Adding additional inputs with corresponding output into the neural network as the auxiliary input

In my project, I would like to build the neural network 2 inputs with single output. In the other input, I wish to add the corresponding output into the neural network too.

However, based the last question (https://uk.mathworks.com/matlabcentral/answers/355286-how-to-give-multiple-inputs-to-the-train-function-of-neural-network), it allows me to add inputs only in the neural network instead of the corresponding outputs. Please find the example from the link below.

x1 =  [4 5 6];
x2 =  [0 1 0];
x = {x1;x2};
t = [0 0 1];
net = feedforwardnet;
net.numinputs = 2;
net.inputConnect = [1 0; 0 1];
net = configure(net,x);
net = train(net,x,t);
view(net)

In my case, I would like to train the below data (this is just the example, the real data size has thousand sets) with 4 hidden layers.

x1 =  [4 5 6 9 10];
t1 = [0 0 1 5 3];

After that, I would like to add below additional data (diffferent matrics from first set of data) into 3rd hidden layer as the auxialiary input.

x2 = [5 6 7 8];
t2 = [2 8 9 4];

However, the previous solution doesn't work for my case as it mentioned the error about the different matrics.

Therefore, in my case, I would like to build the neural network structure which allow me to add the additional inputs and the corresponding outputs as the auxialiary input of the neural network.

I hope you guys can help me on this! Thank you!

Answer 1

[Srivardhan Gadila]

You can refer to Multiple-Input and Multiple-Output Networks and create and deep neural network with inputLayer as imageInputLayer & for the hidden layers you can use the fullyConnectedLayer.

For list of available layers refer to List of Deep Learning Layers & Deep Learning Toolbox.