Module-8: Neural Networks, Computer Vision and Deep Learning

Deep Learning:Neural Networks.

• History of Neural networks and Deep Learning. 25 mins
• How Biological Neurons work? 8 mins
• Growth of biological neural networks 17 mins
• Diagrammatic representation: Logistic Regression and Perceptron 17 mins
• Multi-Layered Perceptron (MLP). 23 mins
• Notation 18 mins
• Training a single-neuron model. 28 mins
• Training an MLP: Chain Rule 40 mins
• Training an MLP:Memoization 14 mins
• Backpropagation. 26 mins
• Activation functions 17 mins
• Vanishing Gradient problem. 23 mins
• Bias-Variance tradeoff. 10 mins
• Decision surfaces: Playground 15 mins

Deep Learning: Deep Multi-layer perceptrons

• Deep Multi-layer perceptrons:1980s to 2010s 16 mins
• Dropout layers & Regularization. 21 mins
• Rectified Linear Units (ReLU). 28 mins
• Weight initialization. 24 mins
• Batch Normalization. 21 mins
• Optimizers:Hill-descent analogy in 2D 19 mins
• Optimizers:Hill descent in 3D and contours. 13 mins
• SGD Recap 18 mins
• Batch SGD with momentum. 25 mins
• Nesterov Accelerated Gradient (NAG) 8 mins
• Optimizers:AdaGrad 15 mins
• Optimizers : Adadelta andRMSProp 10 mins
• Adam 11 mins
• Which algorithm to choose when? 5 mins
• Gradient Checking and clipping 10 mins
• Softmax and Cross-entropy for multi-class classification. 25 mins
• How to train a Deep MLP? 8 mins
• Auto Encoders. 27 mins
• Word2Vec :CBOW 19 mins
• Word2Vec: Skip-gram 14 mins
• Word2Vec :Algorithmic Optimizations. 12mins

Deep Learning: Tensorflow and Keras.

• Tensorflow and Keras overview 23 mins
• GPU vs CPU for Deep Learning. 23 mins
• Google Colaboratory. 5 mins
• Install TensorFlow 6 mins
• Online documentation and tutorials 6 mins
• Softmax Classifier on MNIST dataset 32 mins
• MLP: Initialization 11 mins
• Model 1: Sigmoid activation 22 mins
• Model 2: ReLU activation. 6 mins
• Model 3: Batch Normalization. 8 mins
 • Model 4 : Dropout. 5mins
• MNIST classification in Keras. 18mins
• Hyperparameter tuning in Keras. 11 mins

Deep Learning: Convolutional Neural Nets.

• Biological inspiration: Visual Cortex 18 mins
• Convolution:Edge Detection on images. 28 mins
• Convolution:Padding and strides 19 mins
• Convolution over RGB images. 11 mins
• Convolutional layer. 23 mins
• Max-pooling. 12 mins
• CNN Training: Optimization 9 mins
• Receptive Fields and Effective Receptive Fields 8 mins
• Example CNN: LeNet [1998] 10 mins
• ImageNet dataset. 6 mins
• Data Augmentation. 8 mins
• Convolution Layers in Keras 17 mins
• AlexNet 13 mins
• VGGNet 11 mins
• Residual Network. 22 mins
• Inception Network. 19 mins
• What is Transfer learning. 23 mins
• Code example: Cats vs Dogs. 15 mins
• Code Example: MNIST dataset. 6 mins
• [Interview Question] How to build a face recognition system? 1 mins

Deep Learning: Long Short-term memory (LSTMs)

• Why RNNs? 23 mins
• Recurrent Neural Network. 29 mins
• Training RNNs: Backprop. 16 mins
• Types of RNNs. 14 mins
• Need for LSTM/GRU. 10 mins
• LSTM. 35 mins
• GRUs. 7 mins
• Deep RNN. 7 mins
• Bidirectional RNN. 12 mins
• Code example : IMDB Sentiment classification 33 mins

Deep Learning: Generative Adversarial Networks (GANs)

• Live session on Generative Adversarial Networks (GAN) 124 mins

Encoder-Decoder Models
• LIVE: Encoder-Decoder Models 82 mins

Attention Models in Deep Learning

• Attention Models in Deep Learning 84 mins

Deep Learning: Transformers and BERT

• Transformers and BERT 112 mins

Deep Learning: Image Segmentation

• Live session on Image Segmentation 95 mins
Interview Questions on Deep Learning
• Questions and Answers 30 mins

Deep Learning: Object Detection

• Object Detection 123 mins
• Object Detection YOLO V3 103 mins

Module 8: Live Sessions