Google Machine Learning BootCamp (4)

Deep Neural Networks

Summary

• shallow learning model vs deep learning model
• deep learning model has multiple hidden layers
• notation summary
  • example for 4 layers (3 hidden) NN
    • L : num of layers

      L = 4
      

    • n[l] : num of units in layer l

      n[1] = 5
      n[2] = 5
      n[3] = 3
      n[4] = n[L] = 1
      n[0] = nx = 3
      

    • a[l] : activation in layer l (gl)
    • w[l]
    • b[l]
    • x = a[0]
    • y = a[L]
• forward propagation and its vectorizing result
• in deep learning, right calculating for dimension is very important. it can less bugs
  • z[l], a[l] : (n[l], 1)
  • Z[l], A[l] : (n[l], 1)
  • dz[l], da[l] : (n[l], m)
• building block of deep NN
  • concenpt of cache (z[l])
• parameters and hyperparameters