-
Notifications
You must be signed in to change notification settings - Fork 29
/
test_botvinick.py
293 lines (251 loc) · 17 KB
/
test_botvinick.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
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
import numpy as np
import pytest
import psyneulink as pnl
# This script implements Figure 1 of Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen, J. D. (2001).
# Conflict monitoring and cognitive control. Psychological Review, 108, 624–652.
# http://dx.doi.org/10.1037/0033-295X.108.3.624
# Figure 1 plots the ENERGY computed by a conflict mechanism. It is highest for incongruent trials,
# and similar for congruent and neutral trials.
# Noise is turned of and for each condition we ran one trial only. A response threshold was not defined. Responses were
# made at the marked * signs in the figure.
# Note that this script implements a slightly different Figure than in the original Figure in the paper.
# However, this implementation is identical with a plot we created with an old MATLAB code which was used for the
# conflict monitoring simulations.
import psyneulink.core.components.functions.objectivefunctions
import psyneulink.core.components.functions.transferfunctions
@pytest.mark.model
@pytest.mark.benchmark
@pytest.mark.parametrize("reps", [1,
pytest.param(10, marks=pytest.mark.stress),
pytest.param(100, marks=pytest.mark.stress)])
@pytest.mark.parametrize("mode", ['Python',
pytest.param('LLVM', marks=pytest.mark.llvm),
pytest.param('LLVMExec', marks=pytest.mark.llvm),
pytest.param('LLVMRun', marks=pytest.mark.llvm),
pytest.param('PTXExec', marks=[pytest.mark.llvm, pytest.mark.cuda]),
pytest.param('PTXRun', marks=[pytest.mark.llvm, pytest.mark.cuda])])
def test_botvinick_model(benchmark, mode, reps):
benchmark.group = "Botvinick (scale " + str(reps / 100) + ")"
# SET UP MECHANISMS ----------------------------------------------------------------------------------------------------
# Linear input layer
# colors: ('red', 'green'), words: ('RED','GREEN')
colors_input_layer = pnl.TransferMechanism(size=3,
function=psyneulink.core.components.Linear,
name='COLORS_INPUT')
words_input_layer = pnl.TransferMechanism(size=3,
function=psyneulink.core.components.Linear,
name='WORDS_INPUT')
task_input_layer = pnl.TransferMechanism(size=2,
function=psyneulink.core.components.Linear,
name='TASK_INPUT')
# Task layer, tasks: ('name the color', 'read the word')
task_layer = pnl.RecurrentTransferMechanism(size=2,
function=psyneulink.core.components.Logistic,
hetero=-2,
integrator_mode=True,
integration_rate=0.01,
name='TASK_LAYER')
# Hidden layer
# colors: ('red','green', 'neutral') words: ('RED','GREEN', 'NEUTRAL')
colors_hidden_layer = pnl.RecurrentTransferMechanism(size=3,
function=psyneulink.core.components.Logistic(x_0=4.0), # bias 4.0 is -4.0 in the paper see Docs for description
integrator_mode=True,
hetero=-2,
integration_rate=0.01, # cohen-huston text says 0.01
name='COLORS_HIDDEN')
words_hidden_layer = pnl.RecurrentTransferMechanism(size=3,
function=psyneulink.core.components.Logistic(x_0=4.0),
integrator_mode=True,
hetero=-2,
integration_rate=0.01,
name='WORDS_HIDDEN')
# Response layer, responses: ('red', 'green')
response_layer = pnl.RecurrentTransferMechanism(size=2,
function=psyneulink.core.components.Logistic,
hetero=-2.0,
integrator_mode=True,
integration_rate=0.01,
output_ports = [pnl.RESULT,
{pnl.NAME: 'DECISION_ENERGY',
pnl.VARIABLE: (pnl.OWNER_VALUE,0),
pnl.FUNCTION: psyneulink.core.components.Stability(
default_variable = np.array([0.0, 0.0]),
metric = pnl.ENERGY,
matrix = np.array([[0.0, -4.0],
[-4.0, 0.0]]))}],
name='RESPONSE', )
# Mapping projections---------------------------------------------------------------------------------------------------
color_input_weights = pnl.MappingProjection(matrix=np.array([[1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0]]))
