/
isolated_stimuli_experiments.py
executable file
·63 lines (45 loc) · 2.43 KB
/
isolated_stimuli_experiments.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
# This contains the experimental code for the Cicling 2019 submission for the isolated stimuli.
from encoding_pipeline.isolated_pipeline import SingleInstancePipeline
from read_dataset.read_words_data import WordsReader
from read_dataset.read_stories_data import StoryDataReader
from language_models.elmo_encoder import ElmoEncoder
from language_models.random_encoder import RandomEncoder
from mapping_models.ridge_regression_mapper import RegressionMapper
import logging
# Make sure to get the Mitchell data and the Kaplan data and adjust the paths
user_dir = "USERDIR/"
mitchell_dir = user_dir + "Corpora/mitchell/"
kaplan_dir = user_dir + "Corpora/Kaplan_data/"
save_dir = user_dir + "fmriExperiments/isolated/"
# Use this to set up Mitchell and Kaplan experiments
if __name__ == '__main__':
logging.basicConfig(level=logging.INFO)
# Set the readers
mitchell_reader = WordsReader(data_dir=mitchell_dir)
kaplan_reader = StoryDataReader(data_dir=kaplan_dir)
# Set the mapping model
mapper = RegressionMapper(alpha=10.0)
# Set the language models
stimuli_encoder = ElmoEncoder(save_dir)
random_encoder = RandomEncoder(save_dir)
# Try different language models
for encoder in [ stimuli_encoder, random_encoder]:
# Set up the pipelines
mitchell_pipeline_name = "Words" + encoder.__class__.__name__
mitchell_pipeline = SingleInstancePipeline(mitchell_reader, encoder, mapper, mitchell_pipeline_name, save_dir=save_dir)
stories_pipeline_name = "Stories" + encoder.__class__.__name__
stories_pipeline = SingleInstancePipeline(kaplan_reader, encoder, mapper, stories_pipeline_name, save_dir=save_dir)
# Set voxel selection
voxel_selections = ["none", "on_train_ev"]
for v_selection in voxel_selections:
mitchell_pipeline.voxel_selection = v_selection
# Run Words experiments
#mitchell_pipeline.run_voxelwiseevaluation_cv("crossvalidation_" + v_selection)
#mitchell_pipeline.run_pairwise_procedure( "pairwise_" +v_selection )
# Run Stories experiments
stories_pipeline.voxel_selection = v_selection
stories_pipeline.run_voxelwiseevaluation_cv("crossvalidation_" + v_selection)
stories_pipeline.run_pairwise_procedure("pairwise_" + v_selection)
# Run representational similarity analysis
mitchell_pipeline.runRSA("rsa")
stories_pipeline.runRSA("rsa")