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

tfa.losses.pinball_loss

Computes the pinball loss between y_true and y_pred.

@tf.function
tfa.losses.pinball_loss(
    y_true: tfa.types.TensorLike,
    y_pred: tfa.types.TensorLike,
    tau: tfa.types.FloatTensorLike = 0.5
) -> tf.Tensor

loss = maximum(tau * (y_true - y_pred), (tau - 1) * (y_true - y_pred))

In the context of regression this loss yields an estimator of the tau conditional quantile.

See: https://en.wikipedia.org/wiki/Quantile_regression

Usage:

loss = tfa.losses.pinball_loss([0., 0., 1., 1.],
[1., 1., 1., 0.], tau=.1)
loss
<tf.Tensor: shape=(), dtype=float32, numpy=0.475>

y_true Ground truth values. shape = [batch_size, d0, .. dN]
y_pred The predicted values. shape = [batch_size, d0, .. dN]
tau (Optional) Float in [0, 1] or a tensor taking values in [0, 1] and shape = [d0,..., dn]. It defines the slope of the pinball loss. In the context of quantile regression, the value of tau determines the conditional quantile level. When tau = 0.5, this amounts to l1 regression, an estimator of the conditional median (0.5 quantile).

pinball_loss 1-D float Tensor with shape [batch_size].

References: