def test_spn_construction_by_init_and_evaluation():
# building the same levels
input_layer = build_spn_indicator_layer(vars)
sum_layer, prod_layer = build_spn_layers(input_layer)
spn = Spn(input_layer=input_layer, layers=[sum_layer, prod_layer])
res = spn.eval(I)
print('First evaluation')
print(res)
assert_log_array_almost_equal(root_vals, res)
def test_spn_construction_by_add_and_evaluation():
spn = Spn()
# building the same levels
input_layer = build_spn_indicator_layer(vars)
sum_layer, prod_layer = build_spn_layers(input_layer)
# adding all layers to the spn
spn.set_input_layer(input_layer)
spn.add_layer(sum_layer)
spn.add_layer(prod_layer)
res = spn.eval(I)
print('First evaluation')
print(res)
assert_log_array_almost_equal(root_vals, res)
def test_spn_construction_by_add_and_evaluation_II():
spn = Spn()
# print('empty spn')
# print(spn)
input_layer = build_spn_smoothed_layer(vars, dicts, alpha)
prod_layer = build_spn_layers_II(input_layer)
# adding all layers to the spn
spn.set_input_layer(input_layer)
spn.add_layer(prod_layer)
# print('created spn')
# print(spn)
res = spn.eval(I)
print('First smoothed evaluation')
print(res)
assert_log_array_almost_equal(root_vals, res)
def test_spn_set_get_weights():
# create a simple spn
root_node = SumNode()
root_layer = SumLayer([root_node])
prod_node_1 = ProductNode()
prod_node_2 = ProductNode()
root_node.add_child(prod_node_1, 0.5)
root_node.add_child(prod_node_2, 0.5)
prod_layer = ProductLayer([prod_node_1,
prod_node_2])
sum_node_1 = SumNode()
sum_node_2 = SumNode()
sum_node_3 = SumNode()
prod_node_1.add_child(sum_node_1)
prod_node_1.add_child(sum_node_2)
prod_node_2.add_child(sum_node_2)
prod_node_2.add_child(sum_node_3)
sum_layer = SumLayer([sum_node_1, sum_node_2,
sum_node_3])
ind_node_1 = CategoricalIndicatorNode(var=0, var_val=1)
ind_node_2 = CategoricalIndicatorNode(var=0, var_val=1)
ind_node_3 = CategoricalIndicatorNode(var=0, var_val=1)
ind_node_4 = CategoricalIndicatorNode(var=0, var_val=1)
ind_node_5 = CategoricalIndicatorNode(var=0, var_val=1)
input_layer = CategoricalInputLayer(nodes=[ind_node_1,
ind_node_2,
ind_node_3,
ind_node_4,
ind_node_5])
sum_node_1.add_child(ind_node_1, 0.2)
sum_node_1.add_child(ind_node_2, 0.2)
sum_node_2.add_child(ind_node_2, 0.2)
sum_node_2.add_child(ind_node_3, 0.2)
sum_node_2.add_child(ind_node_4, 0.2)
sum_node_3.add_child(ind_node_4, 0.2)
sum_node_3.add_child(ind_node_5, 0.2)
spn = Spn(input_layer=input_layer,
layers=[sum_layer, prod_layer, root_layer])
print(spn)
# storing these weights
curr_weights = spn.get_weights()
# setting the new weights
spn.set_weights(weights_ds)
# getting them again
new_weights = spn.get_weights()
# comparing them
assert new_weights == weights_ds
# now setting back the previous one
spn.set_weights(curr_weights)
# getting them back again
old_weights = spn.get_weights()
# and checking
assert old_weights == curr_weights
def create_valid_toy_spn():
# root layer
whole_scope = frozenset({0, 1, 2, 3})
root_node = SumNode(var_scope=whole_scope)
root_layer = SumLayer([root_node])
# prod layer
prod_node_1 = ProductNode(var_scope=whole_scope)
prod_node_2 = ProductNode(var_scope=whole_scope)
prod_layer_1 = ProductLayer([prod_node_1, prod_node_2])
root_node.add_child(prod_node_1, 0.5)
root_node.add_child(prod_node_2, 0.5)
# sum layer
scope_1 = frozenset({0, 1})
scope_2 = frozenset({2})
scope_3 = frozenset({3})
scope_4 = frozenset({2, 3})
sum_node_1 = SumNode(var_scope=scope_1)
sum_node_2 = SumNode(var_scope=scope_2)
sum_node_3 = SumNode(var_scope=scope_3)
sum_node_4 = SumNode(var_scope=scope_4)
prod_node_1.add_child(sum_node_1)
prod_node_1.add_child(sum_node_2)
prod_node_1.add_child(sum_node_3)
prod_node_2.add_child(sum_node_1)
prod_node_2.add_child(sum_node_4)
sum_layer_1 = SumLayer([sum_node_1, sum_node_2,
sum_node_3, sum_node_4])
# another product layer
prod_node_3 = ProductNode(var_scope=scope_1)
prod_node_4 = ProductNode(var_scope=scope_1)
prod_node_5 = ProductNode(var_scope=scope_4)
prod_node_6 = ProductNode(var_scope=scope_4)
sum_node_1.add_child(prod_node_3, 0.5)
sum_node_1.add_child(prod_node_4, 0.5)
sum_node_4.add_child(prod_node_5, 0.5)
sum_node_4.add_child(prod_node_6, 0.5)
prod_layer_2 = ProductLayer([prod_node_3, prod_node_4,
