Deep Learner

Uses a Neural Network to solve a Supervised Learning problem.

It employs three Neural Networks:

  • Training Network - Used to adjust the weights and biases with the help of a user-defined optimizer to minimize the loss function given in the supervised learning problem (e.g., Mean Square Error).

  • Validation Network - Used to measure the improvement (in terms of loss function values) of the training network on the validation data given in the supervised learning problem.

  • Test Network - The result of the optimisation procedure which can be used to evaluate the neural network on a test set using the test() function.