Backpropagation concept in ML Python - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to calculate the error between predicted and actual output.

ML Python

error = predicted - [1]

Drag options to blanks, or click blank then click option'

Aactual

Bpredicted

Cweights

Dbias

Attempts:

3 left

💡 Hint

Common Mistakes

Using predicted instead of actual causes zero error.

Using weights or bias here is incorrect.

The error is the difference between the predicted value and the actual value.

2fill in blank

medium

Complete the code to compute the gradient of the loss with respect to weights.

ML Python

gradient = error * [1]

Drag options to blanks, or click blank then click option'

Abias

Binput

Coutput

Dlearning_rate

Attempts:

3 left

💡 Hint

Common Mistakes

Using bias or learning rate instead of input.

Using output instead of input.

The gradient is calculated by multiplying the error by the input to the neuron.

3fill in blank

hard

Fix the error in updating the weights using gradient descent.

ML Python

weights = weights - [1] * gradient

Drag options to blanks, or click blank then click option'

Ainput

Bgradient

Cerror

Dlearning_rate

Attempts:

3 left

💡 Hint

Common Mistakes

Using gradient alone causes too large updates.

Using error or input directly is incorrect.

Weights are updated by subtracting the learning rate times the gradient.

4fill in blank

hard

Fill both blanks to calculate the derivative of the sigmoid activation function.

ML Python

sigmoid_derivative = [1] * (1 - [2])

Drag options to blanks, or click blank then click option'

Asigmoid_output

Binput

Cweights

Dbias

Attempts:

3 left

💡 Hint

Common Mistakes

Using input, weights, or bias instead of sigmoid output.

Using different variables in the two blanks.

The derivative of sigmoid is sigmoid_output times (1 minus sigmoid_output).

5fill in blank

hard

Fill all three blanks to update bias using gradient descent.

ML Python

bias = bias - [1] * [2] * [3]

Drag options to blanks, or click blank then click option'

Alearning_rate

Berror

C1

Dinput

Attempts:

3 left

💡 Hint

Common Mistakes

Using input instead of 1 for bias update.

Forgetting to multiply by learning rate or error.

Bias update uses learning rate times error times 1 (since bias input is always 1).