prod_node_5, prod_node_6])
# last sum one
scope_5 = frozenset({0})
scope_6 = frozenset({1})
sum_node_5 = SumNode(var_scope=scope_5)
sum_node_6 = SumNode(var_scope=scope_6)
sum_node_7 = SumNode(var_scope=scope_5)
sum_node_8 = SumNode(var_scope=scope_6)
sum_node_9 = SumNode(var_scope=scope_2)
sum_node_10 = SumNode(var_scope=scope_3)
sum_node_11 = SumNode(var_scope=scope_2)
sum_node_12 = SumNode(var_scope=scope_3)
prod_node_3.add_child(sum_node_5)
prod_node_3.add_child(sum_node_6)
prod_node_4.add_child(sum_node_7)
prod_node_4.add_child(sum_node_8)
prod_node_5.add_child(sum_node_9)
prod_node_5.add_child(sum_node_10)
prod_node_6.add_child(sum_node_11)
prod_node_6.add_child(sum_node_12)
sum_layer_2 = SumLayer([sum_node_5, sum_node_6,
sum_node_7, sum_node_8,
sum_node_9, sum_node_10,
sum_node_11, sum_node_12])
# input layer
vars = [2, 3, 2, 2]
input_layer = CategoricalIndicatorLayer(vars=vars)
last_sum_nodes = [sum_node_2, sum_node_3,
sum_node_5, sum_node_6,
sum_node_7, sum_node_8,
sum_node_9, sum_node_10,
sum_node_11, sum_node_12]
for sum_node in last_sum_nodes:
(var_scope,) = sum_node.var_scope
for input_node in input_layer.nodes():
if input_node.var == var_scope:
sum_node.add_child(input_node, 1.0)
spn = Spn(input_layer=input_layer,
layers=[sum_layer_2, prod_layer_2,
sum_layer_1, prod_layer_1,
root_layer])
# print(spn)
return spn
def test_spn_mpe_eval_and_traversal():
# create initial layer
node1 = Node()
node2 = Node()
node3 = Node()
node4 = Node()
node5 = Node()
input_layer = CategoricalInputLayer([node1, node2,
node3, node4,
node5])
# top layer made by 3 sum nodes
sum1 = SumNode()
sum2 = SumNode()
sum3 = SumNode()
# linking to input nodes
weight11 = 0.3
sum1.add_child(node1, weight11)
weight12 = 0.3
sum1.add_child(node2, weight12)
weight13 = 0.4
sum1.add_child(node3, weight13)
weight22 = 0.15
sum2.add_child(node2, weight22)
weight23 = 0.15
sum2.add_child(node3, weight23)
weight24 = 0.7
sum2.add_child(node4, weight24)
weight33 = 0.4
sum3.add_child(node3, weight33)
weight34 = 0.25
sum3.add_child(node4, weight34)
weight35 = 0.35
sum3.add_child(node5, weight35)
sum_layer = SumLayer([sum1, sum2, sum3])
# another layer with two product nodes
prod1 = ProductNode()
prod2 = ProductNode()
prod1.add_child(sum1)
prod1.add_child(sum2)
prod2.add_child(sum2)
prod2.add_child(sum3)
prod_layer = ProductLayer([prod1, prod2])
# root layer, double sum
root1 = SumNode()
root2 = SumNode()
weightr11 = 0.5
root1.add_child(prod1, weightr11)
weightr12 = 0.5
root1.add_child(prod2, weightr12)
weightr21 = 0.9
root2.add_child(prod1, weightr21)
weightr22 = 0.1
root2.add_child(prod2, weightr22)
root_layer = SumLayer([root1, root2])
# create the spn
spn = Spn(input_layer=input_layer,
layers=[sum_layer, prod_layer, root_layer])
print('===================')
print(spn)
print('===================')
# setting the input values
val1 = 0.0
node1.set_val(val1)
val2 = 0.5
node2.set_val(val2)
val3 = 0.3
node3.set_val(val3)
val4 = 1.0
node4.set_val(val4)
val5 = 0.0
node5.set_val(val5)
# evaluating the spn with MPE inference
res = spn.test_mpe_eval()
print('spn eval\'d', res)
# testing it
#
# testing the max layer
max1 = max(val1 * weight11,
val2 * weight12,
val3 * weight13)
max2 = max(val2 * weight22,
val3 * weight23,
val4 * weight24)
max3 = max(val3 * weight33,
val4 * weight34,
val5 * weight35)
log_max1 = log(max1) if not numpy.isclose(max1, 0) else LOG_ZERO
log_max2 = log(max2) if not numpy.isclose(max2, 0) else LOG_ZERO
log_max3 = log(max3) if not numpy.isclose(max3, 0) else LOG_ZERO
print('expected max vals {0}, {1}, {2}'.format(log_max1,
log_max2,
log_max3))
print('found max vals {0}, {1}, {2}'.format(sum1.log_val,
sum2.log_val,
sum3.log_val))
if IS_LOG_ZERO(log_max1):
assert IS_LOG_ZERO(sum1.log_val)
else:
assert_almost_equal(log_max1, sum1.log_val)
if IS_LOG_ZERO(log_max2):
assert IS_LOG_ZERO(sum2.log_val)
else:
assert_almost_equal(log_max2, sum2.log_val)
if IS_LOG_ZERO(log_max3):
assert IS_LOG_ZERO(sum3.log_val)
else:
assert_almost_equal(log_max3, sum3.log_val)
# product layer is assumed to be fine, but let's check
# it anyways
prod_val1 = max1 * max2
prod_val2 = max2 * max3
prod_log_val1 = log_max1 + log_max2
prod_log_val2 = log_max2 + log_max3
print('exp prod vals {0}, {1}'.format(prod_log_val1,
