-
Notifications
You must be signed in to change notification settings - Fork 11
/
Copy pathreproducibility_additional_experiments_all.sh
373 lines (310 loc) · 45.1 KB
/
reproducibility_additional_experiments_all.sh
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
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
echo "***WARNING: we heavily discourage running this script, as in our estimates replicating all of our additional experiments might take more than 2 months of uninterrupted computation time to complete. We suggest reproducing only a selection of our experiment by running the reproducibility_run_additional_experiments_selection.sh script instead.***" && \
##################################################
### NECESSARY ACCEPTANCE THRESHOLD EXPERIMENTS ###
##################################################
# Kelpie Necessary ComplEx FB15k; Relevance Threshold 1
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_fb15k_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx FB15k; Relevance Threshold 10
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_fb15k_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18; Relevance Threshold 1
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_wn18_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18; Relevance Threshold 10
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_wn18_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx FB15k-237; Relevance Threshold 1
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_fb15k237_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx FB15k-237; Relevance Threshold 10
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_fb15k237_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18RR; Relevance Threshold 1
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_wn18rr_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18RR; Relevance Threshold 10
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_wn18rr_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx YAGO3-10; Relevance Threshold 1
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_yago310_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx YAGO3-10; Relevance Threshold 10
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_complex_yago310_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE FB15k; Relevance Threshold 1
python3 scripts/conve/explain.py --dataset FB15k --max_epochs 1000 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k.pt --facts_to_explain_path input_facts/conve_fb15k_random_nec.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/conve/verify_explanations.py --dataset FB15k --max_epochs 1000 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_fb15k_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE FB15k; Relevance Threshold 10
python3 scripts/conve/explain.py --dataset FB15k --max_epochs 1000 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k.pt --facts_to_explain_path input_facts/conve_fb15k_random_nec.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/conve/verify_explanations.py --dataset FB15k --max_epochs 1000 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_fb15k_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE WN18; Relevance Threshold 1
python3 scripts/conve/explain.py --dataset WN18 --max_epochs 150 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18.pt --facts_to_explain_path input_facts/conve_wn18_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/conve/verify_explanations.py --dataset WN18 --max_epochs 150 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_wn18_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE WN18; Relevance Threshold 10
python3 scripts/conve/explain.py --dataset WN18 --max_epochs 150 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18.pt --facts_to_explain_path input_facts/conve_wn18_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/conve/verify_explanations.py --dataset WN18 --max_epochs 150 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_wn18_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE FB15k-237; Relevance Threshold 1
python3 scripts/conve/explain.py --dataset FB15k-237 --max_epochs 60 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k-237.pt --facts_to_explain_path input_facts/conve_fb15k237_random.csv --mode necessary --relevance_threshold 1 && \
python scripts/conve/verify_explanations.py --dataset FB15k-237 --max_epochs 60 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k-237.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_fb15k237_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE FB15k-237; Relevance Threshold 10
python3 scripts/conve/explain.py --dataset FB15k-237 --max_epochs 60 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k-237.pt --facts_to_explain_path input_facts/conve_fb15k237_random.csv --mode necessary --relevance_threshold 10 && \
python scripts/conve/verify_explanations.py --dataset FB15k-237 --max_epochs 60 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_FB15k-237.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_fb15k237_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE WN18RR; Relevance Threshold 1
python3 scripts/conve/explain.py --dataset WN18RR --max_epochs 90 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18RR.pt --facts_to_explain_path input_facts/conve_wn18rr_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/conve/verify_explanations.py --dataset WN18RR --max_epochs 90 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18RR.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_wn18rr_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE WN18RR; Relevance Threshold 10
