Feed Forward Neural Network Visualizer

Feed Forward Algorithm (Simplified)

        
        // Initialize weights and biases randomly
        weights_input_hidden = random_matrix(inputNodes, hiddenNodes)
        weights_hidden_output = random_matrix(hiddenNodes, outputNodes)
        bias_hidden = random_vector(hiddenNodes)
        bias_output = random_vector(outputNodes)

        function feedForward(inputValues):
            // Hidden layer calculation
            hidden_layer_input = dot_product(inputValues, weights_input_hidden) + bias_hidden
            hidden_layer_output = sigmoid(hidden_layer_input) // Activation function

            // Output layer calculation
            output_layer_input = dot_product(hidden_layer_output, weights_hidden_output) + bias_output
            output_layer_output = sigmoid(output_layer_input) // Activation function

            return output_layer_output

        function epoch():
            // For visualization purpose, we are simulating learning.
            // In real scenario, this would involve backpropagation and weight updates.

            current_outputs = feedForward(current_inputs)

            // In a real learning scenario:
            // Calculate error (loss)
            // Adjust weights and biases to reduce error (using backpropagation)

            // For visualization, we might just nudge the output values
            // towards some target or show a change.