-
Notifications
You must be signed in to change notification settings - Fork 5
/
wl_equivalence_classes.m
81 lines (68 loc) · 2.15 KB
/
wl_equivalence_classes.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
% WL_EQUIVALENCE_CLASSES computes 1d WL "color classses" for a graph.
%
% Given an (optionally labeled) graph G, this computes the
% Weisfeiler--Lehman equivalence classes of the nodes of G by
% iterating the one-dimensional Weisfeiler--Lehman graph
% transformation until stability is reached.
%
% This implementation uses a fast hashing-based implementation to
% compute the Weisfeiler--Lehman graph transformation; see the
% following paper for more information:
%
% Kersting, K., Mladenov, M., Garnett, R., and Grohe, M. Power
% Iterated Color Refinement. (2014). AAAI Conference on Artificial
% Intelligence (AAAI 2014).
%
% This implementation supports an initial labeling of the nodes; the
% traditional color refinement algorithm begins by labeling each node
% with the same initial color. Omitting the "labels" argument acheives
% this behavior.
%
% Usage:
%
% equivalence_classes = wl_equivalence_classes(A, labels)
%
% Inputs:
%
% A: an (n x n) unweighted adjacency matrix for the desired
% graph.
% labels: (optional) an (n x 1) vector of positive integer labels
% for each node in the graph (default: ones(n, 1))
%
% Outputs:
%
% equivalence_classes: a vector of the color classes for each node in
% the graph.
%
% See also WL_TRANSFORMATION.
% Copyright (c) 2014--2017 Roman Garnett.
function equivalence_classes = wl_equivalence_classes(A, labels)
% if no labels given, use initially equal labels
if (nargin < 2)
labels = ones(size(A, 1), 1);
end
equivalence_classes = zeros(size(labels));
% iterate WL transformation until stability
while (~is_equivalent(labels, equivalence_classes))
equivalence_classes = labels;
labels = wl_transformation(A, labels);
end
end
% checks whether two labelings are equivalent under permutation
function x = is_equivalent(labels_1, labels_2)
% until proven otherwise
x = false;
m = max(labels_1);
% different number of labels
if (m ~= max(labels_2))
return;
end
% check whether every equivalence class remains unchanged
for i = 1:m
y = labels_2(labels_1 == i);
if (any(y(1) ~= y))
return;
end
end
x = true;
end