-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathsvm_analysis.py
155 lines (124 loc) · 4.92 KB
/
svm_analysis.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
from imports import *
from utils import to_cpu
from sklearn.svm import SVC
from itertools import product
from parameters import par, update_parameters
from stimulus import Stimulus
from model import Model
# List save files to analyze
savefns = [
'verify_500neuron_var120delay_v0_data_iter001900',
'verify_500neuron_var120delay_v1_data_iter001900',
'verify_500neuron_var120delay_v5_data_iter001900',
'verify_500neuron_var120delay_v6_data_iter001900',
'verify_500neuron_var120delay_v10_data_iter001900',
'verify_500neuron_var120delay_v11_data_iter001900',
'verify_500neuron_var120delay_v15_data_iter001900',
'verify_500neuron_var120delay_v16_data_iter001900',
]
# Make plot
fig, ax = plt.subplots(1, 2, figsize=(14,6))
# Iterate over provided save files
for num_fn, fn in enumerate(savefns):
print('Processing file {} of {}.'.format(num_fn+1, len(savefns)))
# Load data
data = pickle.load(open('./savedir/{}.pkl'.format(fn), 'rb'))
# Update parameters with current weights
update_parameters(data['par'], verbose=False)
update_parameters({'batch_size':128}, verbose=False)
update_parameters({k+'_init':v for k,v in data['weights'].items()}, \
verbose=False, update_deps=False)
end_dead_time = par['dead_time']//par['dt']
end_fix_time = end_dead_time + par['fix_time']//par['dt']
end_sample_time = end_fix_time + par['sample_time']//par['dt']
end_delay_time = end_sample_time + par['delay_time']//par['dt']
# Make a new model and stimulus (which use the loaded parameters)
print('\nLoading and running model.')
model = Model()
stim = Stimulus()
runs = 8
c_all = []
d_all = []
v_all = []
s_all = []
# Run a couple batches to generate sufficient data points
for i in range(runs):
print('R:{:>2}'.format(i), end='\r')
trial_info = stim.make_batch(var_delay=False)
model.run_model(trial_info, testing=True)
c_all.append(trial_info['sample_cat'])
d_all.append(trial_info['sample_dir'])
v_all.append(to_cpu(model.v))
s_all.append(to_cpu(model.s))
del model
del stim
batch_size = runs*par['batch_size']
c = np.concatenate(c_all, axis=0)
d = np.concatenate(d_all, axis=0)
v = np.concatenate(v_all, axis=1)
s = np.concatenate(s_all, axis=1)
print('Model run complete.')
print('Performing SVM decoding on {} trials.\n'.format(batch_size))
# Initialize linear classifiers
args = {'kernel':'linear', 'decision_function_shape':'ovr', \
'shrinking':False, 'tol':1e-3}
lin_clf_cv = SVC(**args)
lin_clf_cs = SVC(**args)
lin_clf_dv = SVC(**args)
lin_clf_ds = SVC(**args)
c_score_v = np.zeros([par['num_time_steps']])
c_score_s = np.zeros([par['num_time_steps']])
d_score_v = np.zeros([par['num_time_steps']])
d_score_s = np.zeros([par['num_time_steps']])
# Choose training and testing indices
train_pct = 0.75
num_train_inds = int(batch_size * train_pct)
shuffled = np.random.permutation(batch_size)
train_inds = shuffled[:num_train_inds]
test_inds = shuffled[num_train_inds:]
# Fit the classifiers for each time step,
# and judge their accuracy at predicting
# category encoding.
for t in range(end_dead_time, par['num_time_steps']):
print('T:{:>4}'.format(t), end='\r')
lin_clf_cv.fit(v[t,train_inds,:], c[train_inds])
lin_clf_cs.fit(s[t,train_inds,:], c[train_inds])
lin_clf_dv.fit(v[t,train_inds,:], d[train_inds])
lin_clf_ds.fit(s[t,train_inds,:], d[train_inds])
dec_cv = lin_clf_cv.predict(v[t,test_inds,:])
dec_cs = lin_clf_cs.predict(s[t,test_inds,:])
dec_dv = lin_clf_dv.predict(v[t,test_inds,:])
dec_ds = lin_clf_ds.predict(s[t,test_inds,:])
c_score_v[t] = np.mean(c[test_inds]==dec_cv)
c_score_s[t] = np.mean(c[test_inds]==dec_cs)
d_score_v[t] = np.mean(d[test_inds]==dec_dv)
d_score_s[t] = np.mean(d[test_inds]==dec_ds)
# Plot classification scores
if num_fn == 0:
ax[0].plot(c_score_v, c=[241/255, 153/255, 1/255], label='Voltage')
ax[0].plot(c_score_s, c=[58/255, 100/255, 65/255], label='Syn. Eff.')
else:
ax[0].plot(c_score_v, c=[241/255, 153/255, 1/255])
ax[0].plot(c_score_s, c=[58/255, 100/255, 65/255])
ax[1].plot(d_score_v, c=[241/255, 153/255, 1/255])
ax[1].plot(d_score_s, c=[58/255, 100/255, 65/255])
print('Processing of file {} complete.'.format(num_fn+1))
# Decorate and save plots
ax[0].axhline(0.5, c='k', ls='--')
ax[1].axhline(0.125, c='k', ls='--')
ax[0].axvline(trial_info['timings'][0,0], c='k', ls='--')
ax[0].axvline(trial_info['timings'][1,0], c='k', ls='--')
ax[1].axvline(trial_info['timings'][0,0], c='k', ls='--')
ax[1].axvline(trial_info['timings'][1,0], c='k', ls='--')
ax[0].set_title('Sample Category')
ax[1].set_title('Sample Direction')
for i in range(2):
ax[i].set_xlabel('Time')
ax[i].set_ylabel('Decoding Accuracy')
ax[i].set_xlim(0,par['num_time_steps']-1)
ax[i].set_yticks([0., 0.25, 0.5, 0.75, 1.])
ax[i].grid()
ax[0].legend(loc='lower right')
fig.suptitle('SVM Task Decoding from Neural Population')
plt.savefig('./analysis/svm_decoding.png', bbox_inches='tight')
plt.savefig('./analysis/svm_decoding.pdf', bbox_inches='tight')