-
Notifications
You must be signed in to change notification settings - Fork 2
/
experiment_run_soum.py
144 lines (118 loc) · 5.02 KB
/
experiment_run_soum.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
import sys
platform_os = sys.platform
if platform_os == "linux" or platform_os == "linux2":
sys.path.insert(0, '/dss/dsshome1/lxc04/ra93yex2/shapiq')
sys.path.insert(0, '/dss/dsshome1/lxc04/ra93yex2/shapiq/games')
sys.path.insert(0, '/dss/dsshome1/lxc04/ra93yex2/shapiq/approximators')
import random
import os
import time
import numpy as np
import pandas as pd
from approximators import SHAPIQEstimator, PermutationSampling, RegressionEstimator
from experiment import run_top_order_experiment
from games import ParameterizedSparseLinearModel
from utils_experiment import get_gt_values_for_game
import warnings
warnings.filterwarnings("ignore")
def run_soum_experiment(max_budget: int, soum_params: dict):
# CONSTANTS ------------------------------------------------------------------------------------
MAX_BUDGET = max_budget
BUDGET_STEPS = list(np.arange(0.05, 1.05, 0.05)) # step size of computation budgets
PAIRING = False
STRATIFICATION = False
SAMPLING_KERNEL = "ksh"
# Setup games ----------------------------------------------------------------------------------
# initialize games
game_list = []
for i in range(NUMBER_OF_RUNS):
game = ParameterizedSparseLinearModel(**soum_params)
game_list.append(game)
# get number of players
n = game_list[0].n
N = set(range(n))
# define folder name and save path
SAVE_FOLDER = os.path.join("results", '_'.join((data_folder, str(n))), interaction_index)
file_name = '_'.join((
f"n-{n}", f"runs-{NUMBER_OF_RUNS}", f"s0-{order}", f"top-order-{RUN_TOP_ORDER}",
f"pairing-{PAIRING}", f"stratification-{STRATIFICATION}", f"weights-{SAMPLING_KERNEL}"
))
if singleton_run:
file_name += f"_{time.time()}"
file_name += ".json"
SAVE_PATH = os.path.join(SAVE_FOLDER, file_name)
print("Loaded games.")
print("Number of games: ", len(game_list))
# Initialize estimators ------------------------------------------------------------------------
# SHAP-IQ estimator for all three indices
shapiq_estimator = SHAPIQEstimator(
N=N, order=order, interaction_type=interaction_index, top_order=RUN_TOP_ORDER)
# get baseline estimator
if interaction_index in ['SII', 'STI']:
# for SII and STI we use the permutation sampling estimator
baseline_estimator = PermutationSampling(
N=N, order=order, interaction_type=interaction_index, top_order=RUN_TOP_ORDER)
else:
# for SFI we use the regression estimator
baseline_estimator = RegressionEstimator(
N=N, max_order=order)
print("Initialized estimators.")
# Pre-compute the gt values --------------------------------------------------------------------
print("Precomputing gt values.")
all_gt_values = {}
for n, game in enumerate(game_list, start=1):
gt_values = get_gt_values_for_game(game=game, shapiq=shapiq_estimator, order=order)
all_gt_values[n] = gt_values
# Run experiments ------------------------------------------------------------------------------
print("Starting experiments.")
RESULTS = run_top_order_experiment(
top_order=RUN_TOP_ORDER,
game_list=game_list,
shapiq_estimator=shapiq_estimator,
baseline_estimator=baseline_estimator,
all_gt_values=all_gt_values,
order=order,
max_budget=MAX_BUDGET,
pairing=PAIRING,
stratification=STRATIFICATION,
sampling_kernel=SAMPLING_KERNEL,
budget_steps=BUDGET_STEPS,
save_path=SAVE_PATH,
save_folder=SAVE_FOLDER
)
# Save results ---------------------------------------------------------------------------------
print("Saving results.")
if not os.path.exists(SAVE_FOLDER):
os.makedirs(SAVE_FOLDER)
results_df = pd.DataFrame(RESULTS)
results_df.to_json(SAVE_PATH)
print("Done.")
if __name__ == "__main__":
warnings.filterwarnings("ignore")
run_parameters = sys.argv
has_no_param = not len(run_parameters) > 1
print("run parameters", run_parameters)
interaction_index = "FSI" if has_no_param else run_parameters[1]
order = 2 if has_no_param else int(run_parameters[2])
NUMBER_OF_RUNS = 1 if has_no_param else int(run_parameters[3])
RUN_TOP_ORDER = True if has_no_param else bool(run_parameters[4])
singleton_run = False if has_no_param else bool(run_parameters[5])
MAX_BUDGET = 2**14
data_folder = "SOUM"
soum_params = {
"n": 30,
"weighting_scheme": "uniform",
"n_interactions": 50,
"max_interaction_size": 30,
"min_interaction_size": 1,
"n_non_important_features": 0
}
if singleton_run:
for i in range(NUMBER_OF_RUNS):
seconds_sleep = random.randint(1, 7)
print(f"Sleeping for {seconds_sleep} seconds.")
time.sleep(seconds_sleep)
NUMBER_OF_RUNS = 1
run_soum_experiment(max_budget=MAX_BUDGET, soum_params=soum_params)
else:
run_soum_experiment(max_budget=MAX_BUDGET, soum_params=soum_params)