word_input_weights = pnl.MappingProjection(matrix=np.array([[1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0]]))
task_input_weights = pnl.MappingProjection(matrix=np.array([[1.0, 0.0],
[0.0, 1.0]]))
color_task_weights = pnl.MappingProjection(matrix=np.array([[4.0, 0.0],
[4.0, 0.0],
[4.0, 0.0]]))
task_color_weights = pnl.MappingProjection(matrix=np.array([[4.0, 4.0, 4.0],
[0.0, 0.0, 0.0]]))
response_color_weights = pnl.MappingProjection(matrix=np.array([[1.5, 0.0, 0.0],
[0.0, 1.5, 0.0]]))
response_word_weights = pnl.MappingProjection(matrix=np.array([[2.5, 0.0, 0.0],
[0.0, 2.5, 0.0]]))
color_response_weights = pnl.MappingProjection(matrix=np.array([[1.5, 0.0],
[0.0, 1.5],
[0.0, 0.0]]))
word_response_weights = pnl.MappingProjection(matrix=np.array([[2.5, 0.0],
[0.0, 2.5],
[0.0, 0.0]]))
word_task_weights = pnl.MappingProjection(matrix=np.array([[0.0, 4.0],
[0.0, 4.0],
[0.0, 4.0]]))
task_word_weights = pnl.MappingProjection(matrix=np.array([[0.0, 0.0, 0.0],
[4.0, 4.0, 4.0]]))
# CREATE Composition
comp = pnl.Composition()
# Add mechanisms
comp.add_node(colors_input_layer)
comp.add_node(colors_hidden_layer)
comp.add_node(words_input_layer)
comp.add_node(words_hidden_layer)
comp.add_node(task_input_layer)
comp.add_node(task_layer)
comp.add_node(response_layer)
# Add projections
comp.add_projection(task_input_weights, task_input_layer, task_layer)
# Color process
comp.add_projection(color_input_weights, colors_input_layer, colors_hidden_layer)
comp.add_projection(color_response_weights, colors_hidden_layer, response_layer)
comp.add_projection(response_color_weights, response_layer, colors_hidden_layer)
# Word process
comp.add_projection(word_input_weights, words_input_layer, words_hidden_layer)
comp.add_projection(word_response_weights, words_hidden_layer, response_layer)
comp.add_projection(response_word_weights, response_layer, words_hidden_layer)
# Color task process
comp.add_projection(task_color_weights, task_layer, colors_hidden_layer)
comp.add_projection(color_task_weights, colors_hidden_layer, task_layer)
# Word task process
comp.add_projection(task_word_weights, task_layer, words_hidden_layer)
comp.add_projection(word_task_weights, words_hidden_layer, task_layer)
def trial_dict(red_color, green_color, neutral_color, red_word, green_word, neutral_word, CN, WR):
trialdict = {
colors_input_layer: [red_color, green_color, neutral_color],
words_input_layer: [red_word, green_word, neutral_word],
task_input_layer: [CN, WR]
}
return trialdict
# Define initialization trials separately
CN_trial_initialize_input = trial_dict(0, 0, 0, 0, 0, 0, 1, 0) #red_color, green color, red_word, green word, CN, WR
CN_incongruent_trial_input = trial_dict(1, 0, 0, 0, 1, 0, 1, 0) #red_color, green color, red_word, green word, CN, WR
CN_congruent_trial_input = trial_dict(1, 0, 0, 1, 0, 0, 1, 0) #red_color, green color, red_word, green word, CN, WR
CN_control_trial_input = trial_dict(1, 0, 0, 0, 0, 1, 1, 0) #red_color, green color, red_word, green word, CN, WR
Stimulus = [[CN_trial_initialize_input, CN_congruent_trial_input],
[CN_trial_initialize_input, CN_incongruent_trial_input],
[CN_trial_initialize_input, CN_control_trial_input]]
# should be 500 and 1000
ntrials0 = 5 * reps
ntrials = 10 * reps
def run(bin_execute):
results = []
for i, stim in enumerate(Stimulus):
# RUN the COMPOSITION to initialize --------------------------------
exec_id = "exec_" + str(i)
comp.run(inputs=stim[0], num_trials=ntrials0, bin_execute=bin_execute, context=exec_id)
comp.run(inputs=stim[1], num_trials=ntrials, bin_execute=bin_execute, context=exec_id)
# Comp results include concatenation of both the above runs
results.append(comp.results)
return results
res = run(mode)
# the corresponding output port indices in composition results
# these were 0 and 1 in the prior version of the test
response_results_index = 3
response_decision_energy_index = 4
if reps == 1:
res2d = [x[response_results_index] for r in res for x in r]