python3 scripts/conve/explain.py --dataset WN18RR --max_epochs 90 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18RR.pt --facts_to_explain_path input_facts/conve_wn18rr_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/conve/verify_explanations.py --dataset WN18RR --max_epochs 90 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_WN18RR.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_wn18rr_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE YAGO3-10; Relevance Threshold 1
python3 scripts/conve/explain.py --dataset YAGO3-10 --max_epochs 500 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_YAGO3-10.pt --facts_to_explain_path input_facts/conve_yago_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/conve/verify_explanations.py --dataset YAGO3-10 --max_epochs 500 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_YAGO3-10.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_yago310_1.csv && \
rm output_*.csv && \
# Kelpie Necessary ConvE YAGO3-10; Relevance Threshold 10
python3 scripts/conve/explain.py --dataset YAGO3-10 --max_epochs 500 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_YAGO3-10.pt --facts_to_explain_path input_facts/conve_yago_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/conve/verify_explanations.py --dataset YAGO3-10 --max_epochs 500 --batch_size 128 --learning_rate 0.003 --dimension 200 --input_dropout 0.2 --hidden_dropout 0.3 --feature_map_dropout 0.2 --decay_rate 0.995 --model_path stored_models/ConvE_YAGO3-10.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_conve_yago310_10.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE FB15k; Relevance Threshold 1
python3 scripts/transe/explain.py --dataset FB15k --max_epochs 200 --batch_size 2048 --learning_rate 0.01 --dimension 200 --negative_samples_ratio 5 --regularizer_weight 2.0 --margin 2 --model_path stored_models/TransE_FB15k.pt --facts_to_explain_path input_facts/transe_fb15k_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/transe/verify_explanations.py --dataset FB15k --max_epochs 200 --batch_size 2048 --learning_rate 0.00003 --dimension 200 --negative_samples_ratio 5 --regularizer_weight 2.0 --margin 2 --model_path stored_models/TransE_FB15k.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_fb15k_1.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE FB15k; Relevance Threshold 10
python3 scripts/transe/explain.py --dataset FB15k --max_epochs 200 --batch_size 2048 --learning_rate 0.01 --dimension 200 --negative_samples_ratio 5 --regularizer_weight 2.0 --margin 2 --model_path stored_models/TransE_FB15k.pt --facts_to_explain_path input_facts/transe_fb15k_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/transe/verify_explanations.py --dataset FB15k --max_epochs 200 --batch_size 2048 --learning_rate 0.00003 --dimension 200 --negative_samples_ratio 5 --regularizer_weight 2.0 --margin 2 --model_path stored_models/TransE_FB15k.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_fb15k_10.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE WN18; Relevance Threshold 1
python3 scripts/transe/explain --dataset WN18 --max_epochs 250 --batch_size 2048 --learning_rate 0.01 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 0 --margin 2 --facts_to_explain_path input_facts/transe_wn18_random.csv --model_path stored_models/TransE_WN18.pt --mode necessary --relevance_threshold 1 && \
python3 end_to_end_testing/transe/verify_explanations.py --dataset WN18 --max_epochs 250 --batch_size 2048 --learning_rate 0.0002 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 0 --margin 2 --model_path stored_models/TransE_WN18.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_wn18_1.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE WN18; Relevance Threshold 10
python3 scripts/transe/explain --dataset WN18 --max_epochs 250 --batch_size 2048 --learning_rate 0.01 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 0 --margin 2 --facts_to_explain_path input_facts/transe_wn18_random.csv --model_path stored_models/TransE_WN18.pt --mode necessary --relevance_threshold 10 && \
python3 end_to_end_testing/transe/verify_explanations.py --dataset WN18 --max_epochs 250 --batch_size 2048 --learning_rate 0.0002 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 0 --margin 2 --model_path stored_models/TransE_WN18.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_wn18_10.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE FB15k-237; Relevance Threshold 1
python3 scripts/transe/explain.py --dataset FB15k-237 --max_epochs 100 --batch_size 2048 --learning_rate 0.01 --dimension 50 --negative_samples_ratio 15 --regularizer_weight 1.0 --margin 5 --facts_to_explain_path input_facts/transe_fb15k237_random.csv --model_path stored_models/TransE_FB15k-237.pt --mode necessary --relevance_threshold 1 && \
python3 scripts/transe/verify_explanations.py --dataset FB15k-237 --max_epochs 100 --batch_size 2048 --learning_rate 0.0004 --dimension 50 --negative_samples_ratio 15 --regularizer_weight 1.0 --margin 5 --model_path stored_models/TransE_FB15k-237.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_fb15k237_1.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE FB15k-237; Relevance Threshold 10
python3 scripts/transe/explain.py --dataset FB15k-237 --max_epochs 100 --batch_size 2048 --learning_rate 0.01 --dimension 50 --negative_samples_ratio 15 --regularizer_weight 1.0 --margin 5 --facts_to_explain_path input_facts/transe_fb15k237_random.csv --model_path stored_models/TransE_FB15k-237.pt --mode necessary --relevance_threshold 10 && \
