# ---------------------------------------------------------------#
# astroNN.nn.losses: losses
# ---------------------------------------------------------------#
import tensorflow as tf
from tensorflow import keras as tfk
from tensorflow.python.ops.losses import util as tf_losses_util
from astroNN.config import MAGIC_NUMBER
from astroNN.nn import nn_obj_lookup
epsilon = tfk.backend.epsilon
Model = tfk.models.Model
def magic_num_check(x):
# check for magic num and nan
return tf.logical_or(tf.equal(x, MAGIC_NUMBER), tf.math.is_nan(x))
[docs]
def magic_correction_term(y_true):
"""
Calculate a correction term to prevent the loss being "lowered" by magic_num or NaN
:param y_true: Ground Truth
:type y_true: tf.Tensor
:return: Correction Term
:rtype: tf.Tensor
:History:
| 2018-Jan-30 - Written - Henry Leung (University of Toronto)
| 2018-Feb-17 - Updated - Henry Leung (University of Toronto)
"""
num_nonmagic = tf.reduce_sum(
tf.cast(tf.logical_not(magic_num_check(y_true)), tf.float32), axis=-1
)
num_magic = tf.reduce_sum(tf.cast(magic_num_check(y_true), tf.float32), axis=-1)
# If no magic number, then num_zero=0 and whole expression is just 1 and get back our good old loss
# If num_nonzero is 0, that means we don't have any information, then set the correction term to ones
return (num_nonmagic + num_magic) / num_nonmagic
def weighted_loss(losses, sample_weight=None):
"""
Calculate sample-weighted losses from losses
:param losses: Losses
:type losses: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Weighted loss
:rtype: tf.Tensor
:History: 2021-Feb-02 - Written - Henry Leung (University of Toronto)
"""
if sample_weight is None:
return losses
else:
return tf.math.multiply(losses, sample_weight)
def median(x, axis=None):
"""
Calculate median
:param x: Data
:type x: tf.Tensor
:param axis: Axis
:type axis: int
:return: Variance
:rtype: tf.Tensor
:History: 2021-Aug-13 - Written - Henry Leung (University of Toronto)
"""
def median_internal(_x):
shape = tf.shape(_x)[0]
if shape % 2 == 1:
_median = tf.nn.top_k(_x, shape // 2 + 1).values[-1]
else:
_median = (
tf.nn.top_k(_x, shape // 2).values[-1]
+ tf.nn.top_k(_x, shape // 2 + 1).values[-1]
) / 2
return _median
if axis is None:
x_flattened = tf.reshape(x, [-1])
median = median_internal(x_flattened)
return median
else:
x_unstacked = tf.unstack(tf.transpose(x), axis=axis)
median = tf.stack([median_internal(_x) for _x in x_unstacked])
return median
[docs]
def mean_squared_error(y_true, y_pred, sample_weight=None):
"""
Calculate mean square error losses
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Mean Squared Error
:rtype: tf.Tensor
:History: 2017-Nov-16 - Written - Henry Leung (University of Toronto)
"""
losses = tf.reduce_mean(
tf.where(
magic_num_check(y_true), tf.zeros_like(y_true), tf.square(y_true - y_pred)
),
axis=-1,
) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
def mean_squared_reconstruction_error(y_true, y_pred, sample_weight=None):
"""
Calculate mean square reconstruction error losses
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Mean Squared Error
:rtype: tf.Tensor
:History: 2022-May-05 - Written - Henry Leung (University of Toronto)
"""
raw_loss = tf.where(
magic_num_check(y_true), tf.zeros_like(y_true), tf.square(y_true - y_pred)
)
losses = weighted_loss(tf.reduce_mean(raw_loss, axis=-1), sample_weight)
return tf.reduce_mean(tf.reduce_sum(losses, axis=-1))
[docs]
def mse_lin_wrapper(var, labels_err):
"""
Calculate predictive variance, and takes account of labels error in Bayesian Neural Network
:param var: Predictive Variance
:type var: Union(tf.Tensor, tf.Variable)
:param labels_err: Known labels error, give zeros if unknown/unavailable
:type labels_err: Union(tf.Tensor, tf.Variable)
:return: Robust MSE function for labels prediction neurones, which matches Keras losses API
:rtype: function
:Returned Funtion Parameter:
| **function(y_true, y_pred)**
| - **y_true** (*tf.Tensor*): Ground Truth
