-
Notifications
You must be signed in to change notification settings - Fork 0
/
RandomFeatures_ParameterTuning.py
183 lines (156 loc) · 8.26 KB
/
RandomFeatures_ParameterTuning.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
import numpy as np
import pickle
import os
import time
import argparse
from src import RandomFeatures, Config, LinearFunctionApproximator
from src import IDBD, Adam, AutoStep, SGD, SIDBD
class Experiment:
def __init__(self, exp_arguments, results_path, tunable_parameter_values):
self.results_path = results_path
self.verbose = exp_arguments.verbose
self.tunable_parameter_values = tunable_parameter_values
self.stepsize_method = exp_arguments.stepsize_method
self.noisy = exp_arguments.noisy
self.config = Config()
""" Environment Setup """
self.config.num_true_features = exp_arguments.num_true_features
self.config.num_obs_features = exp_arguments.num_true_features
self.config.max_num_features = 200 # arbitrary since it's not used
self.training_data = exp_arguments.training_data_size
self.checkpoint = 10
assert self.training_data % self.checkpoint == 0
""" Experiment Setup """
self.sample_size = exp_arguments.sample_size
""" Stepsize adaptation settings"""
self.config.parameter_size = self.config.num_true_features
if self.stepsize_method == 'idbd':
# non-tunable parameters
self.config.init_beta = np.log(0.001)
self.parameter_name = 'meta_stepsize'
self.stepsize_method_class = IDBD
elif self.stepsize_method == 'sidbd':
self.config.init_beta = -np.log((1/0.001) - 1) # equivalent to starting with a stepsize of 0.001
self.parameter_name = 'meta_stepsize'
self.stepsize_method_class = SIDBD
elif self.stepsize_method == 'adam':
# non-tunable parameters
self.config.beta1 = 0.9
self.config.beta2 = 0.99
self.config.eps = 1e-08
self.parameter_name = 'initial_stepsize'
self.stepsize_method_class = Adam
elif self.stepsize_method == 'autostep':
# non-tunable parameters
self.config.tau = 10000.0
self.config.init_stepsize = 0.001
self.parameter_name = 'meta_stepsize'
self.stepsize_method_class = AutoStep
elif self.stepsize_method in ['sgd', 'rescaled_sgd']:
# non-tunable parameters
self.parameter_name = 'stepsize'
self.stepsize_method_class = SGD
self.config.rescale = (self.stepsize_method == 'rescaled_sgd')
else:
raise ValueError("Unrecognized stepsize adaptation method.")
def _print(self, astring):
if self.verbose:
print(astring)
def set_tunable_parameter_value(self, val):
if self.stepsize_method in ['idbd', 'sidbd']:
self.config.theta = val
elif self.stepsize_method == 'adam':
self.config.init_alpha = val
elif self.stepsize_method == 'autostep':
self.config.mu = val
elif self.stepsize_method in ['sgd', 'rescaled_sgd']:
self.config.alpha = val
else:
raise ValueError("Unrecognized stepsize adaptation method.")
