-
Notifications
You must be signed in to change notification settings - Fork 5
/
compas.py
283 lines (232 loc) · 8.67 KB
/
compas.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
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
"""Experiments on COMPAS
Algorithms (--alg): erm, cvar, cvar_doro, chisq, chisq_doro
Use --data_mat to specify a user training set (.mat file)
Use --remove_outliers to remove outliers from the dataset
Use --noise to add noise to the dataset
"""
from dro import *
import os
import argparse
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
import scipy.io as sio
import torch
import torch.nn
from torch import Tensor
from torch import optim
from torch.nn.modules.module import Module
from torch.utils.data import DataLoader
from torch.utils.data.dataset import Dataset
import torch.nn.functional as F
import torch.backends.cudnn as cudnn
def build_model(input_dim: int) -> Module:
"""Two-layer feed-forward ReLU neural network"""
model = torch.nn.Sequential(torch.nn.Linear(input_dim, 10, bias=True),
torch.nn.ReLU(inplace=True),
torch.nn.Linear(10, 1, bias=True))
return model
def preprocess_compas(df: pd.DataFrame):
"""Preprocess dataset"""
columns = ['juv_fel_count', 'juv_misd_count', 'juv_other_count', 'priors_count',
'age',
'c_charge_degree',
'sex', 'race', 'is_recid']
target_variable = 'is_recid'
df = df[['id'] + columns].drop_duplicates()
df = df[columns]
race_dict = {'African-American': 1, 'Caucasian': 0}
df['race'] = df.apply(lambda x: race_dict[x['race']] if x['race'] in race_dict.keys() else 2, axis=1).astype(
'category')
sex_map = {'Female': 0, 'Male': 1}
df['sex'] = df['sex'].map(sex_map)
c_charge_degree_map = {'F': 0, 'M': 1}
df['c_charge_degree'] = df['c_charge_degree'].map(c_charge_degree_map)
X = df.drop([target_variable], axis=1)
y = df[target_variable]
return X, y
class MyDataset(Dataset):
def __init__(self, X, y):
super(MyDataset, self).__init__()
self.X = X
self.y = y
def __getitem__(self, item):
return self.X[item], self.y[item]
def __len__(self):
return len(self.X)
class MyLoss(object):
def __init__(self, reduction='mean'):
self.reduction = reduction
def __call__(self, outputs: Tensor, targets: Tensor) -> Tensor:
outputs = outputs.view(-1)
loss = -targets * F.logsigmoid(outputs) - (1 - targets) * F.logsigmoid(-outputs)
if self.reduction == 'mean':
loss = loss.mean()
elif self.reduction == 'sum':
loss = loss.sum()
return loss
def main():
parser = argparse.ArgumentParser()
# Dataset args
parser.add_argument('--data_mat', type=str)
parser.add_argument('--noise', default=0.0, type=float)
# Remove outliers: this only output a clean dataset but does not train on it
parser.add_argument('--remove_outliers', default=False, action='store_true')
parser.add_argument('--trim_times', default=5, type=int)
parser.add_argument('--trim_num', default=200, type=int)
# Train args
parser.add_argument('--alg', default='erm', type=str)
parser.add_argument('--batch_size', default=128, type=int)
parser.add_argument('--device', default='cpu', type=str)
parser.add_argument('--lr', default=0.01, type=float)
parser.add_argument('--alpha', type=float)
parser.add_argument('--eps', type=float)
parser.add_argument('--epochs', default=300, type=int)
parser.add_argument('--save_file', type=str)
parser.add_argument('--seed', type=int)
args = parser.parse_args()
device = args.device
if args.save_file is not None:
d = os.path.dirname(os.path.abspath(args.save_file))
if not os.path.isdir(d):
os.makedirs(d)
# Prepare the dataset
df = pd.read_csv('compas-scores-two-years.csv')
X, y = preprocess_compas(df)
input_dim = len(X.columns)
X, y = X.to_numpy().astype('float32'), y.to_numpy()
X[:, 4] /= 10
X[X[:, 7] > 0, 7] = 1 # Race: White (0) and Others (1)
domain_fn = [
lambda x: x[:, 7] == 0, # White
lambda x: x[:, 7] == 1, # Others
lambda x: x[:, 6] == 0, # Female
lambda x: x[:, 6] == 1, # Male
]
# Split the dataset: train-test = 70-30
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3,
random_state=42, shuffle=True)
if args.data_mat is not None:
