-
Notifications
You must be signed in to change notification settings - Fork 0
/
decode_k.py
187 lines (165 loc) · 6.52 KB
/
decode_k.py
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
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
import os
import csv
import scipy.sparse
import numpy as np
import pandas as pd
from tqdm import tqdm
from sklearn.metrics import precision_score, recall_score, precision_recall_curve, average_precision_score, \
roc_auc_score, accuracy_score, f1_score, precision_recall_fscore_support
from sklearn.preprocessing import normalize, minmax_scale
import torch
import torch.nn.functional as F
# import torch_geometric.utils
from torch_geometric.nn.models import GCN, GraphSAGE, GAT
from torch_geometric.utils import k_hop_subgraph, to_dense_adj, subgraph, to_scipy_sparse_matrix, is_undirected
import torch_geometric.transforms as T
from configs import get_arguments
from load_datasets import get_dataset
from attacks.link_attack import Attacker, auc_node
from attacks.link_teller import LinkTeller
from models import GAT_ad
args = get_arguments()
dataset_name = args.dataset.lower()
epsilon = args.epsilon
if args.eval_true or args.defense != 1:
epsilon = None
dataset = get_dataset('./datasets', dataset_name, epsilon=epsilon)
data = dataset.data
normalize_feat = T.NormalizeFeatures()
data = normalize_feat(data)
print(is_undirected(data.edge_index))
print(data.num_edges/2)
if args.model.lower() == 'gcn':
gnn = GCN(in_channels=dataset.num_node_features,
hidden_channels=args.hidden_channels,
num_layers=args.num_layers,
out_channels=dataset.num_classes,
dropout=args.dropout,
jk='last')
elif args.model.lower() == 'sage':
gnn = GraphSAGE(in_channels=dataset.num_node_features,
hidden_channels=args.hidden_channels,
num_layers=args.num_layers,
out_channels=dataset.num_classes,
dropout=args.dropout,
jk='last',
aggr='max')
elif args.model.lower() == 'gat':
gnn = GAT(in_channels=dataset.num_node_features,
hidden_channels=args.hidden_channels,
num_layers=args.num_layers,
out_channels=dataset.num_classes,
dropout=args.dropout,
jk='last',
heads=8)
else:
raise NotImplementedError('GNN not implemented!')
model_dir = './src'
if not os.path.exists(model_dir):
os.mkdir(model_dir)
model_name = dataset_name + '_' + args.model.lower() + '_l' + str(args.num_layers)
if args.defense == 1:
model_name += '_ep' + '%.1f' % args.epsilon
elif args.defense == 2:
gnn = GAT_ad(in_channels=dataset.num_node_features,
hidden_channels=args.hidden_channels,
num_layers=args.num_layers,
out_channels=dataset.num_classes,
dropout=args.dropout,
jk='last')
model_name += '_ad' + '%.1f' % args.beta
gnn.load_state_dict(torch.load(os.path.join(model_dir, model_name + '.pt')))
gnn.eval()
out = gnn(data.x, data.edge_index)
pred = out.argmax(dim=-1)
correct = (pred[data.test_mask] == data.y[data.test_mask]).sum()
acc = torch.div(correct / data.test_mask.sum(), 1e-4, rounding_mode='floor') * 1e-4
print(f'Accuracy: {acc:.4f}\n')
result_dir = os.path.join('influence_values', dataset_name)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
attacker = Attacker(gnn, data.edge_index, data.x, sample=args.sample)
attacker.update_influence_matrix(result_dir, defense_ep=epsilon)
if args.attacker == 0:
if args.combo:
res_name = os.path.join(result_dir, 'attacker' + str(args.attacker) + '_' +
model_name + '_c.npz')
elif args.efficient:
res_name = os.path.join(result_dir, 'attacker' + str(args.attacker) + '_' +
model_name + '_e.npz')
else:
res_name = os.path.join(result_dir, 'attacker' + str(args.attacker) + '_' +
model_name + '.npz')
elif args.attacker == 1:
res_name = os.path.join(result_dir, 'attacker' + str(args.attacker) + '_' +
model_name + '.npz')
elif args.attacker == 2:
res_name = os.path.join(result_dir, 'attacker' + str(args.attacker) + 'pos' + str(args.pos_type) + '_' +
model_name + '.npz')
else:
raise ValueError(f'attacker={args.attacker} error')
is_whole_graph = False
if args.targets is None:
targets = range(data.x.shape[0])
is_whole_graph = True
elif type(args.targets) is int:
targets = [args.targets]
elif len(args.targets) == 2:
targets = range(args.targets[0], args.targets[1])
else:
raise ValueError()
del gnn.lin
adj_true = to_dense_adj(data.edge_index, max_num_nodes=data.num_nodes).squeeze(0).detach().numpy().astype(bool)
emb = gnn(data.x, data.edge_index)
adj = torch.matmul(emb, emb.t()) * torch.tensor(attacker.influence_matrix)
adj_ori = adj.detach().numpy()
adj_true = adj_true[targets, :]
adj_true = adj_true[:, targets]
adj_ori = adj_ori[targets, :]
adj_ori = adj_ori[:, targets]
adj_topk = adj_ori + adj_ori.transpose()
adj_value = normalize(adj_ori, norm='max', axis=1)
adj_norm = adj_value + adj_value.transpose()
topk = [0.25, 0.5, 0.75, 1, 1.25, 1.5]
precision_topk = [100]
precision_norm = [100]
recall_topk = [0]
recall_norm = [0]
for i, k in enumerate(topk):
print('k:', k)
print('---')
'''top-k'''
print('Top-k')
y_idx = np.argpartition(adj_topk, -int(data.num_edges * k),
axis=None)[-int(data.num_edges * k):]
y_pred = np.zeros_like(adj_topk).reshape(-1)
y_pred[y_idx] = 1
y = adj_true.reshape(-1)
result = precision_recall_fscore_support(y, y_pred)
p = result[0][1]
r = result[1][1]
precision_topk.append(p * 100)
recall_topk.append(r * 100)
print('Precision, recall, F1:', result)
print('Precision, recall: {:.2f}, {:.2f}'.format(p*100//0.01*0.01, r*100//0.01*0.01))
'''normalization'''
print('\nNormalization')
y_idx = np.argpartition(adj_norm, -int(data.num_edges * k),
axis=None)[-int(data.num_edges * k):]
y_pred = np.zeros_like(adj_norm).reshape(-1)
y_pred[y_idx] = 1
y = adj_true.reshape(-1)
result = precision_recall_fscore_support(y, y_pred)
p = result[0][1]
r = result[1][1]
precision_norm.append(p * 100)
recall_norm.append(r * 100)
print('Precision, recall, F1:', result)
print('Precision, recall: {:.2f}, {:.2f}'.format(p*100//0.01*0.01, r*100//0.01*0.01))
print()
float_formatter = "{:.1f}".format
np.set_printoptions(formatter={'float_kind': float_formatter})
print('Precision topk:', precision_topk)
print('Recall topk:', recall_topk)
print('Precision norm:', precision_norm)
print('Recall norm:', recall_norm)