def run(self):
results = {'parameter_name': self.parameter_name,
'sample_size': self.sample_size,
'parameter_values': self.tunable_parameter_values,
'avg_l2_norm_diff': np.zeros(len(self.tunable_parameter_values)),
'avg_mse': np.zeros(len(self.tunable_parameter_values))}
for j, pv in enumerate(self.tunable_parameter_values):
self._print("Parameter value: {0}".format(pv))
self.set_tunable_parameter_value(pv)
l2_norm_diff_per_run = np.zeros(self.sample_size)
avg_mse_per_run = np.zeros(self.sample_size)
for i in range(self.sample_size):
self._print("\tRun number: {0}".format(i+1))
env = RandomFeatures(self.config)
approximator = LinearFunctionApproximator(self.config)
stepsize_method = self.stepsize_method_class(self.config)
run_mse = np.zeros(self.training_data // self.checkpoint)
mse = 0
current_checkpoint = 0
for k in range(self.training_data):
target, observable_features, best_approximation = env.sample_observation(noisy=self.noisy)
prediction = approximator.get_prediction(observable_features)
error = target - prediction
_, _, new_weights = stepsize_method.update_weight_vector(error, observable_features,
approximator.get_weight_vector())
approximator.update_weight_vector(new_weights)
mse += np.square(error) / self.checkpoint
if (k + 1) % self.checkpoint == 0:
run_mse[current_checkpoint] += mse
mse *= 0
current_checkpoint += 1
theta_star = env.theta
approx_theta = approximator.get_weight_vector()
l2_norm_diff = np.sqrt(np.sum(np.square(theta_star - approx_theta)))
l2_norm_diff_per_run[i] += l2_norm_diff
avg_mse_per_run[i] += np.average(run_mse)
self._print("\t\tL2 Norm Difference: {0:.4f}".format(l2_norm_diff))
self._print("\t\tAverage MSE: {0:.4f}".format(np.average(run_mse)))
results['avg_l2_norm_diff'][j] += np.average(l2_norm_diff_per_run)
results['avg_mse'][j] += np.average(avg_mse_per_run)
self._print("Average L2 Norm Difference: {0:.4f}".format(np.average(l2_norm_diff_per_run)))
self._print("Average MSE: {0:.4f}".format(np.average(avg_mse_per_run)))
self.store_results(results)
def store_results(self, results):
file_path = os.path.join(self.results_path, 'parameter_tuning_results.p')
with open(file_path, mode='wb') as results_file:
pickle.dump(results, results_file)
print("Results successfully stored.")
def main():
""" Experiment Parameters """
parser = argparse.ArgumentParser()
parser.add_argument('-ss', '--sample_size', action='store', default=1, type=int)
parser.add_argument('-tds', '--training_data_size', action='store', default=100000, type=int)
parser.add_argument('-ntf', '--num_true_features', action='store', default=3, type=int)
parser.add_argument('-ssm', '--stepsize_method', action='store', default='sgd', type=str,
choices=['sgd', 'adam', 'idbd', 'autostep', 'rescaled_sgd', 'sidbd'])
parser.add_argument('-tpv', '--tunable_parameter_values', action='store', nargs='+', type=float, required=True)
parser.add_argument('--noisy', action='store_true', default=False)
parser.add_argument('-v', '--verbose', action='store_true')
exp_parameters = parser.parse_args()
task_name = 'random_features_task'
if exp_parameters.noisy:
task_name = 'noisy_' + task_name
results_path = os.path.join(os.getcwd(), 'results', 'parameter_tuning', task_name,
"num_true_features_" + str(exp_parameters.num_true_features),
exp_parameters.stepsize_method)
os.makedirs(results_path, exist_ok=True)
init_time = time.time()
exp = Experiment(exp_parameters, results_path, exp_parameters.tunable_parameter_values)
exp.run()
finish_time = time.time()
elapsed_time = (finish_time - init_time) / 60
print("Running time in minutes: {0:.4f}".format(elapsed_time))
if __name__ == '__main__':
main()
# SGD parameter values: stepsize in {0.1, 0.05, 0.01, 0.005, 0.001, 0.0005, 0.0001}
# - 0.005 had the lowest MSVE
# IDBD parameter values: theta in {1.0, 0.5, 0.1, 0.05, 0.01, 0.005, 0.001}
# - 0.5 had the lowest MSVE
# Adam parameter values: initial stepsize in {0.1, 0.05, 0.01, 0.005, 0.001, 0.0005, 0.0001}
# - 0.005 had the lowest MSVE
# AutoStep parameter values: mu in {0.5, 0.1, 0.05, 0.01, 0.005, 0.001, 0.0005}
# - 0.05 had the lowest MSVE
# Rescaled SGD parameter values: stepsize in {0.5, 0.1, 0.05, 0.01, 0.005, 0.001}
# - 0.01 had the lowest MSVE
# SIDBD parameter values: theta in {5.0 1.0 0.5 0.1 0.05 0.01 0.005}
# - 0.5 had the lowest MSVE