| - **y_pred** (*tf.Tensor*): Prediction
| Return (*tf.Tensor*): Robust Mean Squared Error
:History: 2017-Nov-16 - Written - Henry Leung (University of Toronto)
"""
def mse_lin(y_true, y_pred, sample_weight=None):
return robust_mse(y_true, y_pred, var, labels_err, sample_weight)
mse_lin.__name__ = (
"mse_lin_wrapper" # set the name to be the same as parent so it can be found
)
return mse_lin
[docs]
def mse_var_wrapper(lin, labels_err):
"""
Calculate predictive variance, and takes account of labels error in Bayesian Neural Network
:param lin: Prediction
:type lin: Union(tf.Tensor, tf.Variable)
:param labels_err: Known labels error, give zeros if unknown/unavailable
:type labels_err: Union(tf.Tensor, tf.Variable)
:return: Robust MSE function for predictive variance neurones which matches Keras losses API
:rtype: function
:Returned Funtion Parameter:
| **function(y_true, y_pred)**
| - **y_true** (*tf.Tensor*): Ground Truth
| - **y_pred** (*tf.Tensor*): Predictive Variance
| Return (*tf.Tensor*): Robust Mean Squared Error
:History: 2017-Nov-16 - Written - Henry Leung (University of Toronto)
"""
def mse_var(y_true, y_pred, sample_weight=None):
return robust_mse(y_true, lin, y_pred, labels_err, sample_weight)
mse_var.__name__ = (
"mse_var_wrapper" # set the name to be the same as parent so it can be found
)
return mse_var
[docs]
def robust_mse(y_true, y_pred, variance, labels_err, sample_weight=None):
"""
Calculate predictive variance, and takes account of labels error in Bayesian Neural Network
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param variance: Log Predictive Variance
:type variance: Union(tf.Tensor, tf.Variable)
:param labels_err: Known labels error, give zeros if unknown/unavailable
:type labels_err: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Robust Mean Squared Error, can be used directly with Tensorflow
:rtype: tf.Tensor
:History: 2018-April-07 - Written - Henry Leung (University of Toronto)
"""
# labels_err still contains magic_number
labels_err_y = tf.where(
magic_num_check(y_true), tf.zeros_like(y_true), tf.cast(labels_err, tf.float32)
)
# Neural Net is predicting log(var), so take exp, takes account the target variance, and take log back
y_pred_corrected = tf.math.log(tf.exp(variance) + tf.square(labels_err_y))
wrapper_output = tf.where(
magic_num_check(y_true),
tf.zeros_like(y_true),
0.5 * tf.square(y_true - y_pred) * (tf.exp(-y_pred_corrected))
+ 0.5 * y_pred_corrected,
)
losses = tf.reduce_mean(wrapper_output, axis=-1) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
[docs]
def mean_absolute_error(y_true, y_pred, sample_weight=None):
"""
Calculate mean absolute error, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Mean Absolute Error
:rtype: tf.Tensor
:History: 2018-Jan-14 - Written - Henry Leung (University of Toronto)
"""
losses = tf.reduce_mean(
tf.where(
magic_num_check(y_true), tf.zeros_like(y_true), tf.abs(y_true - y_pred)
),
axis=-1,
) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
[docs]
def mean_absolute_percentage_error(y_true, y_pred, sample_weight=None):
"""
Calculate mean absolute percentage error, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:return: Mean Absolute Percentage Error
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:rtype: tf.Tensor
:History: 2018-Feb-17 - Written - Henry Leung (University of Toronto)
"""
tf_inf = tf.cast(tf.constant(1) / tf.constant(0), tf.float32)
epsilon_tensor = tf.cast(tf.constant(tfk.backend.epsilon()), tf.float32)
diff = tf.abs(
(y_true - y_pred) / tf.clip_by_value(tf.abs(y_true), epsilon_tensor, tf_inf)
)
diff_corrected = tf.where(magic_num_check(y_true), tf.zeros_like(y_true), diff)
losses = (
100.0 * tf.reduce_mean(diff_corrected, axis=-1) * magic_correction_term(y_true)
)
return weighted_loss(losses, sample_weight)
def median_absolute_percentage_error(y_true, y_pred, sample_weight=None):
"""