prod_log_val2))
print('rea prod vals {0}, {1}'.format(prod1.log_val,
prod2.log_val))
if IS_LOG_ZERO(prod_log_val1):
assert IS_LOG_ZERO(prod1.log_val)
else:
assert_almost_equal(prod_log_val1, prod1.log_val)
if IS_LOG_ZERO(prod_log_val2):
assert IS_LOG_ZERO(prod2.log_val)
else:
assert_almost_equal(prod_log_val2, prod2.log_val)
# root layer, again a sum layer
root_val1 = max(prod_val1 * weightr11,
prod_val2 * weightr12)
root_val2 = max(prod_val1 * weightr21,
prod_val2 * weightr22)
root_log_val1 = log(root_val1) if not numpy.isclose(
root_val1, 0) else LOG_ZERO
root_log_val2 = log(root_val2) if not numpy.isclose(
root_val2, 0) else LOG_ZERO
print('exp root vals {0}, {1}'.format(root_log_val1,
root_log_val2))
print('found ro vals {0}, {1}'.format(root1.log_val,
root2.log_val))
if IS_LOG_ZERO(root_log_val1):
assert IS_LOG_ZERO(root1.log_val)
else:
assert_almost_equal(root_log_val1, root1.log_val)
if IS_LOG_ZERO(root_log_val2):
assert IS_LOG_ZERO(root2.log_val)
else:
assert_almost_equal(root_log_val2, root2.log_val)
# now we are traversing top down the net
print('mpe traversing')
for i, j, k in spn.mpe_traversal():
print(i, j, k)
def test_spn_backprop():
# create initial layer
node1 = Node()
node2 = Node()
node3 = Node()
node4 = Node()
node5 = Node()
input_layer = CategoricalInputLayer([node1, node2,
node3, node4,
node5])
# top layer made by 3 sum nodes
sum1 = SumNode()
sum2 = SumNode()
sum3 = SumNode()
# linking to input nodes
weight11 = 0.3
sum1.add_child(node1, weight11)
weight12 = 0.3
sum1.add_child(node2, weight12)
weight13 = 0.4
sum1.add_child(node3, weight13)
weight22 = 0.15
sum2.add_child(node2, weight22)
weight23 = 0.15
sum2.add_child(node3, weight23)
weight24 = 0.7
sum2.add_child(node4, weight24)
weight33 = 0.4
sum3.add_child(node3, weight33)
weight34 = 0.25
sum3.add_child(node4, weight34)
weight35 = 0.35
sum3.add_child(node5, weight35)
sum_layer = SumLayer([sum1, sum2, sum3])
# another layer with two product nodes
prod1 = ProductNode()
prod2 = ProductNode()
prod1.add_child(sum1)
prod1.add_child(sum2)
prod2.add_child(sum2)
prod2.add_child(sum3)
prod_layer = ProductLayer([prod1, prod2])
# root layer, double sum
root1 = SumNode()
root2 = SumNode()
weightr11 = 0.5
root1.add_child(prod1, weightr11)
weightr12 = 0.5
root1.add_child(prod2, weightr12)
weightr21 = 0.9
root2.add_child(prod1, weightr21)
weightr22 = 0.1
root2.add_child(prod2, weightr22)
root_layer = SumLayer([root1, root2])
# root_layer = SumLayer([root1])
# create the spn
spn = Spn(input_layer=input_layer,
layers=[sum_layer, prod_layer, root_layer])
# setting the input values
val1 = 0.0
node1.set_val(val1)
val2 = 0.5
node2.set_val(val2)
val3 = 0.3
node3.set_val(val3)
val4 = 1.0
node4.set_val(val4)
val5 = 0.0
node5.set_val(val5)
# evaluating the spn
res = spn.test_eval()
print('spn eval\'d', res)
# backprop
spn.backprop()
# computing derivatives by hand
# topdown: root layer
root_der = 1.0
log_root_der = log(root_der)
# print('root ders', root1.log_der, root2.log_der)
print('root ders', root1.log_der)
assert_almost_equal(log_root_der, root1.log_der)
assert_almost_equal(log_root_der, root2.log_der)
# product layer
prod_der1 = (root_der * weightr11 +
root_der * weightr21)
prod_der2 = (root_der * weightr12 +
root_der * weightr22)
# prod_der1 = (root_der * weightr11)
# prod_der2 = (root_der * weightr12)
log_prod_der1 = log(prod_der1) if prod_der1 > 0.0 else LOG_ZERO
log_prod_der2 = log(prod_der2) if prod_der2 > 0.0 else LOG_ZERO
print('found prod ders', prod1.log_der, prod2.log_der)
print('expect prod ders', log_prod_der1, log_prod_der2)
if IS_LOG_ZERO(log_prod_der1):
assert IS_LOG_ZERO(prod1.log_der)
else:
assert_almost_equal(log_prod_der1, prod1.log_der)
if IS_LOG_ZERO(log_prod_der2):
assert IS_LOG_ZERO(prod2.log_der)
else:
assert_almost_equal(log_prod_der2, prod2.log_der)
# sum layer
sum_der1 = (
prod_der1 * (weight22 * val2 +
weight23 * val3 +
weight24 * val4))
log_sum_der1 = log(sum_der1) if sum_der1 > 0.0 else LOG_ZERO
sum_der2 = (prod_der1 * (weight11 * val1 +
weight12 * val2 +
weight13 * val3) +
prod_der2 * (weight33 * val3 +
weight34 * val4 +
weight35 * val5))
log_sum_der2 = log(sum_der2) if sum_der2 > 0.0 else LOG_ZERO
sum_der3 = (prod_der2 * (weight22 * val2 +
