In Java

(2-2). (4 points) (due: 11:59pm,2/14 ) This homework is a simple example of a supervised classification problem. The cube is given. Eight corners of this cube are classified into two groups. The four corners A, E, H, and D belong to one group, and the four corners B, C, G, and F belong to the other group. The labelled data is given below. This programming assignment is to simulate Backpropagation algorithm to solve this non-linear separable problem. You can use your favorite computer language. Do not use Machine learning libraries for this question. Train your 3-layer neural network to classify this problem. You will need 3 input units in the input layer. One output unit is enough to solve this problem for two classifications (0 and 1) using a logistic sigmoid function. For this assignment, the architecture is given as 3 input units, 3 hidden units, and 1 output unit. Since the activation function is a sigmoid function, the actual output cannot be exactly either one or zero like the target output. In programming, when the actual output of output unit is larger than 0.9, we accept the output as 1 . When the actual output of output unit is less than 0.1, we accept the output as 0 . You will start with your small random initial weights for this problem. You should try different learning coefficients to run your program. As the result of training, your program should print vour weighted sum and actual output (produced from Sigmoid function) for each input (000,100,,111) to verify your actual output is close enough to target output. Your program also should print the final connections weights as the result of training. This programming assignment is to simulate Backpropagation algorithm to solve this non-linear separable problem. You can use your favorite computer language. Do not use Machine learning libraries for this question. Train your 3-layer neural network to classify this problem. You will need 3 input units in the input layer. One output unit is enough to solve this problem for two classifications ( 0 and 1) using a logisti sigmoid function. For this assignment, the architecture is given as 3 input units, 3 hidden units, and 1 output unit. Since the activation function is a sigmoid function, the actual output cannot be exactly either one or zero like the target output. In programming, when the actual output of output unit is larger than 0.9, we accept the output as 1 . When the actual output of output unit is less than 0.1, we accept the output as 0. You will start with your small random initial weights for this problem. You should try different learning coefficients to run your program. As the result of training, your program should print vour weighted sum and actual output (produced from Sigmoid function) for each input (000,100,,111) to verify your actual output is close enough to target output. Your program also should print the final connections weights as the result of training