Questions? 91-7034206363, 7034256363, 7034286363, 6238743273

Registered with Ministry Of Corporate Affairs - Government of India,
(CIN: U72900KL2016PTC046479, Dated: 05/08/2016)

Deep Learning Syllabus

Deep Learning (1 Month)
  • Perceptron, Perceptron connections, Activation functions, Heavy sided, Sigmoid, logit
  • Perceptron learning algorithm, rate and epoch, Binary classification with Perceptron
  • Document classification with Perceptron, Limitations of Perceptron, Linearly seperable concept
  • ANN: Artificial Neural networks, Neural Net, Neural Decision boundaries, ANN for XOR
  • Feedforward and feedback ANN, Multilayer Perceptron, Hidden layer, Network Diagram, MSE cost function
  • Minimising cost function, Backpropagation, Training FF with BP, Forward Propagation steps, Error calculation
  • Updating weights, Classifying handwritten digits, Keras, Features of Keras, Keras backend
  • Keras Neural Networks, Layers, Deep learning trends, probabilistic modeling, Logistic regression using keras
  • Kernel methods, Decision boundary, Kernel function, Learning Deep, Keras workflow, sequntial layers
  • Compilation, Fit, Using Layers, optimiser, Preprocessing, Preparing the labels, Overfitting
  • Tensors, 1D, 2D and 3D tensors, Image processing, Time Series and uses, Video data, tensor operations
  • Elementwise operations, Broad casting, Tensor dot, dot operation, Tensor re-shaping, Transpose
  • Geometric interpretation of tensor operations, Deep learning algorithm, engine of neural networks
  • Derivatives, SGD, mini-batch, Issues with SGD, batch SGD, loss surface, GD intuition, Optimizers
  • Momentum, Local and Global Minimum, Naive implementation, Chaining derivatives, Topology selection
  • Choosing Loss function, Projects: Classifying reviews, Classifying news wires, Predicting house prices
  • MLPs, RNN, CNN, ReLU, Regulariztion in Deep Networks, Dropout, Softmax activation, tanh, Sigmoid
  • Choices of Lossfunction, Plotting the model, Conv2D, Convolution, Convolution process, Pooling
  • MaxPoolin2D, Pooling and Compression, Performance evaluation, RNN digit classification, SimpleRNN, LSTM Model
  • Sentiment Analysis using LSTM, GRU, Stacked RNN, Keras Functional API, Flattend and Dropout
  • Tensorflow, Dimensions, Immutability, Variables, Computational Graph, Sessions, Correct the Initialization errors
  • Global Variable Initializser, Assign values to Variables, Guidelines and Tensor types, Feed values, run graph
  • Eval and run, Evaluating multiple nodes, Dependencies between nodes, Solutions to multiple runs
  • clossing session, NN Architecture, using activation function in Tensorflow, Swish Activation
  • Project: US Census Bureau data sets, re-structure for iterations, Project: Fashion MINIST data set classification