To implement a multi-layer feedforward back-propagation algorithm (we can just call it ANN) that can Teach a Neural Network for Classification. Requirement: (on picture)

To implement a multi-layer feedforward back-propagation algorithm (we can just call it ANN) that can Teach a Neural Network for Classification". Requirement: 1) You must have three layers (input, hidden, and output) in your architecture. 2) Use Sigmoid function as your transfer function. Implement the ANN (NOT perceptron) to recognize 4 blocks image (such as one image shown below) to make a decision if an image is BRIGHT or DARK. Remember that we have 16 samples for the input. You must have at least three layers in your ANN: input layer, hidden layer, and output layer. If it contains 2, 3 or 4 white pixels, it is "bright" If it contains O or 1 white pixels, it is "dark" You may write your code in a contemporary language of your choice; typical languages would include C/C++, Java, Ada, Pascal, Smalltalk, Lisp, and Prolog. A GUI interface is preferred. Submit a hard copy of your well-commented source program and your printed output if any. In addition, please give a demo on your program in the classroom in the due day. Please do not take codes from the website. Try to understand the algorithm and implement the algorithm by your own