Session 1: Introduction to Python

No lectures available

Session 2: Introduction to Logistic Regression

No lectures available

Session 3: Introduction to Artificial Neural Network

1. History of Neural networks and Deep Learning
2. How Biological Neurons work?
3. Growth of biological neural networks
4. Diagrammatic representation: Logistic Regression and Perceptron
5. Multi-Layered Perceptron (MLP)
6. Notation
7. Training a single-neuron model
8. Training an MLP: Chain Rule
9. Training an MLP:Memoization
10. Backpropagation
11. Activation functions
12. Vanishing Gradient problem
13. Bias-Variance tradeoff

Session 4: Deep Multi-layer perceptrons

1. Deep Multi-layer perceptrons:1980s to 2010s
2. Dropout layers & Regularization
3. Rectified Linear Units (ReLU)
4. Weight initialization
5. Batch Normalization
6. Optimizers:Hill-descent analogy in 2D
7. Optimizers:Hill descent in 3D and contours
8. SGD Recap
9. Batch SGD with momentum
10. Nesterov Accelerated Gradient (NAG)
11. Optimizers:AdaGrad
12. Optimizers : Adadelta andRMSProp
13. Adam
14. Which algorithm to choose when?
15. Gradient Checking and clipping
16. Softmax and Cross-entropy for multi-class classification
17. How to train a Deep MLP?

Session 5: Convolutional Neural Network

1. Biological inspiration: Visual Cortex
2. Convolution:Edge Detection on images
3. Convolution:Padding and strides
4. Convolution over RGB images
5. Convolutional layer
6. Max-pooling
7. CNN Training: Optimization
8. Receptive Fields and Effective Receptive Fields
9. ImageNet dataset
10. Data Augmentation
11. Convolution Layers in Keras
12. AlexNet
13. VGGNet
14. Residual Network
15. Inception Network
16. What is Transfer learning

Session 6: Recurrent Neural Network

1. Why RNNs?
2. Recurrent Neural Network
3. Training RNNs: Backprop
4. Types of RNNs
5. Need for LSTM/GRU
6. LSTM
7. GRUs
8. Deep RNN
9. Bidirectional RNN