Calculate median absolute percentage error, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:return: Median Absolute Percentage Error
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:rtype: tf.Tensor
:History: 2020-Aug-13 - Written - Henry Leung (University of Toronto)
"""
tf_inf = tf.cast(tf.constant(1) / tf.constant(0), tf.float32)
epsilon_tensor = tf.cast(tf.constant(tfk.backend.epsilon()), tf.float32)
diff = tf.abs(
(y_true - y_pred) / tf.clip_by_value(tf.abs(y_true), epsilon_tensor, tf_inf)
)
diff_corrected = tf.where(magic_num_check(y_true), tf.zeros_like(y_true), diff)
losses = 100.0 * median(diff_corrected, axis=None) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
[docs]
def mean_squared_logarithmic_error(y_true, y_pred, sample_weight=None):
"""
Calculate mean squared logarithmic error, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Mean Squared Logarithmic Error
:rtype: tf.Tensor
:History: 2018-Feb-17 - Written - Henry Leung (University of Toronto)
"""
tf_inf = tf.cast(tf.constant(1) / tf.constant(0), tf.float32)
epsilon_tensor = tf.cast(tf.constant(tfk.backend.epsilon()), tf.float32)
first_log = tf.math.log(tf.clip_by_value(y_pred, epsilon_tensor, tf_inf) + 1.0)
second_log = tf.math.log(tf.clip_by_value(y_true, epsilon_tensor, tf_inf) + 1.0)
log_diff = tf.where(
magic_num_check(y_true),
tf.zeros_like(y_true),
tf.square(first_log - second_log),
)
losses = tf.reduce_mean(log_diff, axis=-1) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
[docs]
def mean_error(y_true, y_pred, sample_weight=None):
"""
Calculate mean error as a way to get the bias in prediction, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Mean Error
:rtype: tf.Tensor
:History: 2018-May-22 - Written - Henry Leung (University of Toronto)
"""
losses = tf.reduce_mean(
tf.where(magic_num_check(y_true), tf.zeros_like(y_true), y_true - y_pred),
axis=-1,
) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
[docs]
def mean_percentage_error(y_true, y_pred, sample_weight=None):
"""
Calculate mean percentage error, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Mean Percentage Error
:rtype: tf.Tensor
:History: 2018-Jun-06 - Written - Henry Leung (University of Toronto)
"""
tf_inf = tf.cast(tf.constant(1) / tf.constant(0), tf.float32)
epsilon_tensor = tf.cast(tf.constant(tfk.backend.epsilon()), tf.float32)
diff = y_true - y_pred / tf.clip_by_value(y_true, epsilon_tensor, tf_inf)
diff_corrected = tf.where(magic_num_check(y_true), tf.zeros_like(y_true), diff)
losses = (
100.0 * tf.reduce_mean(diff_corrected, axis=-1) * magic_correction_term(y_true)
)
return weighted_loss(losses, sample_weight)
def median_percentage_error(y_true, y_pred, sample_weight=None):
"""
Calculate median percentage error, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Median Percentage Error
:rtype: tf.Tensor
:History: 2020-Aug-13 - Written - Henry Leung (University of Toronto)
"""
tf_inf = tf.cast(tf.constant(1) / tf.constant(0), tf.float32)
epsilon_tensor = tf.cast(tf.constant(tfk.backend.epsilon()), tf.float32)
diff = y_true - y_pred / tf.clip_by_value(y_true, epsilon_tensor, tf_inf)
diff_corrected = tf.where(magic_num_check(y_true), tf.zeros_like(y_true), diff)
losses = 100.0 * median(diff_corrected, axis=None) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
[docs]
def categorical_crossentropy(y_true, y_pred, sample_weight=None, from_logits=False):
"""
Categorical cross-entropy between an output tensor and a target tensor, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:param from_logits: From logits space or not. If you want to use logits, please use from_logits=True
:type from_logits: boolean
:return: Categorical Cross-Entropy
:rtype: tf.Tensor
:History: 2018-Jan-14 - Written - Henry Leung (University of Toronto)
"""
# calculate correction term first
correction = magic_correction_term(y_true)
# Deal with magic number
y_true = tf.where(magic_num_check(y_true), tf.zeros_like(y_true), y_true)
