feat(numpy): update exercise 6 subject and audit

- Make exercise relevant to the topic of broadcasting

subjects/ai/numpy/README.md

@@ -156,7 +156,7 @@ The goal of this exercise is to learn to concatenate and reshape arrays.
 The goal of this exercise is to learn to access values of n-dimensional arrays efficiently.
 
 1. Create an 2-dimensional array size 9,9 of 1s. Each value has to be an `int8`.
-2. Using **slicing**, output this array:
+2. Using a combination of **slicing** and **broadcasting**, output this array:
 
    ```python
    array([[1, 1, 1, 1, 1, 1, 1, 1, 1],

subjects/ai/numpy/audit/README.md

@@ -186,15 +186,17 @@ https://jakevdp.github.io/PythonDataScienceHandbook/ (section: The Basics of Num
         [1, 1, 1, 1, 1, 1, 1, 1, 1]], dtype=int8)
 ```
 
-##### The solution of question 2 is not accepted if the values of the array have been changed one by one manually. The usage of the for loop is not allowed neither.
+##### The solution of question 2 is not accepted if the values of the array have been changed one by one manually.
 
 Here is an example of a possible solution:
 
-```console
-        x[1:8,1:8] = 0
-        x[2:7,2:7] = 1
-        x[3:6,3:6] = 0
-        x[4,4] = 1
+```python
+for step in range(1, 5):
+    array_len = 9 - step * 2
+    if step % 2 == 1:
+        x[step:-step, step:-step] -= np.ones(array_len, dtype='int8')
+    else:
+        x[step:-step, step:-step] += np.ones(array_len, dtype='int8')
 ```
 
 ---