weight23 * val3 +
weight24 * val4))
log_sum_der3 = log(sum_der3) if sum_der3 > 0.0 else LOG_ZERO
print('expected sum ders', log_sum_der1,
log_sum_der2,
log_sum_der3)
print('found sum ders', sum1.log_der,
sum2.log_der,
sum3.log_der)
if IS_LOG_ZERO(log_sum_der1):
assert IS_LOG_ZERO(sum1.log_der)
else:
assert_almost_equal(log_sum_der1, sum1.log_der)
if IS_LOG_ZERO(log_sum_der2):
assert IS_LOG_ZERO(sum2.log_der)
else:
assert_almost_equal(log_sum_der2, sum2.log_der)
if IS_LOG_ZERO(log_sum_der3):
assert IS_LOG_ZERO(sum3.log_der)
else:
assert_almost_equal(log_sum_der3, sum3.log_der)
# final level, the first one
try:
log_der1 = log(sum_der1 * weight11)
except:
log_der1 = LOG_ZERO
try:
log_der2 = log(sum_der1 * weight12 +
sum_der2 * weight22)
except:
log_der2 = LOG_ZERO
try:
log_der3 = log(sum_der1 * weight13 +
sum_der2 * weight23 +
sum_der3 * weight33)
except:
log_der3 = LOG_ZERO
try:
log_der4 = log(sum_der2 * weight24 +
sum_der3 * weight34)
except:
log_der4 = LOG_ZERO
try:
log_der5 = log(sum_der3 * weight35)
except:
log_der5 = LOG_ZERO
# printing, just in case
print('child log der', node1.log_der, node2.log_der,
node3.log_der, node4.log_der, node5.log_der)
print('exact log der', log_der1, log_der2, log_der3,
log_der4, log_der5)
if IS_LOG_ZERO(log_der1):
assert IS_LOG_ZERO(node1.log_der)
else:
assert_almost_equal(log_der1, node1.log_der, 15)
if IS_LOG_ZERO(log_der2):
assert IS_LOG_ZERO(node2.log_der)
else:
assert_almost_equal(log_der2, node2.log_der, 15)
if IS_LOG_ZERO(log_der3):
assert IS_LOG_ZERO(node3.log_der)
else:
assert_almost_equal(log_der3, node3.log_der, 15)
if IS_LOG_ZERO(log_der4):
assert IS_LOG_ZERO(node4.log_der)
else:
assert_almost_equal(log_der4, node4.log_der, 15)
if IS_LOG_ZERO(log_der5):
assert IS_LOG_ZERO(node5.log_der)
else:
assert_almost_equal(log_der5, node5.log_der, 15)
def test_linked_to_theano_indicator():
# creating single nodes
root = SumNode()
prod1 = ProductNode()
prod2 = ProductNode()
prod3 = ProductNode()
sum1 = SumNode()
sum2 = SumNode()
sum3 = SumNode()
sum4 = SumNode()
ind1 = CategoricalIndicatorNode(var=0, var_val=0)
ind2 = CategoricalIndicatorNode(var=0, var_val=1)
ind3 = CategoricalIndicatorNode(var=1, var_val=0)
ind4 = CategoricalIndicatorNode(var=1, var_val=1)
ind5 = CategoricalIndicatorNode(var=2, var_val=0)
ind6 = CategoricalIndicatorNode(var=2, var_val=1)
ind7 = CategoricalIndicatorNode(var=2, var_val=2)
ind8 = CategoricalIndicatorNode(var=3, var_val=0)
ind9 = CategoricalIndicatorNode(var=3, var_val=1)
ind10 = CategoricalIndicatorNode(var=3, var_val=2)
ind11 = CategoricalIndicatorNode(var=3, var_val=3)
prod4 = ProductNode()
prod5 = ProductNode()
prod6 = ProductNode()
prod7 = ProductNode()
# linking nodes
root.add_child(prod1, 0.3)
root. add_child(prod2, 0.3)
root.add_child(prod3, 0.4)
prod1.add_child(sum1)
prod1.add_child(sum2)
prod2.add_child(ind7)
prod2.add_child(ind8)
prod2.add_child(ind11)
prod3.add_child(sum3)
prod3.add_child(sum4)
sum1.add_child(ind1, 0.3)
sum1.add_child(ind2, 0.3)
sum1.add_child(prod4, 0.4)
sum2.add_child(ind2, 0.5)
sum2.add_child(prod4, 0.2)
sum2.add_child(prod5, 0.3)
sum3.add_child(prod6, 0.5)
sum3.add_child(prod7, 0.5)
sum4.add_child(prod6, 0.5)
sum4.add_child(prod7, 0.5)
prod4.add_child(ind3)
prod4.add_child(ind4)
prod5.add_child(ind5)
prod5.add_child(ind6)
prod6.add_child(ind9)
prod6.add_child(ind10)
prod7.add_child(ind9)
prod7.add_child(ind10)
# building layers from nodes
root_layer = SumLayerLinked([root])
prod_layer = ProductLayerLinked([prod1, prod2, prod3])
sum_layer = SumLayerLinked([sum1, sum2, sum3, sum4])
aprod_layer = ProductLayerLinked([prod4, prod5, prod6, prod7])
ind_layer = CategoricalIndicatorLayer(nodes=[ind1, ind2,
ind3, ind4,
ind5, ind6,
ind7, ind8,
ind9, ind10,
ind11])
# creating the linked spn
spn_linked = SpnLinked(input_layer=ind_layer,
layers=[aprod_layer,
sum_layer,
prod_layer,
root_layer])
print(spn_linked)
# converting to theano repr
spn_theano = SpnFactory.linked_to_theano(spn_linked)
print(spn_theano)
# time for some inference comparison
for instance in I:
print('linked')
res_l = spn_linked.eval(instance)
print(res_l)
print('theano')
res_t = spn_theano.eval(instance)
print(res_t)
assert_array_almost_equal(res_l, res_t)
def test_spn_sampling():
