How to use a self-made loss function for a simple Neural Network in Matlab?

I have been using

net = feedforwardnet(10) %or

net = fitnet(10)

for my regression problem in Matlab. I am using simple networks with 1 or 2 layers and ReLU activation function (net.transferFcn = 'poslin').

But now, I have to use a self-made custom loss functions instead of 'mse' (mean squared error). Could you please let me know how can I do this.

I have found the following document regarding using custom layers and loss functions:https://www.mathworks.com/help/deeplearning/ug/define-custom-regression-output-layer.html

But this is regarding to complex Neural Networks like CNN. I could not understand how to simplify this for a normal deep neural network.

Answer 1

[yanqi liu]

yes，sir，may be it same on：https://ww2.mathworks.cn/matlabcentral/answers/1618945-how-to-use-a-self-made-loss-function-for-a-simple-neural-net

net=newff([0,1],[5,1],{'tansig','logsig'},'traingd')

Warning: NEWFF used in an obsolete way.

          See help for NEWFF to update calls to the new argument list.
 

net =

    Neural Network
 
              name: 'Custom Neural Network'
          userdata: (your custom info)
 
    dimensions:
 
         numInputs: 1
         numLayers: 2
        numOutputs: 1
    numInputDelays: 0
    numLayerDelays: 0
 numFeedbackDelays: 0
 numWeightElements: 16
        sampleTime: 1
 
    connections:
 
       biasConnect: [1; 1]
      inputConnect: [1; 0]
      layerConnect: [0 0; 1 0]
     outputConnect: [0 1]
 
    subobjects:
 
             input: Equivalent to inputs{1}
            output: Equivalent to outputs{2}
 
            inputs: {1x1 cell array of 1 input}
            layers: {2x1 cell array of 2 layers}
           outputs: {1x2 cell array of 1 output}
            biases: {2x1 cell array of 2 biases}
      inputWeights: {2x1 cell array of 1 weight}
      layerWeights: {2x2 cell array of 1 weight}
 
    functions:
 
          adaptFcn: 'adaptwb'
        adaptParam: (none)
          derivFcn: 'defaultderiv'
         divideFcn: (none)
       divideParam: (none)
        divideMode: 'sample'
           initFcn: 'initlay'
        performFcn: 'mse'
      performParam: .regularization, .normalization
          plotFcns: {'plotperform', 'plottrainstate',
                    'plotregression'}
        plotParams: {1x3 cell array of 3 params}
          trainFcn: 'traingd'
        trainParam: .showWindow, .showCommandLine, .show, .epochs,
                    .time, .goal, .min_grad, .max_fail, .lr
 
    weight and bias values:
 
                IW: {2x1 cell} containing 1 input weight matrix
                LW: {2x2 cell} containing 1 layer weight matrix
                 b: {2x1 cell} containing 2 bias vectors
 
    methods:
 
             adapt: Learn while in continuous use
         configure: Configure inputs & outputs
            gensim: Generate Simulink model
              init: Initialize weights & biases
           perform: Calculate performance
               sim: Evaluate network outputs given inputs
             train: Train network with examples
              view: View diagram
       unconfigure: Unconfigure inputs & outputs
 

net.performFcn

ans = 'mse'

we can find the default is mse

so，make the self definition function，https://www.mathworks.com/matlabcentral/answers/64319-neural-network-how-to-use-a-custom-performance-function

Answer 2

[pathakunta]

I have been using net = feedforwardnet(10) %or net = fitnet(10) for my regression problem in Matlab. I am using simple networks with 1 or 2 layers and ReLU activation function (net.transferFcn = 'poslin'). But now, I have to use a self-made custom loss functions instead of 'mse' (mean squared error). Could you please let me know how can I do this. I have found the following document regarding using custom layers and loss functions:https://www.mathworks.com/help/deeplearning/ug/define-custom-regression-output-layer.html But this is regarding to complex Neural Networks like CNN. I could not understand how to simplify this for a normal deep neural network.