Help protect the Great Barrier Reef with TensorFlow on Kaggle Join Challenge

tfr.keras.losses.ApproxMRRLoss

Computes approximate MRR loss between y_true and y_pred.

Implementation of ApproxMRR loss (Qin et al, 2008). This loss is an approximation for tfr.keras.metrics.MRRMetric. It replaces the non-differentiable ranking function in MRR with a differentiable approximation based on the logistic function.

For each list of scores s in y_pred and list of labels y in y_true:

loss = sum_i (1 / approxrank(s_i)) * y_i
approxrank(s_i) = 1 + sum_j (1 / (1 + exp(-(s_j - s_i) / temperature)))

Standalone usage:

y_true = [[1., 0.]]
y_pred = [[0.6, 0.8]]
loss = tfr.keras.losses.ApproxMRRLoss()
loss(y_true, y_pred).numpy()
-0.53168947
# Using ragged tensors
y_true = tf.ragged.constant([[1., 0.], [0., 1., 0.]])
y_pred = tf.ragged.constant([[0.6, 0.8], [0.5, 0.8, 0.4]])
loss = tfr.keras.losses.ApproxMRRLoss(ragged=True)
loss(y_true, y_pred).numpy()
-0.73514676

Usage with the compile() API:

model.compile(optimizer='sgd', loss=tfr.keras.losses.ApproxMRRLoss())

Definition:

\[ \mathcal{L}(\{y\}, \{s\}) = -\sum_{i} \frac{1}{\text{approxrank}_i} y_i \]

where:

\[ \text{approxrank}_i = 1 + \sum_{j \neq i} \frac{1}{1 + \exp\left(\frac{-(s_j - s_i)}{\text{temperature} }\right)} \]

References:

reduction Type of tf.keras.losses.Reduction to apply to loss. Default value is AUTO. AUTO indicates that the reduction option will be determined by the usage context. For almost all cases this defaults to SUM_OVER_BATCH_SIZE. When used with tf.distribute.Strategy, outside of built-in training loops such as tf.keras compile and fit, using AUTO or SUM_OVER_BATCH_SIZE will raise an error. Please see this custom training tutorial for more details.
name Optional name for the instance.

Methods

from_config

View source

Instantiates a Loss from its config (output of get_config()).

Args
config Output of get_config().

Returns
A Loss instance.

get_config

View source

Returns the config dictionary for a Loss instance.

__call__

View source

See tf.keras.losses.Loss.