from collections import Counter
from spn.factory import linked_categorical_input_to_indicators
#
# building a small mixture model
features = [2, 2, 2, 2]
n_features = len(features)
#
# different categorical vars groups as leaves
input_nodes_1 = [
CategoricalSmoothedNode(i, features[i], alpha=0.0, freqs=[0, 1])
for i in range(n_features)
]
input_nodes_2 = [
CategoricalSmoothedNode(i, features[i], alpha=0.0, freqs=[1, 0])
for i in range(n_features)
]
input_nodes_3 = [CategoricalSmoothedNode(i, features[i], alpha=0.0,
freqs=[1, 0]) for i in range(n_features // 2)] + \
[CategoricalSmoothedNode(i, features[i], alpha=0.0,
freqs=[0, 1]) for i in range(n_features // 2, n_features)]
input_nodes_4 = [CategoricalSmoothedNode(i, features[i], alpha=0.0,
freqs=[0, 1]) for i in range(n_features // 2)] + \
[CategoricalSmoothedNode(i, features[i], alpha=0.0,
freqs=[1, 0]) for i in range(n_features // 2, n_features)]
input_layer = CategoricalSmoothedLayer(
nodes=input_nodes_1 + input_nodes_2 + input_nodes_3 + input_nodes_4)
#
# one product node for each group
prod_node_1 = ProductNode()
for leaf in input_nodes_1:
prod_node_1.add_child(leaf)
prod_node_2 = ProductNode()
for leaf in input_nodes_2:
prod_node_2.add_child(leaf)
prod_node_3 = ProductNode()
for leaf in input_nodes_3:
prod_node_3.add_child(leaf)
prod_node_4 = ProductNode()
for leaf in input_nodes_4:
prod_node_4.add_child(leaf)
prod_layer = ProductLayer(
nodes=[prod_node_1, prod_node_2, prod_node_3, prod_node_4])
#
# one root as a mixture
root = SumNode()
root.add_child(prod_node_1, 0.5)
root.add_child(prod_node_2, 0.1)
root.add_child(prod_node_3, 0.2)
root.add_child(prod_node_4, 0.2)
root_layer = SumLayer(nodes=[root])
spn = Spn(input_layer=input_layer, layers=[prod_layer, root_layer])
print(spn)
n_instances = 1000
#
# sampling some instances
sample_start_t = perf_counter()
samples = spn.sample(n_instances=n_instances, verbose=False)
sample_end_t = perf_counter()
print('Sampled in {} secs'.format(sample_end_t - sample_start_t))
if n_instances < 20:
print(samples)
#
# some statistics
tuple_samples = [tuple(s) for s in samples]
if n_instances < 20:
print(tuple_samples)
sample_counter = Counter(tuple_samples)
print(sample_counter)
#
# transforming into an spn with indicator nodes
print('Into indicator nodes')
ind_start_t = perf_counter()
spn = linked_categorical_input_to_indicators(spn)
ind_end_t = perf_counter()
print('Done in ', ind_end_t - ind_start_t)
sample_start_t = perf_counter()
samples = spn.sample(n_instances=n_instances,
verbose=False,
one_hot_encoding=True)
sample_end_t = perf_counter()
print('Sampled in {} secs'.format(sample_end_t - sample_start_t))
if n_instances < 20:
print(samples)
#
# some statistics
tuple_samples = [tuple(s) for s in samples]
if n_instances < 20:
print(tuple_samples)
sample_counter = Counter(tuple_samples)
print(sample_counter)
def test_linked_to_theano_categorical():
vars = [2, 2, 3, 4]
freqs = [{'var': 0, 'freqs': [1, 2]},
{'var': 1, 'freqs': [2, 2]},
{'var': 0, 'freqs': [3, 2]},
{'var': 1, 'freqs': [0, 3]},
{'var': 2, 'freqs': [1, 0, 2]},
{'var': 3, 'freqs': [1, 2, 1, 2]},
{'var': 3, 'freqs': [3, 4, 0, 1]}]
# create input layer first
input_layer = CategoricalSmoothedLayer(vars=vars,
node_dicts=freqs)
# get nodes
ind_nodes = [node for node in input_layer.nodes()]
root_node = ProductNode()
sum1 = SumNode()
sum2 = SumNode()
prod1 = ProductNode()
prod2 = ProductNode()
sum3 = SumNode()
sum4 = SumNode()
# linking
root_node.add_child(sum1)
root_node.add_child(sum2)
root_node.add_child(ind_nodes[0])
root_node.add_child(ind_nodes[1])
sum1.add_child(ind_nodes[2], 0.4)
sum1.add_child(ind_nodes[3], 0.6)
sum2.add_child(ind_nodes[3], 0.2)
sum2.add_child(prod1, 0.5)
sum2.add_child(prod2, 0.3)
prod1.add_child(ind_nodes[4])
prod1.add_child(sum3)
prod1.add_child(sum4)
prod2.add_child(sum3)
prod2.add_child(sum4)
sum3.add_child(ind_nodes[5], 0.5)
sum3.add_child(ind_nodes[6], 0.5)
sum4.add_child(ind_nodes[5], 0.4)
sum4.add_child(ind_nodes[6], 0.6)
# creating layers
root_layer = ProductLayerLinked([root_node])
sum_layer = SumLayerLinked([sum1, sum2])
prod_layer = ProductLayerLinked([prod1, prod2])
sum_layer2 = SumLayerLinked([sum3, sum4])
# create the linked spn
spn_linked = SpnLinked(input_layer=input_layer,
