA: Jennil Thiyam)

Week 1Introduction to Machine Learning

Reference Books:

  1. Machine Learning by Tom M. Mitchell click
  2. Introduction to Machine Learning by Alex Smola and S.V.N. Vishwanathan click

Lessons

  1. Lesson 1: Course Introduction PPT
  2. Lesson 2: Introduction to Machine Learning
    1. What is Machine Learning? PPT
    2. My First Machine Learning Model?? PPT
  3. Lesson 3: Different Classifier Methods
    1. Bayesian and Naive Bayes Classifiers PPT
    2. k-nearest neighbor and Centroid based classifier Classifier PPT
    3. Decision Tree PPT
      1. Some of the examples and figures are taken from the book Tom M. Mitchell, Machine Learning, McGraw-Hill, 1997 and slides from Allan Neymark CS157B – Spring 2007
    4. Support Vector Machine PPT

Sample Programs

  1. Python Tutorial - Part I code in .ipynb format)
  2. Python Tutorial - Part II code in .ipynb format)
  3. Building Classifiers using Scikit-learn code in .ipynb format)
  4. Image Classification code

External Reading Resources

  1. What is machine learning? click
  2. Numpy Installation click
  3. Numpy Manual click
  4. Basic operations on Numpy arrays click
  5. Installation of Anaconda on a Windows system:click
  6. Installation of Anaconda on a Linux system:click
  7. Scikit learn installation:click
  8. Various supervised learning models provided by the Scikit learn click
Week 2Introduction to Neural Networks

Reference Books:

  1. Neural Networks and Learning Machines by Simon Haykin click

Lessons

  1. Lesson 1: Neural Network and Human Brain PPT
  2. Lesson 2: Multilayer Perceptron PPT
  3. Lesson 3: Parameters of Multi-layer perceptron PPT
  4. Lesson 4: Understanding Backpropagation PPT
  5. Lesson 5: Loss Functions and Their Gradient estimates
    1. Mean Square Loss function and its gradient PPT
    2. Cross entropy loss functions and its gradient PPT
  6. Lesson 6: Activation Functions and Theirs Gradient Estimates PPT

Sample Codes

  1. Understanding Forward Pass in MLP click
  2. Understanding different Activation Functions click
  3. Understanding dirivatives of different Activation Functions click
  4. Understanding Backpropagation through program click
Week 3Tensors and TensorFlow Programming

Reference Links:

  1. TensorFlow click
  2. TensorFlow Tutorial click

Lessons

  1. Lesson 1: Introduction to Tensorflow PPT
  2. Lesson 2: Introduction to Tensors
    1. What are Tensors and how to define them? PPT
  3. Lesson 3: Operations on Tensors
    1. Operations on Tensors PPT
    2. Compatibility of two tensors PPT
  4. Lesson 4: Variable Type Tensors PPT
  5. Lesson 5: Reshaping Tensors PPT
  6. Lesson 6: Extracting Data from a Tensor
    1. Extract Contiguous elements from a Tensor PPT
    2. Extract Non-contiguous elements from a Tensor PPT
    3. Extract Minimum/Maximum element from a Tensor, Join tensors PPT
  7. Lesson 7: Implementing Neural Network
    1. Implement Simple two layer MLP with and without Tensorflow PPT
    2. Implement MLP with Tensorflow – Sequential and Functional Model PPT

Sample Codes

  1. Tensor Definition click
  2. Operation on Tensor click
  3. More on Tensor click
  4. Matrix Operation on Tensor click
  5. Accessing Data from Tensor click
  6. MLP with Keras click
Week 4Multilayer Perceptrons (MLP) with Real-World Applications

Reference Links:

  1. TensorFlow click
  2. TensorFlow Tutorial click

Lessons

  1. TensorFlow click
  2. TensorFlow Tutorial click

Lessons

  1. Lesson 1: Vehicle Type Classification with MLP – Explicit Features PPT
  2. Lesson 2: Classification with MLP - Image PPT
  3. Lesson 3: Regression with MLP PPT
  4. Lesson 4: Representation Learning with MLP and Clustering PPT

Demo Programs click

Deep Learning Basics · 4 weeks

Instructor: Sanasam Ranbir Singh (TA: Jennil Thiyam)

Week 1Sequential Neural Models

Reference Books:

  1. Deep Learning, Ian Goodfellow and Yoshua Bengio and Aaron Courville Click

Lessons

  1. Introduction to Course slide
  2. Bias in Neural Network slide
  3. Recurrent Neural Network (RNN) slide
  4. Backpropagation in RNN slide
  5. Long Short Term Memory (LSTM) slide
  6. LSTM Forward Pass slide
  7. Gated Recurrent Unit slide
  8. Implementation slide

Sample Programs slide

Live Sesson: Backpropagation in LSTM slide

Week 2Convolutional Neural Networks and Encoder–Decoder Models

Reference Books:

  1. Deep Learning, Ian Goodfellow and Yoshua Bengio and Aaron Courville Click

Lessons

  1. Introduction to Convolutinal Neural Network slide
  2. CNN as Sparsedly Connected MLP slide
  3. CNN Backpropagation slide
  4. Encoder Decoder Model slide

Live Session

  1. 1D, 2D, 3D Convolution
  2. Up-pooling and Upsampling

Sample Programs

  1. Image Classification with CNN slide
Week 3Representation Learning

Reference Books:

  1. Deep Learning, Ian Goodfellow and Yoshua Bengio and Aaron Courville Click

Lessons

  1. Introduction to Representation Learning slide
  2. Word Embedding slide
  3. Network Embedding slide
  4. Principal Component Analysis slide
  5. Image Embedding slide

Live Session

  1. Graph Neural Network

Sample Programs

  1. Word Embedding Download
Week 4Real-World Applications

Reference Links:

  1. TensorFlow click
  2. TensorFlow Tutorial click

Lessons

  1. Lesson 1: Image Embedding and Various Image classification models slides
  2. Lesson 2: Text Embedding PPT
  3. Lesson 3: Fake News Detection - Image slides
  4. Lesson 4: POS PPT

Sample Programs

  1. Image embedding and classification click

Advanced Deep Learning · 4 weeks

Instructor: Sanasam Ranbir Singh (TA: Jennil Thiyam)

Week 1Attention and Transformers

Slides

  1. Click

Sample Code

  1. Attention Click
  2. Neural Machine Translation/Transliteration Click
  3. GAN Click
  4. Face swap Click
Week 2Transfer Learning and Multitask Learning

Slides

  1. Click

Sample Code

  1. Attention Click
  2. Neural Machine Translation/Transliteration Click
  3. GAN Click
  4. Face swap Click
Week 3Generative Adversarial Networks

Slides

  1. Click

Sample Code

  1. Attention Click
  2. Neural Machine Translation/Transliteration Click
  3. GAN Click
  4. Face swap Click
Week 4Solving Real-World Problems

Slides

  1. Click

Sample Code

  1. Attention Click
  2. Neural Machine Translation/Transliteration Click
  3. GAN Click
  4. Face swap Click