# Note: tf.nn.softmax_cross_entropy_with_logits expects logits, we expects probabilities by default.
if not from_logits:
epsilon_tensor = tf.cast(tf.constant(tfk.backend.epsilon()), tf.float32)
# scale preds so that the class probas of each sample sum to 1
y_pred /= tf.reduce_sum(y_pred, len(y_pred.get_shape()) - 1, True)
# manual computation of crossentropy
y_pred = tf.clip_by_value(y_pred, epsilon_tensor, 1.0 - epsilon_tensor)
losses = (
-tf.reduce_sum(y_true * tf.math.log(y_pred), len(y_pred.get_shape()) - 1)
* correction
)
return weighted_loss(losses, sample_weight)
else:
losses = (
tf.nn.softmax_cross_entropy_with_logits(labels=y_true, logits=y_pred)
* correction
)
return weighted_loss(losses, sample_weight)
[docs]
def binary_crossentropy(y_true, y_pred, sample_weight=None, from_logits=False):
"""
Binary cross-entropy between an output tensor and a target tensor, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param from_logits: From logits space or not. If you want to use logits, please use from_logits=True
:type from_logits: boolean
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Binary Cross-Entropy
:rtype: tf.Tensor
:History: 2018-Jan-14 - Written - Henry Leung (University of Toronto)
"""
# Note: tf.nn.sigmoid_cross_entropy_with_logits expects logits, we expects probabilities by default.
if not from_logits:
epsilon_tensor = tf.cast(tf.constant(tfk.backend.epsilon()), tf.float32)
# transform back to logits
y_pred = tf.clip_by_value(y_pred, epsilon_tensor, 1.0 - epsilon_tensor)
y_pred = tf.math.log(y_pred / (1.0 - y_pred))
cross_entropy = tf.nn.sigmoid_cross_entropy_with_logits(
labels=y_true, logits=y_pred
)
corrected_cross_entropy = tf.where(
magic_num_check(y_true), tf.zeros_like(cross_entropy), cross_entropy
)
losses = tf.reduce_mean(corrected_cross_entropy, axis=-1) * magic_correction_term(
y_true
)
return weighted_loss(losses, sample_weight)
[docs]
def bayesian_categorical_crossentropy_wrapper(logit_var):
"""
| Categorical crossentropy between an output tensor and a target tensor for Bayesian Neural Network
| equation (12) of arxiv:1703.04977
:param logit_var: Predictive variance
:type logit_var: Union(tf.Tensor, tf.Variable)
:return: Robust categorical_crossentropy function for predictive variance neurones which matches Keras losses API
:rtype: function
:Returned Function Parameter:
| **function(y_true, y_pred)**
| - **y_true** (*tf.Tensor*): Ground Truth
| - **y_pred** (*tf.Tensor*): Prediction in logits space
| Return (*tf.Tensor*): Robust categorical crossentropy
:History: 2018-Mar-15 - Written - Henry Leung (University of Toronto)
"""
# y_pred is logits
def bayesian_crossentropy(y_true, y_pred, sample_weight=None):
return robust_categorical_crossentropy(y_true, y_pred, logit_var, sample_weight)
# set the name to be the same as parent so it can be found
bayesian_crossentropy.__name__ = "bayesian_categorical_crossentropy_wrapper"
return bayesian_crossentropy
[docs]
def bayesian_categorical_crossentropy_var_wrapper(logits):
"""
| Categorical crossentropy between an output tensor and a target tensor for Bayesian Neural Network
| equation (12) of arxiv:1703.04977
:param logits: Prediction in logits space
:type logits: Union(tf.Tensor, tf.Variable)
:return: Robust categorical_crossentropy function for predictive variance neurones which matches Keras losses API
:rtype: function
:Returned Function Parameter:
| **function(y_true, y_pred)**
