Example 1:
Input: mat = [[5, 10, 8], [4, 25, 7], [3, 9, 6]]
Output: [1, 1]
Explanation: The value at index [1, 1] is 25, which is a peak because all its neighbors (10, 7, 4, 9) are smaller.

Example 2:
Input: mat = [[1, 2, 3], [6, 5, 4], [7, 8, 9]]
Output: [2, 2]
Explanation: The value at index [2, 2] is 9, which is a peak as it is greater than its neighbors (8, 4).

Disclaimer: It's highly recommend trying to solve it before looking at the solution.

Approach

Algorithm

To solve this problem we use the binary search approach.
The key idea comes from how we find a peak in a 1-D array:

For any middle position (mid), we check if it’s larger than both its neighbors, if it is, we’ve found a peak.
If mid is smaller than the element on its left, that means a peak must be somewhere to the left, so we can discard the right half.
If mid is smaller than the element on its right, then a peak must lie to the right, allowing us to discard the left half.
This method reduces the number of elements we need to consider in every step, improving efficiency.

For a 2-D array,

The search will cover the column range from 0 to col-1, where col is the total number of columns.
We choose a middle column and identify the row with the largest element in that column.
We apply similar logic as in 1-D: if this element is bigger than both its side neighbors, we’ve found the peak.
If the left neighbor is bigger, we only search the left part; if the right neighbor is bigger, we search the right part.