layers=[sum_layer2, prod_layer,
sum_layer, root_layer])
print(spn_linked)
# converting to theano repr
spn_theano = SpnFactory.linked_to_theano(spn_linked)
print(spn_theano)
# time for some inference comparison
for instance in I:
print('linked')
res_l = spn_linked.eval(instance)
print(res_l)
print('theano')
res_t = spn_theano.eval(instance)
print(res_t)
assert_array_almost_equal(res_l, res_t)
def test_toy_spn_numpy_linked():
input_vec = numpy.array([[0., 0., 0.], [0., 0., 0.], [0., 1., 1.],
[MARG_IND, MARG_IND, MARG_IND]]).T
ind_node_1 = CategoricalIndicatorNode(var=0, var_val=0)
ind_node_2 = CategoricalIndicatorNode(var=0, var_val=1)
ind_node_3 = CategoricalIndicatorNode(var=1, var_val=0)
ind_node_4 = CategoricalIndicatorNode(var=1, var_val=1)
ind_node_5 = CategoricalIndicatorNode(var=2, var_val=0)
ind_node_6 = CategoricalIndicatorNode(var=2, var_val=1)
input_layer = CategoricalInputLayer(nodes=[
ind_node_1, ind_node_2, ind_node_3, ind_node_4, ind_node_5, ind_node_6
])
n_nodes_layer_1 = 6
layer_1_sum_nodes = [SumNode() for i in range(n_nodes_layer_1)]
layer_1_sum_nodes[0].add_child(ind_node_1, 0.6)
layer_1_sum_nodes[0].add_child(ind_node_2, 0.4)
layer_1_sum_nodes[1].add_child(ind_node_1, 0.3)
layer_1_sum_nodes[1].add_child(ind_node_2, 0.7)
layer_1_sum_nodes[2].add_child(ind_node_3, 0.1)
layer_1_sum_nodes[2].add_child(ind_node_4, 0.9)
layer_1_sum_nodes[3].add_child(ind_node_3, 0.7)
layer_1_sum_nodes[3].add_child(ind_node_4, 0.3)
layer_1_sum_nodes[4].add_child(ind_node_5, 0.5)
layer_1_sum_nodes[4].add_child(ind_node_6, 0.5)
layer_1_sum_nodes[5].add_child(ind_node_5, 0.2)
layer_1_sum_nodes[5].add_child(ind_node_6, 0.8)
layer_1 = SumLayer(layer_1_sum_nodes)
n_nodes_layer_2 = 4
layer_2_prod_nodes = [ProductNode() for i in range(n_nodes_layer_2)]
layer_2_prod_nodes[0].add_child(layer_1_sum_nodes[0])
layer_2_prod_nodes[0].add_child(layer_1_sum_nodes[2])
layer_2_prod_nodes[0].add_child(layer_1_sum_nodes[4])
layer_2_prod_nodes[1].add_child(layer_1_sum_nodes[1])
layer_2_prod_nodes[1].add_child(layer_1_sum_nodes[3])
layer_2_prod_nodes[1].add_child(layer_1_sum_nodes[5])
layer_2_prod_nodes[2].add_child(layer_1_sum_nodes[0])
layer_2_prod_nodes[2].add_child(layer_1_sum_nodes[2])
layer_2_prod_nodes[2].add_child(layer_1_sum_nodes[5])
layer_2_prod_nodes[3].add_child(layer_1_sum_nodes[1])
layer_2_prod_nodes[3].add_child(layer_1_sum_nodes[3])
layer_2_prod_nodes[3].add_child(layer_1_sum_nodes[4])
layer_2 = ProductLayer(layer_2_prod_nodes)
root = SumNode()
root.add_child(layer_2_prod_nodes[0], 0.2)
root.add_child(layer_2_prod_nodes[1], 0.4)
root.add_child(layer_2_prod_nodes[2], 0.15)
root.add_child(layer_2_prod_nodes[3], 0.25)
layer_3 = SumLayer([root])
spn = Spn(input_layer=input_layer, layers=[layer_1, layer_2, layer_3])
res = spn.eval(input_vec)
print('First evaluation')
print(res)
def test_mini_spn_fit_em():
vars = numpy.array([2, 2, 2, 2])
input_layer = CategoricalIndicatorLayer(vars=vars)
print(input_layer)
ind1 = input_layer._nodes[0]
ind2 = input_layer._nodes[1]
ind3 = input_layer._nodes[2]
ind4 = input_layer._nodes[3]
ind5 = input_layer._nodes[4]
ind6 = input_layer._nodes[5]
ind7 = input_layer._nodes[6]
ind8 = input_layer._nodes[7]
# creating a sum layer of 4 nodes
sum1 = SumNode()
sum2 = SumNode()
sum3 = SumNode()
sum4 = SumNode()
sum1.add_child(ind1, 0.6)
sum1.add_child(ind2, 0.4)
sum2.add_child(ind3, 0.5)
sum2.add_child(ind4, 0.5)
sum3.add_child(ind5, 0.7)
sum3.add_child(ind6, 0.3)
sum4.add_child(ind7, 0.4)
sum4.add_child(ind8, 0.6)
sum_layer = SumLayer(nodes=[sum1, sum2, sum3, sum4])
# and a top layer of 3 products
prod1 = ProductNode()
prod2 = ProductNode()
prod3 = ProductNode()
prod1.add_child(sum1)
prod1.add_child(sum2)
prod2.add_child(sum2)
prod2.add_child(sum3)
prod3.add_child(sum3)
prod3.add_child(sum4)
prod_layer = ProductLayer(nodes=[prod1, prod2, prod3])
# root layer
root = SumNode()
root.add_child(prod1, 0.4)
root.add_child(prod2, 0.25)
root.add_child(prod3, 0.35)
root_layer = SumLayer(nodes=[root])
spn = Spn(input_layer=input_layer,
layers=[sum_layer, prod_layer, root_layer])
print(spn)
# training on obs
spn.fit_em(train=syn_train_data, valid=syn_val_data, test=None, hard=True)
def test_linked_to_theano_indicator():