| - **y_true** (*tf.Tensor*): Ground Truth
| - **y_pred** (*tf.Tensor*): Predictive variance in logits space
| Return (*tf.Tensor*): Robust categorical crossentropy
:History: 2018-Mar-15 - Written - Henry Leung (University of Toronto)
"""
# y_pred is predictive entropy
def bayesian_crossentropy(y_true, y_pred, sample_weight=None):
return robust_categorical_crossentropy(y_true, logits, y_pred, sample_weight)
# set the name to be the same as parent so it can be found
bayesian_crossentropy.__name__ = "bayesian_categorical_crossentropy_var_wrapper"
return bayesian_crossentropy
[docs]
def robust_categorical_crossentropy(y_true, y_pred, logit_var, sample_weight):
"""
Calculate categorical accuracy, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction in logits space
:type y_pred: Union(tf.Tensor, tf.Variable)
:param logit_var: Predictive variance in logits space
:type logit_var: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: categorical cross-entropy
:rtype: tf.Tensor
:History: 2018-Mar-15 - Written - Henry Leung (University of Toronto)
"""
variance_depressor = tf.reduce_mean(tf.exp(logit_var) - tf.ones_like(logit_var))
undistorted_loss = categorical_crossentropy(
y_true, y_pred, sample_weight, from_logits=True
)
mc_num = 25
batch_size = tf.shape(y_pred)[0]
label_size = tf.shape(y_pred)[-1]
dist = tf.random.normal(shape=[mc_num, batch_size, label_size], mean=y_pred, stddev=logit_var)
mc_result = -tf.nn.elu(
tf.tile(undistorted_loss, [mc_num])
- categorical_crossentropy(
tf.tile(y_true, [mc_num, 1]),
tf.reshape(dist, (batch_size * mc_num, label_size)),
from_logits=True,
)
)
variance_loss = (
tf.reduce_mean(tf.reshape(mc_result, (mc_num, batch_size)), axis=0)
* undistorted_loss
)
losses = (
variance_loss + undistorted_loss + variance_depressor
) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
[docs]
def bayesian_binary_crossentropy_wrapper(logit_var):
"""
| Binary crossentropy between an output tensor and a target tensor for Bayesian Neural Network
| equation (12) of arxiv:1703.04977
:param logit_var: Predictive variance
:type logit_var: Union(tf.Tensor, tf.Variable)
:return: Robust binary_crossentropy function for predictive variance neurones which matches Keras losses API
:rtype: function
:Returned Function Parameter:
| **function(y_true, y_pred)**
| - **y_true** (*tf.Tensor*): Ground Truth
| - **y_pred** (*tf.Tensor*): Prediction in logits space
| Return (*tf.Tensor*): Robust binary crossentropy
:History: 2018-Mar-15 - Written - Henry Leung (University of Toronto)
"""
# y_pred is logits
def bayesian_crossentropy(y_true, y_pred, sample_weight=None):
return robust_binary_crossentropy(y_true, y_pred, logit_var, sample_weight)
# set the name to be the same as parent so it can be found
bayesian_crossentropy.__name__ = "bayesian_binary_crossentropy_wrapper"
return bayesian_crossentropy
[docs]
def bayesian_binary_crossentropy_var_wrapper(logits):
"""
| Binary crossentropy between an output tensor and a target tensor for Bayesian Neural Network
| equation (12) of arxiv:1703.04977
:param logits: Prediction in logits space
:type logits: Union(tf.Tensor, tf.Variable)
:return: Robust binary_crossentropy function for predictive variance neurones which matches Keras losses API
:rtype: function
:Returned Function Parameter:
| **function(y_true, y_pred)**
| - **y_true** (*tf.Tensor*): Ground Truth
| - **y_pred** (*tf.Tensor*): Predictive variance in logits space
| Return (*tf.Tensor*): Robust binary crossentropy
