def fcm_loss(inputs, probs, num_iters=5, m=2):
"""
Args:
inputs: [num_samples, 1, x_dims, 1]
probs: [num_samples, num_clusters, 1, 1].
"""
# centers: [1, num_clusters, x_dims]
k = 1
b = 0
weights = []
delta_probs = []
for i in range(num_iters):
probs_m = tf.pow(probs, m)
centers = cl.reduce_sum(probs_m * inputs, axis=0, keepdims=True) / cl.reduce_sum(probs_m, axis=0, keepdims=True)
# distance matrix with shape [num_samples, num_clusters, 1, 1]
distance_matrix = cl.norm(inputs - centers, axis=(2, 3), keepdims=True)
distance_matrix = tf.pow(distance_matrix, 2 / (m - 1))
probs_plus = 1 / (distance_matrix / cl.reduce_sum(distance_matrix, axis=1, keepdims=True))
delta_probs.append(tf.norm(probs_plus - probs))
weights.append(tf.exp(tf.cast(k * i + b, tf.float32)))
probs = probs_plus
weights = tf.stack(weights, axis=0)
delta_probs = tf.stack(delta_probs, axis=0)
loss = tf.reduce_sum(weights * delta_probs) / tf.reduce_sum(weights)
return loss
def primaryCaps(inputs,
filters,
kernel_size,
strides,
out_caps_dims,
method=None,
name=None):
'''Primary capsule layer.
Args:
inputs: [batch_size, in_height, in_width, in_channels].
filters: Integer, the dimensionality of the output space.
kernel_size: kernel_size
strides: strides
out_caps_dims: A list of 2 integers.
method: the method of calculating probability of entity existence(logistic, norm, None)
Returns:
pose: A 6-D tensor, [batch_size, out_height, out_width, filters] + out_caps_dims
activation: A 4-D tensor, [batch_size, out_height, out_width, filters]
'''
name = "primary_capsule" if name is None else name
with tf.variable_scope(name):
channels = filters * np.prod(out_caps_dims)
channels = channels + filters if method == "logistic" else channels
pose = tf.layers.conv2d(inputs,
channels,
kernel_size=kernel_size,
strides=strides,
activation=None)
shape = cl.shape(pose, name="get_pose_shape")
batch_size = shape[0]
height = shape[1]
width = shape[2]
shape = [batch_size, height, width, filters] + out_caps_dims
if method == 'logistic':
# logistic activation unit
pose, activation_logit = tf.split(pose,
[channels - filters, filters],
axis=-1)
pose = tf.reshape(pose, shape=shape)
activation = tf.sigmoid(activation_logit)
elif method == 'norm' or method is None:
pose = tf.reshape(pose, shape=shape)
squash_on = -2 if out_caps_dims[-1] == 1 else [-2, -1]
pose = cl.ops.squash(pose, axis=squash_on)
activation = cl.norm(pose, axis=(-2, -1))
activation = tf.clip_by_value(activation, 1e-20, 1. - 1e-20)
return (pose, activation)
def squash(inputs, axis=-2, ord="euclidean", name=None):
"""Squashing function.
Args:
inputs: A tensor with shape [batch_size, 1, num_caps, vec_len, 1] or [batch_size, num_caps, vec_len, 1]
ord: Order of the norm. Supported values are 'fro', 'euclidean', 1, 2, np.inf and any positive real number yielding the corresponding p-norm. Default is 'euclidean' which is equivalent to Frobenius norm if tensor is a matrix and equivalent to 2-norm for vectors.
Returns:
A tensor with the same shape as inputs but squashed in `axis` dimension.
"""
name = "squashing" if name is None else name
with tf.name_scope(name):
norm = cl.norm(inputs, ord=ord, axis=axis, keepdims=True)
norm_squared = tf.square(norm)
scalar_factor = norm_squared / (1 + norm_squared)
return scalar_factor * (inputs / norm)
def call(self, inputs):
pose = self.conv2d(inputs)
shape = cl.shape(pose, name="get_pose_shape")
batch_size = shape[0]
height = shape[1]
width = shape[2]
shape = [batch_size, height, width, self.filters] + self.out_caps_dims
if self.method == 'logistic':
# logistic activation unit
num_or_size_splits = [self.channels - self.filters, self.filters]
pose, activation_logit = tf.split(pose,
num_or_size_splits,
axis=-1)
pose = tf.reshape(pose, shape=shape)
activation = tf.sigmoid(activation_logit)
elif self.method == 'norm' or self.method is None:
pose = tf.reshape(pose, shape=shape)
squash_on = -2 if self.out_caps_dims[-1] == 1 else [-2, -1]
pose = cl.ops.squash(pose, axis=squash_on)
activation = cl.norm(pose, axis=(-2, -1))
activation = tf.clip_by_value(activation, 1e-20, 1. - 1e-20)
return (pose, activation)