# creating single nodes
root = SumNode()
prod1 = ProductNode()
prod2 = ProductNode()
prod3 = ProductNode()
sum1 = SumNode()
sum2 = SumNode()
sum3 = SumNode()
sum4 = SumNode()
ind1 = CategoricalIndicatorNode(var=0, var_val=0)
ind2 = CategoricalIndicatorNode(var=0, var_val=1)
ind3 = CategoricalIndicatorNode(var=1, var_val=0)
ind4 = CategoricalIndicatorNode(var=1, var_val=1)
ind5 = CategoricalIndicatorNode(var=2, var_val=0)
ind6 = CategoricalIndicatorNode(var=2, var_val=1)
ind7 = CategoricalIndicatorNode(var=2, var_val=2)
ind8 = CategoricalIndicatorNode(var=3, var_val=0)
ind9 = CategoricalIndicatorNode(var=3, var_val=1)
ind10 = CategoricalIndicatorNode(var=3, var_val=2)
ind11 = CategoricalIndicatorNode(var=3, var_val=3)
prod4 = ProductNode()
prod5 = ProductNode()
prod6 = ProductNode()
prod7 = ProductNode()
# linking nodes
root.add_child(prod1, 0.3)
root.add_child(prod2, 0.3)
root.add_child(prod3, 0.4)
prod1.add_child(sum1)
prod1.add_child(sum2)
prod2.add_child(ind7)
prod2.add_child(ind8)
prod2.add_child(ind11)
prod3.add_child(sum3)
prod3.add_child(sum4)
sum1.add_child(ind1, 0.3)
sum1.add_child(ind2, 0.3)
sum1.add_child(prod4, 0.4)
sum2.add_child(ind2, 0.5)
sum2.add_child(prod4, 0.2)
sum2.add_child(prod5, 0.3)
sum3.add_child(prod6, 0.5)
sum3.add_child(prod7, 0.5)
sum4.add_child(prod6, 0.5)
sum4.add_child(prod7, 0.5)
prod4.add_child(ind3)
prod4.add_child(ind4)
prod5.add_child(ind5)
prod5.add_child(ind6)
prod6.add_child(ind9)
prod6.add_child(ind10)
prod7.add_child(ind9)
prod7.add_child(ind10)
# building layers from nodes
root_layer = SumLayerLinked([root])
prod_layer = ProductLayerLinked([prod1, prod2, prod3])
sum_layer = SumLayerLinked([sum1, sum2, sum3, sum4])
aprod_layer = ProductLayerLinked([prod4, prod5, prod6, prod7])
ind_layer = CategoricalIndicatorLayer(nodes=[
ind1, ind2, ind3, ind4, ind5, ind6, ind7, ind8, ind9, ind10, ind11
])
# creating the linked spn
spn_linked = SpnLinked(
input_layer=ind_layer,
layers=[aprod_layer, sum_layer, prod_layer, root_layer])
print(spn_linked)
# converting to theano repr
spn_theano = SpnFactory.linked_to_theano(spn_linked)
print(spn_theano)
# time for some inference comparison
for instance in I:
print('linked')
res_l = spn_linked.eval(instance)
print(res_l)
print('theano')
res_t = spn_theano.eval(instance)
print(res_t)
assert_array_almost_equal(res_l, res_t)
def test_linked_to_theano_categorical():
vars = [2, 2, 3, 4]
freqs = [{
'var': 0,
'freqs': [1, 2]
}, {
'var': 1,
'freqs': [2, 2]
}, {
'var': 0,
'freqs': [3, 2]
}, {
'var': 1,
'freqs': [0, 3]
}, {
'var': 2,
'freqs': [1, 0, 2]
}, {
'var': 3,
'freqs': [1, 2, 1, 2]
}, {
'var': 3,
'freqs': [3, 4, 0, 1]
}]
# create input layer first
input_layer = CategoricalSmoothedLayer(vars=vars, node_dicts=freqs)
# get nodes
ind_nodes = [node for node in input_layer.nodes()]
root_node = ProductNode()
sum1 = SumNode()
sum2 = SumNode()
prod1 = ProductNode()
prod2 = ProductNode()
sum3 = SumNode()
sum4 = SumNode()
# linking
root_node.add_child(sum1)
root_node.add_child(sum2)
root_node.add_child(ind_nodes[0])
root_node.add_child(ind_nodes[1])
sum1.add_child(ind_nodes[2], 0.4)
sum1.add_child(ind_nodes[3], 0.6)
sum2.add_child(ind_nodes[3], 0.2)
sum2.add_child(prod1, 0.5)
sum2.add_child(prod2, 0.3)
prod1.add_child(ind_nodes[4])
prod1.add_child(sum3)
prod1.add_child(sum4)
prod2.add_child(sum3)
prod2.add_child(sum4)
sum3.add_child(ind_nodes[5], 0.5)
sum3.add_child(ind_nodes[6], 0.5)
sum4.add_child(ind_nodes[5], 0.4)
sum4.add_child(ind_nodes[6], 0.6)
# creating layers
root_layer = ProductLayerLinked([root_node])
sum_layer = SumLayerLinked([sum1, sum2])
prod_layer = ProductLayerLinked([prod1, prod2])
sum_layer2 = SumLayerLinked([sum3, sum4])
# create the linked spn
spn_linked = SpnLinked(
input_layer=input_layer,
layers=[sum_layer2, prod_layer, sum_layer, root_layer])
print(spn_linked)
# converting to theano repr
spn_theano = SpnFactory.linked_to_theano(spn_linked)
print(spn_theano)
# time for some inference comparison
for instance in I:
print('linked')
res_l = spn_linked.eval(instance)
print(res_l)
print('theano')
res_t = spn_theano.eval(instance)
print(res_t)
assert_array_almost_equal(res_l, res_t)
def test_build_linked_spn_from_scope_graph():
#
# creating a region graph as an input scope graph
n_cols = 2
