# tfr.keras.losses.GumbelApproxNDCGLoss

|---|
| [![](https://www.tensorflow.org/images/GitHub-Mark-32px.png) View source on GitHub](https://github.com/tensorflow/ranking/blob/v0.5.3/tensorflow_ranking/python/keras/losses.py#L1126-L1227) |

Computes the Gumbel approximate NDCG loss between `y_true` and `y_pred`.

Inherits From: [`ApproxNDCGLoss`](https://www.tensorflow.org/ranking/api_docs/python/tfr/keras/losses/ApproxNDCGLoss)

    tfr.keras.losses.GumbelApproxNDCGLoss(
        reduction: tf.losses.Reduction = tf.losses.Reduction.AUTO,
        name: Optional[str] = None,
        lambda_weight: Optional[losses_impl._LambdaWeight] = None,
        temperature: float = 0.1,
        sample_size: int = 8,
        gumbel_temperature: float = 1.0,
        seed: Optional[int] = None,
        ragged: bool = False
    )

Implementation of Gumbel ApproxNDCG loss ([Bruch et al, 2020](https://research.google/pubs/pub48689/)).
This loss is the same as [`tfr.keras.losses.ApproxNDCGLoss`](https://www.tensorflow.org/ranking/api_docs/python/tfr/keras/losses/ApproxNDCGLoss) but where logits
are sampled from the Gumbel distribution:

`y_new_pred ~ Gumbel(y_pred, 1 / temperature)`

> [!NOTE]
> **Note:** This loss is stochastic and may return different values for identical inputs.

#### Standalone usage:

    tf.random.set_seed(42)
    y_true = [[1., 0.]]
    y_pred = [[0.6, 0.8]]
    loss = tfr.keras.losses.GumbelApproxNDCGLoss(seed=42)
    loss(y_true, y_pred).numpy()
    -0.8160851

    # Using a higher gumbel temperature
    loss = tfr.keras.losses.GumbelApproxNDCGLoss(gumbel_temperature=2.0,
        seed=42)
    loss(y_true, y_pred).numpy()
    -0.7583889

    # 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.GumbelApproxNDCGLoss(seed=42, ragged=True)
    loss(y_true, y_pred).numpy()
    -0.6987189

Usage with the `compile()` API:

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

#### Definition:

\\\[\\mathcal{L}(\\{y\\}, \\{s\\}) = \\text{ApproxNDCGLoss}(\\{y\\}, \\{z\\})\\\]

where

\\\[
z \\sim \\text{Gumbel}(s, \\beta)\\\\
p(z) = e\^{-t-e\^{-t} }\\\\
t = \\beta(z - s)\\\\
\\beta = \\frac{1}{\\text{temperature} }
\\\]

<br />

<br />

| ## References ||
|---|---|
| <br /> - [A Stochastic Treatment of Learning to Rank Scoring Functions, Bruch et al, 2020](https://research.google/pubs/pub48689/) ||

<br />

<br />

<br />

<br />

| ## Args ||
|---|---|
| `reduction` | Type of [`tf.keras.losses.Reduction`](https://www.tensorflow.org/api_docs/python/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 under a [`tf.distribute.Strategy`](https://www.tensorflow.org/api_docs/python/tf/distribute/Strategy), except via [`Model.compile()`](https://www.tensorflow.org/api_docs/python/tf/keras/Model#compile) and [`Model.fit()`](https://www.tensorflow.org/api_docs/python/tf/keras/Model#fit), using `AUTO` or `SUM_OVER_BATCH_SIZE` will raise an error. Please see this custom training [tutorial](https://www.tensorflow.org/tutorials/distribute/custom_training) for more details. |
| `name` | Optional name for the instance. |

<br />

## Methods

### `from_config`

[View source](https://github.com/tensorflow/ranking/blob/v0.5.3/tensorflow_ranking/python/keras/losses.py#L742-L752)

    @classmethod
    from_config(
        config, custom_objects=None
    )

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

<br />

<br />

<br />

| Args ||
|---|---|
| `config` | Output of `get_config()`. |

<br />

<br />

<br />

<br />

| Returns ||
|---|---|
| A `Loss` instance. ||

<br />

### `get_config`

[View source](https://github.com/tensorflow/ranking/blob/v0.5.3/tensorflow_ranking/python/keras/losses.py#L1209-L1216)

    get_config() -> Dict[str, Any]

Returns the config dictionary for a `Loss` instance.

### `__call__`

[View source](https://github.com/tensorflow/ranking/blob/v0.5.3/tensorflow_ranking/python/keras/losses.py#L1218-L1227)

    __call__(
        y_true: https://www.tensorflow.org/ranking/api_docs/python/tfr/keras/model/TensorLike,
        y_pred: https://www.tensorflow.org/ranking/api_docs/python/tfr/keras/model/TensorLike,
        sample_weight: Optional[utils.TensorLike] = None
    ) -> tf.Tensor

See _RankingLoss.