Graph Pooling Max

Problem

Aggregate all node feature vectors of a graph with elementwise max pooling.

features is a list of equal-length vectors. Return the coordinate-wise maximum, or an empty list when there are no nodes.

Hint

Max pooling preserves the strongest activation in each coordinate.

case 1

in: [[[1,5],[3,2],[0,4]]]

expected: [3,5]

case 2

in: [[[7,1],[2,9]]]

expected: [7,9]

case 3

in: [[]]

expected: []

case 4

in: [[[-1,-2],[-3,-4]]]

expected: [-1,-2]

Starter codePython

def graph_pooling_max(features):
    # Your code here
    pass

4 test cases · run your code against them in the editor