python3 scripts/transe/verify_explanations.py --dataset FB15k-237 --max_epochs 100 --batch_size 2048 --learning_rate 0.0004 --dimension 50 --negative_samples_ratio 15 --regularizer_weight 1.0 --margin 5 --model_path stored_models/TransE_FB15k-237.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_fb15k237_10.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE WN18RR; Relevance Threshold 1
python3 scripts/transe/explain.py --dataset WN18RR --max_epochs 200 --batch_size 2048 --learning_rate 0.01 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 50.0 --margin 2 --model_path stored_models/TransE_WN18RR.pt --facts_to_explain_path input_facts/transe_wn18rr_random.csv --mode necessary --relevance_threshold 1 && \
python scripts/transe/verify_explanations.py --dataset WN18RR --max_epochs 200 --batch_size 2048 --learning_rate 0.0001 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 50.0 --margin 2 --model_path stored_models/TransE_WN18RR.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_wn18rr_1.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE WN18RR; Relevance Threshold 10
python3 scripts/transe/explain.py --dataset WN18RR --max_epochs 200 --batch_size 2048 --learning_rate 0.01 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 50.0 --margin 2 --model_path stored_models/TransE_WN18RR.pt --facts_to_explain_path input_facts/transe_wn18rr_random.csv --mode necessary --relevance_threshold 10 && \
python scripts/transe/verify_explanations.py --dataset WN18RR --max_epochs 200 --batch_size 2048 --learning_rate 0.0001 --dimension 50 --negative_samples_ratio 5 --regularizer_weight 50.0 --margin 2 --model_path stored_models/TransE_WN18RR.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_wn18rr_10.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE YAGO3-10; Relevance Threshold 1
python3 scripts/transe/explain.py --dataset YAGO3-10 --max_epochs 100 --batch_size 2048 --learning_rate 0.01 --dimension 200 --negative_samples_ratio 10 --regularizer_weight 50 --margin 5 --model_path stored_models/TransE_YAGO3-10.pt --facts_to_explain_path input_facts/transe_yago_random.csv --mode necessary --relevance_threshold 1 && \
python3 scripts/transe/verify_explanations.py --dataset YAGO3-10 --max_epochs 100 --batch_size 2048 --learning_rate 0.0001 --dimension 200 --negative_samples_ratio 10 --regularizer_weigh 50 --margin 5 --model_path stored_models/TransE_YAGO3-10.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_yago310_1.csv && \
rm output_*.csv && \
# Kelpie Necessary TransE YAGO3-10; Relevance Threshold 10
python3 scripts/transe/explain.py --dataset YAGO3-10 --max_epochs 100 --batch_size 2048 --learning_rate 0.01 --dimension 200 --negative_samples_ratio 10 --regularizer_weight 50 --margin 5 --model_path stored_models/TransE_YAGO3-10.pt --facts_to_explain_path input_facts/transe_yago_random.csv --mode necessary --relevance_threshold 10 && \
python3 scripts/transe/verify_explanations.py --dataset YAGO3-10 --max_epochs 100 --batch_size 2048 --learning_rate 0.0001 --dimension 200 --negative_samples_ratio 10 --regularizer_weigh 50 --margin 5 --model_path stored_models/TransE_YAGO3-10.pt --mode necessary && \
mv output_end_to_end.csv scripts/experiments/necessary_xsi_threshold/kelpie_transe_yago310_10.csv && \
rm output_*.csv && \
#######################################
### PREFILTER THRESHOLD EXPERIMENTS ###
#######################################
# Kelpie Necessary ComplEx FB15k: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode necessary --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_fb15k_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx FB15k: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode necessary --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_fb15k_30.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx FB15k: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode sufficient --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_fb15k_10.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx FB15k: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode sufficient --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_fb15k_30.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode necessary --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_wn18_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode necessary --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_wn18_30.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx WN18: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode sufficient --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_wn18_10.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx WN18: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode sufficient --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_wn18_30.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx FB15k-237: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode necessary --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_fb15k237_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx FB15k-237: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode necessary --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_fb15k237_30.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx FB15k-237: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode sufficient --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_fb15k237_10.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx FB15k-237: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode sufficient --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_fb15k237_30.