# NOTE: The formatting below provides visual split between
# initialization and runs. Please do not change it.
assert np.allclose(res2d, [[0.497679878752004, 0.497679878752004],
[0.4954000154631831, 0.4954000154631831],
[0.4931599760310996, 0.4931599760310996],
[0.4909593232354856, 0.4909593232354856],
[0.4887976172454234, 0.4887976172454234],
[0.4866744160981826, 0.4866744160981826],
[0.4845913708928323, 0.4845892761509014],
[0.4825481248817388, 0.4825417411478152],
[0.4805443207349158, 0.4805313535747981],
[0.4785796008355799, 0.4785576550393446],
[0.4766536075535322, 0.4766201866281457],
[0.4747659834976122, 0.4747184892433102],
[0.4729163717484405, 0.4728521039182508],
[0.471104416072617, 0.4710205721142167],
[0.4693297611195044, 0.4692234359984253],
[0.497679878752004, 0.497679878752004],
[0.4954000154631831, 0.4954000154631831],
[0.4931599760310996, 0.4931599760310996],
[0.4909593232354856, 0.4909593232354856],
[0.4887976172454234, 0.4887976172454234],
[0.4866744160981826, 0.4866744160981826],
[0.4845900488307297, 0.484590598212878],
[0.4825440921072959, 0.4825457739209331],
[0.4805361215634117, 0.4805395527400902],
[0.478565712169312, 0.4785715436856292],
[0.4766324385709689, 0.4766413555592825],
[0.4747358754098249, 0.4747485972170732],
[0.4728755976222633, 0.4728928778172651],
[0.4710511807198248, 0.4710738070495766],
[0.4692622010511354, 0.4692909953467661],
[0.497679878752004, 0.497679878752004],
[0.4954000154631831, 0.4954000154631831],
[0.4931599760310996, 0.4931599760310996],
[0.4909593232354856, 0.4909593232354856],
[0.4887976172454234, 0.4887976172454234],
[0.4866744160981826, 0.4866744160981826],
[0.4845900488307297, 0.4845892761509014],
[0.4825440987479238, 0.4825417477884414],
[0.4805361483673633, 0.480531380378737],
[0.4785657797808562, 0.4785577226508331],
[0.4766325749933464, 0.476620323050344],
[0.4747361162403673, 0.4747187300733827],
[0.4728759862850626, 0.4728524925799762],
[0.4710517686957791, 0.4710211600879554],
[0.469263048105203, 0.469224283048266]])
res1d = [x[response_decision_energy_index] for r in res for x in r]
assert np.allclose(res1d, [[0.9907410468584376], [0.9816847012836883], [0.9728270478359791], [0.964164228287384], [0.9556924424992138],
[0.9474079491380278], [0.9393111265997224], [0.9313984494721904], [0.9236664515817239], [0.9161117261021626],
[0.9087309255565738], [0.9015207617204032], [0.8944780054345429], [0.8875994863362322], [0.880882092515256],
[0.9907410468584376], [0.9816847012836883], [0.9728270478359791], [0.964164228287384], [0.9556924424992138],
[0.9474079491380278], [0.9393111266035642], [0.9313984495075565], [0.9236664517261579], [0.9161117265115205],
[0.9087309264959722], [0.9015207635977346], [0.8944780088366047], [0.887599492067544], [0.8808821016396065],
[0.9907410468584376], [0.9816847012836883], [0.9728270478359791], [0.964164228287384], [0.9556924424992138],
[0.9474079491380278], [0.939308563971253], [0.9313906911792855], [0.9236507947874026], [0.9160853988421866],
[0.9086910874785843], [0.901464504886388], [0.894402355184426], [0.8875014022102764], [0.8807584692328312]])
if reps == 10:
assert np.allclose(res[0][ntrials0 - 1][response_results_index], [0.42481045, 0.42481045])
assert np.allclose(res[0][-1][response_results_index], [0.43512335, 0.39995991])
assert np.allclose(res[1][ntrials0 - 1][response_results_index], [0.42481045, 0.42481045])
assert np.allclose(res[1][-1][response_results_index], [0.41360321, 0.42121262])
assert np.allclose(res[2][ntrials0 - 1][response_results_index], [0.42481045, 0.42481045])
assert np.allclose(res[2][-1][response_results_index], [0.41621778, 0.40255998])
assert np.allclose(res[0][ntrials0 - 1][response_decision_energy_index], [0.72185566])
assert np.allclose(res[0][-1][response_decision_energy_index], [0.69612758])
assert np.allclose(res[1][ntrials0 - 1][response_decision_energy_index], [0.72185566])
assert np.allclose(res[1][-1][response_decision_energy_index], [0.69685957])
assert np.allclose(res[2][ntrials0 - 1][response_decision_energy_index], [0.72185566])
assert np.allclose(res[2][-1][response_decision_energy_index], [0.67021047])
if reps == 100:
assert np.allclose(res[0][ntrials0 - 1][response_results_index], [0.48590224, 0.48590224])
assert np.allclose(res[0][-1][response_results_index], [0.95967791, 0.21434208])
assert np.allclose(res[1][ntrials0 - 1][response_results_index], [0.48590224, 0.48590224])
assert np.allclose(res[1][-1][response_results_index], [0.55847666, 0.83814112])
assert np.allclose(res[2][ntrials0 - 1][response_results_index], [0.48590224, 0.48590224])
assert np.allclose(res[2][-1][response_results_index], [0.89673726, 0.25100269])
assert np.allclose(res[0][ntrials0 - 1][response_decision_energy_index], [0.94440397])
assert np.allclose(res[0][-1][response_decision_energy_index], [0.82279743])
assert np.allclose(res[1][ntrials0 - 1][response_decision_energy_index], [0.94440397])
assert np.allclose(res[1][-1][response_decision_energy_index], [1.87232903])
assert np.allclose(res[2][ntrials0 - 1][response_decision_energy_index], [0.94440397])
assert np.allclose(res[2][-1][response_decision_energy_index], [0.90033387])
if benchmark.enabled:
benchmark(run, mode)