public/subjects/ai/neural-networks/audit

Exercise 0: Environment and libraries

The exercise is validated if all questions of the exercise are validated.

Activate the virtual environment. If you used conda run conda activate your_env.

Run python --version.

Does it print Python 3.x? x >= 8

Do import jupyter and import numpy run without any error?

---

---

Exercise 1: The neuron

For question 1, does this code:

    neuron = Neuron(0,1,4)
    neuron.feedforward(2,3)

return 0.9990889488055994?

---

---

Exercise 2: Neural network

For question 1, is the output 0.9524917424084265?

---

---

Exercise 3: Log loss

For question 1, is the output 0.5472899351247816?

---

---

Exercise 4: Forward propagation

The exercise is validated if all questions of the exercise are validated

For question 1, is the output the following?

Bob: 0.7855253278357536
Eli: 0.7771516558846259
Tom: 0.8067873659804015
Ryan: 0.7892343955586032

For question 2, is the logloss for the 4 students 0.5485133607757963?

---

---

Exercise 5: Regression

The exercise is validated if all questions of the exercise are validated

For question 1, is the output 7?

For question 2, are the outputs the following?

Bob: 14.918863163724454
Eli: 14.83137890625537
Tom: 15.086662606964074
Ryan: 14.939270885974128

For question 3, is the MSE 10.237608699909138?