# User dataset
mat = sio.loadmat(args.data_mat)
X_train = mat['X_train'].astype('float32')
y_train = mat['y_train'].flatten()
trainset = MyDataset(X_train, y_train)
testset = MyDataset(X_test, y_test)
trainloader = DataLoader(trainset, batch_size=args.batch_size, shuffle=True)
testtrainloader = DataLoader(trainset, batch_size=1024, shuffle=False)
testloader = DataLoader(testset, batch_size=1024, shuffle=False)
if args.seed is not None:
# Fix seed for reproducibility
torch.manual_seed(args.seed)
np.random.seed(args.seed)
torch.set_deterministic(True)
cudnn.benchmark = False
else:
cudnn.benchmark = True
# Label noise
if args.noise > 0.0:
l = len(y_train)
n_noise = int(l * args.noise)
r = np.random.rand(l)
r = np.argsort(r)
y_train[r[:n_noise]] = 1 - y_train[r[:n_noise]]
# Remove outliers
if args.remove_outliers:
assert args.save_file is not None
for t in range(args.trim_times):
print('=====Round {}====='.format(t + 1))
model = build_model(input_dim)
model = model.to(device)
optimizer = optim.ASGD(model.parameters(), lr=0.1)
criterion = MyLoss(reduction='none')
for epoch in range(args.epochs):
print('Epoch {}'.format(epoch + 1))
train(args.alg, model, trainloader, optimizer, criterion, device,
args.alpha, args.eps)
test(model, testloader, criterion, device, domain_fn)
# Trim
r = test(model, testtrainloader, criterion, device,
domain_fn, args.trim_num)
X_train = X_train[r]
y_train = y_train[r]
trainset = MyDataset(X_train, y_train)
trainloader = DataLoader(trainset, batch_size=args.batch_size, shuffle=True)
testtrainloader = DataLoader(trainset, batch_size=1024, shuffle=False)
mat = {
'X_train': X_train,
'y_train': y_train,
}
sio.savemat(args.save_file, mat)
return
# Training
model = build_model(input_dim)
model = model.to(device)
criterion = MyLoss(reduction='none')
optimizer = optim.ASGD(model.parameters(), lr=args.lr)
avg_acc = []
avg_loss = []
group_acc = []
group_loss = []
for epoch in range(args.epochs):
print('Epoch {}'.format(epoch + 1))
train(args.alg, model, trainloader, optimizer, criterion, device,
args.alpha, args.eps)
a, b, c, d = test(model, testloader, criterion, device, domain_fn)
avg_acc.append(a)
avg_loss.append(b)
group_acc.append(c)
group_loss.append(d)
# Save the results
if args.save_file is not None:
mat = {
'avg_acc': np.array(avg_acc),
'avg_loss': np.array(avg_loss),
'group_acc': np.array(group_acc),
'group_loss': np.array(group_loss),
}
sio.savemat(args.save_file, mat)
def test(model: Module, loader: DataLoader, criterion, device: str,
domain_fn, trim_num=None):
"""Test the avg and group acc of the model"""
model.eval()
total_correct = 0
total_loss = 0
total_num = 0
num_domains = len(domain_fn)
group_correct = np.zeros((num_domains,), dtype=np.int)
group_loss = np.zeros((num_domains,), dtype=np.float)
group_num = np.zeros((num_domains,), dtype=np.int)
l_rec = []
with torch.no_grad():
for _, (inputs, targets) in enumerate(loader):
inputs, targets = inputs.to(device), targets.to(device)
outputs = model(inputs).view(-1)
c = ((outputs > 0) & (targets == 1)) | ((outputs < 0) & (targets == 0))
correct = c.sum().item()
l = criterion(outputs, targets).view(-1)
if trim_num is not None:
l_rec.append(l.detach().cpu().numpy())
loss = l.sum().item()
total_correct += correct
total_loss += loss
total_num += len(inputs)
for i in range(num_domains):
g = domain_fn[i](inputs)
group_correct[i] += c[g].sum().item()
group_loss[i] += l[g].sum().item()
group_num[i] += g.sum().item()
print('Acc: {} ({} of {})'.format(total_correct / total_num, total_correct, total_num))
print('Avg Loss: {}'.format(total_loss / total_num))
for i in range(num_domains):
print('Group {}\tAcc: {} ({} of {})'.format(i, group_correct[i]/ group_num[i],
group_correct[i], group_num[i]))
print('Group {}\tAvg Loss: {}'.format(i, group_loss[i] / group_num[i]))
if trim_num is not None:
l_vec = np.concatenate(l_rec)
l = np.argsort(l_vec)[:-trim_num]
return l
return total_correct / total_num, total_loss / total_num, \
group_correct / group_num, group_loss / group_num
if __name__ == '__main__':
main()