pyepo.func.rank
===============

.. py:module:: pyepo.func.rank

.. autoapi-nested-parse::

   Learning to rank Losses



Attributes
----------

.. autoapisummary::

   pyepo.func.rank.lsLTR
   pyepo.func.rank.prLTR
   pyepo.func.rank.ptLTR


Classes
-------

.. autoapisummary::

   pyepo.func.rank.listwiseLearningToRank
   pyepo.func.rank.pairwiseLearningToRank
   pyepo.func.rank.pointwiseLearningToRank


Module Contents
---------------

.. py:class:: listwiseLearningToRank(optmodel: pyepo.model.opt.optModel, processes: int = 1, solve_ratio: float = 1.0, reduction: pyepo.func.abcmodule.Reduction = 'mean', dataset: pyepo.data.dataset.optDataset | None = None)

   Bases: :py:obj:`pyepo.func.abcmodule.optModule`


   Listwise Learning-to-Rank loss over a cached solution pool.

   Models the ranking distribution over the cached pool :math:`\Gamma` as
   SoftMax of predicted-cost scores and minimizes its cross-entropy against
   the true ranking distribution. The full-list formulation captures
   interactions between every pair of solutions in :math:`\Gamma`.
   Pool semantics (``solve_ratio``, ``dataset``) are shared with the other
   LTR variants and with the contrastive methods.

   Reference: Mandi et al. (2022)
   `<https://proceedings.mlr.press/v162/mandi22a.html>`_


   .. py:method:: forward(pred_cost: torch.Tensor, true_cost: torch.Tensor) -> torch.Tensor

      Forward pass



.. py:class:: pairwiseLearningToRank(optmodel: pyepo.model.opt.optModel, processes: int = 1, solve_ratio: float = 1.0, reduction: pyepo.func.abcmodule.Reduction = 'mean', dataset: pyepo.data.dataset.optDataset | None = None)

   Bases: :py:obj:`pyepo.func.abcmodule.optModule`


   Pairwise Learning-to-Rank loss over a cached solution pool.

   Enforces a margin between the true optimum (the best member of
   :math:`\Gamma`) and each suboptimal solution via a ReLU hinge on the
   predicted-cost difference. Lighter than the listwise variant
   (no SoftMax over the full pool) and often a good first choice when the
   pool is large. Pool semantics (``solve_ratio``, ``dataset``) are shared
   with the other LTR variants and with the contrastive methods.

   Reference: Mandi et al. (2022)
   `<https://proceedings.mlr.press/v162/mandi22a.html>`_


   .. py:attribute:: relu


   .. py:method:: forward(pred_cost: torch.Tensor, true_cost: torch.Tensor) -> torch.Tensor

      Forward pass



.. py:class:: pointwiseLearningToRank(optmodel: pyepo.model.opt.optModel, processes: int = 1, solve_ratio: float = 1.0, reduction: pyepo.func.abcmodule.Reduction = 'mean', dataset: pyepo.data.dataset.optDataset | None = None)

   Bases: :py:obj:`pyepo.func.abcmodule.optModule`


   Pointwise Learning-to-Rank loss over a cached solution pool.

   Treats each cached solution :math:`\mathbf{w} \in \Gamma` as an
   independent regression target: the predicted score
   :math:`\hat{\mathbf{c}}^\top \mathbf{w}` is fit toward the true
   score :math:`\mathbf{c}^\top \mathbf{w}` via squared error, averaged
   over the pool. Cheapest of the three LTR variants -- no cross-pool
   interactions. Pool semantics (``solve_ratio``, ``dataset``) are shared
   with the other LTR variants and with the contrastive methods.

   Reference: Mandi et al. (2022)
   `<https://proceedings.mlr.press/v162/mandi22a.html>`_


   .. py:method:: forward(pred_cost: torch.Tensor, true_cost: torch.Tensor) -> torch.Tensor

      Forward pass



.. py:data:: lsLTR

.. py:data:: prLTR

.. py:data:: ptLTR