Deep Learning

Deep Learning is a a sub-field of Machine Learning, using Deep Neural Networks(DNN). the word “Deep” implies refers to the depth of layers in a Artificial Neural Networks (ANN).

Deep Learning is better than traditional machine learning when we have high dimensional data; where we have large inputs and outputs. Deep Learning is the best solution when we try to mimics the human cognition.


Deep Neural Networks are capable of Feature learning.


  • Input layer, Hidden Layer, and Output layer:dl-layers.png
    • The first layer is the input layer. Each node in this layer takes an input and then passes its output as the input to each node in the next layer.
    • Middle layers are called hidden layer.
    • Last layers are output nodes.
  • Backpropagation(backward propagation of errors): This helps to calculate the Gradient of Loss Functions with respect to all the weights in the network. Backward propagation of errors helps to calculate the gradient of a loss function with respect to all the weights in the network.
  • Chain Rule: The chain rule is a way to compute the derivative of a function whose variables are themselves functions of other variables.
  • Activation Functions
  • Weights Initialization

Concerns in building a Deep Learning Model:

  • Performing Experiments
  • Making changes to the model
  • Keeping track of changes and results