n_rows = 2
coarse = 2
#
# create initial region
root_region = Region.create_whole_region(n_rows, n_cols)
region_graph = create_poon_region_graph(root_region, coarse=coarse)
# print(region_graph)
print('# partitions', region_graph.n_partitions())
print('# regions', region_graph.n_scopes())
print(region_graph)
#
#
k = 2
spn = build_linked_spn_from_scope_graph(region_graph, k)
print(spn)
print(spn.stats())
#
# back to the scope graph
root_layer = list(spn.root_layer().nodes())
assert len(root_layer) == 1
root = root_layer[0]
scope_graph = get_scope_graph_from_linked_spn(root)
print(scope_graph)
assert scope_graph == region_graph
#
# building an spn from scratch
#
# building leaf nodes
n_vars = 4
vars = [0, 1, 2, 3]
leaves = [
CategoricalIndicatorNode(var, val) for var in range(n_vars)
for val in [0, 1]
]
input_layer = CategoricalIndicatorLayer(nodes=leaves, vars=vars)
#
# building root
root_node = SumNode(var_scope=frozenset(vars))
root_layer = SumLayer([root_node])
#
# building product nodes
prod_list_1 = [ProductNode(var_scope=vars) for i in range(4)]
prod_list_2 = [ProductNode(var_scope=vars) for i in range(4)]
prod_nodes_1 = prod_list_1 + prod_list_2
product_layer_1 = ProductLayer(prod_nodes_1)
for p in prod_nodes_1:
root_node.add_child(p, 1.0 / len(prod_nodes_1))
#
# build sum nodes
sum_list_1 = [SumNode() for i in range(2)]
sum_list_2 = [SumNode() for i in range(2)]
sum_list_3 = [SumNode() for i in range(2)]
sum_list_4 = [SumNode() for i in range(2)]
sum_layer_2 = SumLayer(sum_list_1 + sum_list_2 + sum_list_3 + sum_list_4)
sum_pairs = []
for s_1 in sum_list_1:
for s_2 in sum_list_2:
sum_pairs.append((s_1, s_2))
for p, (s_1, s_2) in zip(prod_list_1, sum_pairs):
p.add_child(s_1)
p.add_child(s_2)
sum_pairs = []
for s_3 in sum_list_3:
for s_4 in sum_list_4:
sum_pairs.append((s_3, s_4))
for p, (s_3, s_4) in zip(prod_list_2, sum_pairs):
p.add_child(s_3)
p.add_child(s_4)
#
# again product nodes
prod_list_3 = [ProductNode() for i in range(4)]
prod_list_4 = [ProductNode() for i in range(4)]
prod_list_5 = [ProductNode() for i in range(4)]
prod_list_6 = [ProductNode() for i in range(4)]
product_layer_3 = ProductLayer(prod_list_3 + prod_list_4 + prod_list_5 +
prod_list_6)
for s in sum_list_1:
for p in prod_list_3:
s.add_child(p, 1.0 / len(prod_list_3))
for s in sum_list_2:
for p in prod_list_4:
s.add_child(p, 1.0 / len(prod_list_4))
for s in sum_list_3:
for p in prod_list_5:
s.add_child(p, 1.0 / len(prod_list_5))
for s in sum_list_4:
for p in prod_list_6:
s.add_child(p, 1.0 / len(prod_list_6))
#
# build sum nodes
sum_list_5 = [SumNode() for i in range(2)]
sum_list_6 = [SumNode() for i in range(2)]
sum_list_7 = [SumNode() for i in range(2)]
sum_list_8 = [SumNode() for i in range(2)]
sum_layer_4 = SumLayer(sum_list_5 + sum_list_6 + sum_list_7 + sum_list_8)
sum_pairs = []
for s_5 in sum_list_5:
for s_7 in sum_list_7:
sum_pairs.append((s_5, s_7))
for p, (s_5, s_7) in zip(prod_list_3, sum_pairs):
p.add_child(s_5)
p.add_child(s_7)
sum_pairs = []
for s_6 in sum_list_6:
for s_8 in sum_list_8:
sum_pairs.append((s_6, s_8))
for p, (s_6, s_8) in zip(prod_list_4, sum_pairs):
p.add_child(s_6)
p.add_child(s_8)
sum_pairs = []
for s_5 in sum_list_5:
for s_6 in sum_list_6:
sum_pairs.append((s_5, s_6))
for p, (s_5, s_6) in zip(prod_list_5, sum_pairs):
p.add_child(s_5)
p.add_child(s_6)
sum_pairs = []
for s_7 in sum_list_7:
for s_8 in sum_list_8:
sum_pairs.append((s_7, s_8))
for p, (s_7, s_8) in zip(prod_list_6, sum_pairs):
p.add_child(s_7)
p.add_child(s_8)
#
# linking to input layer
for s in sum_list_5:
for i in leaves[0:2]:
s.add_child(i, 0.5)
for s in sum_list_6:
for i in leaves[2:4]:
s.add_child(i, 0.5)
for s in sum_list_7:
for i in leaves[4:6]:
s.add_child(i, 0.5)
for s in sum_list_8:
for i in leaves[6:]:
s.add_child(i, 0.5)
lspn = LinkedSpn(input_layer=input_layer,
layers=[
sum_layer_4, product_layer_3, sum_layer_2,
product_layer_1, root_layer
])
print(lspn)
print(lspn.stats())
#
# trying to evaluate them
input_vec = numpy.array([[1., 1., 1., 0.], [0., 0., 0., 0.],
[0., 1., 1., 0.],
[MARG_IND, MARG_IND, MARG_IND, MARG_IND]]).T
res = spn.eval(input_vec)
print('First evaluation')
print(res)
res = lspn.eval(input_vec)
print('Second evaluation')
print(res)