:History: 2018-Mar-15 - Written - Henry Leung (University of Toronto)
"""
# y_pred is predictive entropy
def bayesian_crossentropy(y_true, y_pred, sample_weight=None):
return robust_binary_crossentropy(y_true, logits, y_pred, sample_weight)
# set the name to be the same as parent so it can be found
bayesian_crossentropy.__name__ = "bayesian_binary_crossentropy_var_wrapper"
return bayesian_crossentropy
[docs]
def robust_binary_crossentropy(y_true, y_pred, logit_var, sample_weight):
"""
Calculate binary accuracy, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction in logits space
:type y_pred: Union(tf.Tensor, tf.Variable)
:param logit_var: Predictive variance in logits space
:type logit_var: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: categorical cross-entropy
:rtype: tf.Tensor
:History: 2018-Mar-15 - Written - Henry Leung (University of Toronto)
"""
variance_depressor = tf.reduce_mean(tf.exp(logit_var) - tf.ones_like(logit_var))
undistorted_loss = binary_crossentropy(y_true, y_pred, from_logits=True)
mc_num = 25
batch_size = tf.shape(y_pred)[0]
label_size = tf.shape(y_pred)[-1]
dist = tf.random.normal(shape=[mc_num, batch_size, label_size], mean=y_pred, stddev=logit_var)
mc_result = -tf.nn.elu(
tf.tile(undistorted_loss, [mc_num])
- binary_crossentropy(
tf.tile(y_true, [mc_num, 1]),
tf.reshape(dist, (batch_size * mc_num, label_size)),
from_logits=True,
)
)
variance_loss = (
tf.reduce_mean(tf.reshape(mc_result, (mc_num, batch_size)), axis=0)
* undistorted_loss
)
losses = (
variance_loss + undistorted_loss + variance_depressor
) * magic_correction_term(y_true)
return weighted_loss(losses, sample_weight)
def nll(y_true, y_pred, sample_weight=None):
"""
Calculate negative log likelihood
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Negative log likelihood
:rtype: tf.Tensor
:History: 2018-Jan-30 - Written - Henry Leung (University of Toronto)
"""
# astroNN binary_cross_entropy gives the mean over the last axis. we require the sum
losses = tf.reduce_sum(binary_crossentropy(y_true, y_pred), axis=-1)
return weighted_loss(losses, sample_weight)
[docs]
def categorical_accuracy(y_true, y_pred):
"""
Calculate categorical accuracy, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:return: Categorical Classification Accuracy
:rtype: tf.Tensor
:History: 2018-Jan-21 - Written - Henry Leung (University of Toronto)
"""
y_true = tf.where(magic_num_check(y_true), tf.zeros_like(y_true), y_true)
return tf.cast(
tf.equal(tf.argmax(y_true, axis=-1), tf.argmax(y_pred, axis=-1)), tf.float32
) * magic_correction_term(y_true)
def __binary_accuracy(from_logits=False):
"""
Calculate binary accuracy, ignoring the magic number
:param from_logits: From logits space or not. If you want to use logits, please use from_logits=True
:type from_logits: boolean
:return: Function for Binary classification accuracy which matches Keras losses API
:rtype: function
:Returned Funtion Parameter:
| **function(y_true, y_pred)**
| - **y_true** (*tf.Tensor*): Ground Truth
| - **y_pred** (*tf.Tensor*): Prediction
| Return (*tf.Tensor*): Binary Classification Accuracy
:History: 2018-Jan-31 - Written - Henry Leung (University of Toronto)
"""
# DO NOT correct y_true for magic number, just let it goes wrong and then times a correction terms
def binary_accuracy_internal(y_true, y_pred):
if from_logits:
y_pred = tf.nn.sigmoid(y_pred)
return tf.reduce_mean(