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18RR: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode necessary --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_wn18rr_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18RR: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode necessary --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_wn18rr_30.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx WN18RR: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode sufficient --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_wn18rr_10.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx WN18RR: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode sufficient --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_wn18rr_30.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx YAGO3-10: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode necessary --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_yago310_10.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx YAGO3-10: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode necessary --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_necessary_complex_yago310_30.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx YAGO3-10: Prefilter Threshold 10
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode sufficient --prefilter_threshold 10 && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_yago310_10.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx YAGO3-10: Prefilter Threshold 30
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode sufficient --prefilter_threshold 30 && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_threshold/kelpie_sufficient_complex_yago310_30.csv && \
rm output_*.csv && \
##################################
### PREFILTER TYPE EXPERIMENTS ###
##################################
# Kelpie Necessary ComplEx FB15k: Typebased Prefilter
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode necessary --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_necessary_complex_fb15k_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx FB15k: Typebased Prefilter
python3 scripts/complex/explain.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --facts_to_explain_path input_facts/complex_fb15k_random.csv --mode sufficient --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset FB15k --model_path stored_models/ComplEx_FB15k.pt --optimizer Adagrad --dimension 2000 --batch_size 100 --max_epochs 200 --learning_rate 0.01 --reg 2.5e-3 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_sufficient_complex_fb15k_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18: Typebased Prefilter
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode necessary --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_necessary_complex_wn18_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx WN18: Typebased Prefilter
python3 scripts/complex/explain.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_wn18_random.csv --mode sufficient --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset WN18 --model_path stored_models/ComplEx_WN18.pt --optimizer Adagrad --dimension 500 --batch_size 1000 --max_epochs 20 --learning_rate 0.1 --reg 5e-2 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_sufficient_complex_wn18_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx FB15k-237: Typebased Prefilter
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode necessary --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_necessary_complex_fb15k237_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx FB15k-237: Typebased Prefilter
python3 scripts/complex/explain.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --facts_to_explain_path input_facts/complex_fb15k237_random.csv --mode sufficient --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset FB15k-237 --model_path stored_models/ComplEx_FB15k-237.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 100 --learning_rate 0.1 --reg 5e-2 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_sufficient_complex_fb15k237_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx WN18RR: Typebased Prefilter
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode necessary --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_necessary_complex_wn18rr_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx WN18RR: Typebased Prefilter
python3 scripts/complex/explain.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --facts_to_explain_path input_facts/complex_wn18rr_random.csv --mode sufficient --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset WN18RR --model_path stored_models/ComplEx_WN18RR.pt --optimizer Adagrad --dimension 500 --batch_size 100 --max_epochs 100 --learning_rate 0.1 --reg 1e-1 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_sufficient_complex_wn18rr_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Necessary ComplEx YAGO3-10: Typebased Prefilter
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode necessary --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode necessary && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_necessary_complex_yago310_typebased_prefilter.csv && \
rm output_*.csv && \
# Kelpie Sufficient ComplEx YAGO3-10: Typebased Prefilter
python3 scripts/complex/explain.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --facts_to_explain_path input_facts/complex_yago_random.csv --mode sufficient --prefilter type_based && \
python3 scripts/complex/verify_explanations.py --dataset YAGO3-10 --model_path stored_models/ComplEx_YAGO3-10.pt --optimizer Adagrad --dimension 1000 --batch_size 1000 --max_epochs 50 --learning_rate 0.1 --reg 5e-3 --mode sufficient && \
mv output_end_to_end.csv scripts/experiments/prefilter_type/kelpie_sufficient_complex_yago310_typebased_prefilter.csv && \
rm output_*.csv