tf.cast(tf.equal(y_true, tf.round(y_pred)), tf.float32), axis=-1
) * magic_correction_term(y_true)
if not from_logits:
binary_accuracy_internal.__name__ = (
"binary_accuracy" # set the name to be displayed in TF/Keras log
)
else:
binary_accuracy_internal.__name__ = "binary_accuracy_from_logits" # set the name to be displayed in TF/Keras log
return binary_accuracy_internal
[docs]
def binary_accuracy(*args, **kwargs):
"""
Calculate binary accuracy, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:return: Binary accuracy
:rtype: tf.Tensor
:History: 2018-Jan-31 - Written - Henry Leung (University of Toronto)
"""
return __binary_accuracy(from_logits=False)(*args, **kwargs)
def binary_accuracy_from_logits(*args, **kwargs):
"""
Calculate binary accuracy from logits, ignoring the magic number
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:return: Binary accuracy
:rtype: tf.Tensor
:History: 2018-Jan-31 - Written - Henry Leung (University of Toronto)
"""
return __binary_accuracy(from_logits=True)(*args, **kwargs)
[docs]
def zeros_loss(y_true, y_pred, sample_weight=None):
"""
Always return zeros
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param sample_weight: Sample weights
:type sample_weight: Union(tf.Tensor, tf.Variable, list)
:return: Zeros
:rtype: tf.Tensor
:History: 2018-May-24 - Written - Henry Leung (University of Toronto)
"""
losses = tf.reduce_mean(
tf.where(magic_num_check(y_true), tf.zeros_like(y_true), y_true) * 0.0
+ 0.0 * y_pred,
axis=-1,
)
return weighted_loss(losses, sample_weight)
def median_error(y_true, y_pred, sample_weight=None, axis=-1):
"""
Calculate median difference
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param axis: Axis
:type axis: int
:return: Variance
:rtype: tf.Tensor
:History: 2021-Aug-13 - Written - Henry Leung (University of Toronto)
"""
# tf.boolean_mask(tf.logical_not(magic_num_check(y_true))
return weighted_loss(median(y_true - y_pred, axis=axis), sample_weight)
def median_absolute_deviation(y_true, y_pred, sample_weight=None, axis=-1):
"""
Calculate median absilute difference
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param axis: Axis
:type axis: int
:return: Variance
:rtype: tf.Tensor
:History: 2021-Aug-13 - Written - Henry Leung (University of Toronto)
"""
return weighted_loss(median(tf.abs(y_true - y_pred), axis=axis), sample_weight)
def mad_std(y_true, y_pred, sample_weight=None, axis=-1):
"""
Calculate 1.4826 * median absilute difference
:param y_true: Ground Truth
:type y_true: Union(tf.Tensor, tf.Variable)
:param y_pred: Prediction
:type y_pred: Union(tf.Tensor, tf.Variable)
:param axis: Axis
:type axis: int
:return: Variance
:rtype: tf.Tensor
:History: 2021-Aug-13 - Written - Henry Leung (University of Toronto)
"""
return weighted_loss(
1.4826 * median_absolute_deviation(y_true, y_pred, axis=axis), sample_weight
)
# Just alias functions
mse = mean_squared_error
mae = mean_absolute_error
mape = mean_absolute_percentage_error
msle = mean_squared_logarithmic_error
me = mean_error
mpe = mean_percentage_error
mad = median_absolute_deviation
# legacy support
mse_lin = mse_lin_wrapper
mse_var = mse_var_wrapper
def losses_lookup(identifier):
"""
Lookup astroNN.nn.losses function by name
:param identifier: identifier
:type identifier: str
:return: Looked up function
:rtype: function
:History: 2018-Apr-28 - Written - Henry Leung (University of Toronto)
"""
return nn_obj_lookup(identifier, module_obj=globals(), module_name=__name__)