Update docs

docs/build/html/_sources/autoapi/pyepo/data/index.rst.txt

@@ -17,6 +17,7 @@ Submodules
 
  dataset/index.rst
  knapsack/index.rst
+ portfolio/index.rst
  shortestpath/index.rst
  tsp/index.rst
 

docs/build/html/_sources/autoapi/pyepo/func/abcmodule/index.rst.txt

@@ -22,18 +22,23 @@ Classes
 
 
 
-.. py:class:: optModule(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: optModule(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`torch.nn.Module`
 
  An abstract module for the learning to rank losses, which measure the difference in how the predicted cost
  vector and the true cost vector rank a pool of feasible solutions.
 
- .. py:method:: forward(pred_cost, true_cost, reduction='mean')
+ .. py:method:: forward(pred_cost, true_cost)
  :abstractmethod:
 
  Forward pass
 
 
+ .. py:method:: _update_solution_pool(sol)
+
+ Add new solutions to solution pool
+
+
 

docs/build/html/_sources/autoapi/pyepo/func/blackbox/index.rst.txt

@@ -56,23 +56,13 @@ Classes
 
  A autograd function for differentiable black-box optimizer
 
- .. py:method:: forward(ctx, pred_cost, lambd, optmodel, processes, pool, solve_ratio, module)
+ .. py:method:: forward(ctx, pred_cost, module)
  :staticmethod:
 
  Forward pass for DBB
 
  :param pred_cost: a batch of predicted values of the cost
  :type pred_cost: torch.tensor
- :param lambd: a hyperparameter for differentiable block-box to control interpolation degree
- :type lambd: float
- :param optmodel: an PyEPO optimization model
- :type optmodel: optModel
- :param processes: number of processors, 1 for single-core, 0 for all of cores
- :type processes: int
- :param pool: process pool object
- :type pool: ProcessPool
- :param solve_ratio: the ratio of new solutions computed during training
- :type solve_ratio: float
  :param module: blackboxOpt module
  :type module: optModule
 
@@ -119,21 +109,13 @@ Classes
 
  A autograd function for differentiable black-box optimizer
 
- .. py:method:: forward(ctx, pred_cost, optmodel, processes, pool, solve_ratio, module)
+ .. py:method:: forward(ctx, pred_cost, module)
  :staticmethod:
 
  Forward pass for NID
 
  :param pred_cost: a batch of predicted values of the cost
  :type pred_cost: torch.tensor
- :param optmodel: an PyEPO optimization model
- :type optmodel: optModel
- :param processes: number of processors, 1 for single-core, 0 for all of cores
- :type processes: int
- :param pool: process pool object
- :type pool: ProcessPool
- :param solve_ratio: the ratio of new solutions computed during training
- :type solve_ratio: float
  :param module: blackboxOpt module
  :type module: optModule
 

docs/build/html/_sources/autoapi/pyepo/func/contrastive/index.rst.txt

@@ -23,7 +23,7 @@ Classes
 
 
 
-.. py:class:: NCE(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: NCE(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -38,13 +38,13 @@ Classes
 
  Reference: <https://www.ijcai.org/proceedings/2021/390>
 
- .. py:method:: forward(pred_cost, true_sol, reduction='mean')
+ .. py:method:: forward(pred_cost, true_sol)
 
  Forward pass
 
 
 
-.. py:class:: contrastiveMAP(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: contrastiveMAP(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -59,7 +59,7 @@ Classes
 
  Reference: <https://www.ijcai.org/proceedings/2021/390>
 
- .. py:method:: forward(pred_cost, true_sol, reduction='mean')
+ .. py:method:: forward(pred_cost, true_sol)
 
  Forward pass
 

docs/build/html/_sources/autoapi/pyepo/func/index.rst.txt

@@ -38,6 +38,7 @@ Classes
  pyepo.func.perturbedOpt
  pyepo.func.perturbedFenchelYoung
  pyepo.func.implicitMLE
+ pyepo.func.adaptiveImplicitMLE
  pyepo.func.NCE
  pyepo.func.contrastiveMAP
  pyepo.func.listwiseLTR
@@ -47,7 +48,7 @@ Classes
 
 
 
-.. py:class:: SPOPlus(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: SPOPlus(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -64,7 +65,7 @@ Classes
 
  Reference: <https://doi.org/10.1287/mnsc.2020.3922>
 
- .. py:method:: forward(pred_cost, true_cost, true_sol, true_obj, reduction='mean')
+ .. py:method:: forward(pred_cost, true_cost, true_sol, true_obj)
 
  Forward pass
 
@@ -142,7 +143,7 @@ Classes
 
 
 
-.. py:class:: perturbedFenchelYoung(optmodel, n_samples=10, sigma=1.0, processes=1, seed=135, solve_ratio=1, dataset=None)
+.. py:class:: perturbedFenchelYoung(optmodel, n_samples=10, sigma=1.0, processes=1, seed=135, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -160,13 +161,13 @@ Classes
 
  Reference: <https://papers.nips.cc/paper/2020/hash/6bb56208f672af0dd65451f869fedfd9-Abstract.html>
 
- .. py:method:: forward(pred_cost, true_sol, reduction='mean')
+ .. py:method:: forward(pred_cost, true_sol)
 
  Forward pass
 
 
 
-.. py:class:: implicitMLE(optmodel, n_samples=10, sigma=1.0, lambd=10, processes=1, distribution=sumGammaDistribution(kappa=5), solve_ratio=1, dataset=None)
+.. py:class:: implicitMLE(optmodel, n_samples=10, sigma=1.0, lambd=10, distribution=sumGammaDistribution(kappa=5), two_sides=False, processes=1, solve_ratio=1, dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -175,11 +176,11 @@ Classes
  an optimal solution in a constrained exponential family distribution via
  Perturb-and-MAP.
 
- For I-LME, it works as black-box combinatorial solvers, in which constraints
+ For I-MLE, it works as black-box combinatorial solvers, in which constraints
  are known and fixed, but the cost vector need to be predicted from
  contextual data.
 
- The I-LME approximate gradient of optimizer smoothly. Thus, allows us to
+ The I-MLE approximate gradient of optimizer smoothly. Thus, allows us to
  design an algorithm based on stochastic gradient descent.
 
  Reference: <https://proceedings.neurips.cc/paper_files/paper/2021/hash/7a430339c10c642c4b2251756fd1b484-Abstract.html>
@@ -190,7 +191,31 @@ Classes
 
 
 
-.. py:class:: NCE(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: adaptiveImplicitMLE(optmodel, n_samples=10, sigma=1.0, distribution=sumGammaDistribution(kappa=5), two_sides=False, processes=1, solve_ratio=1, dataset=None)
+
+
+ Bases: :py:obj:`pyepo.func.abcmodule.optModule`
+
+ An autograd module for Adaptive Implicit Maximum Likelihood Estimator, which
+ adaptively choose hyperparameter λ and yield an optimal solution in a
+ constrained exponential family distribution via Perturb-and-MAP.
+
+ For AI-MLE, it works as black-box combinatorial solvers, in which constraints
+ are known and fixed, but the cost vector need to be predicted from
+ contextual data.
+
+ The AI-MLE approximate gradient of optimizer smoothly. Thus, allows us to
+ design an algorithm based on stochastic gradient descent.
+
+ Reference: <https://ojs.aaai.org/index.php/AAAI/article/view/26103>
+
+ .. py:method:: forward(pred_cost)
+
+ Forward pass
+
+
+
+.. py:class:: NCE(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -205,13 +230,13 @@ Classes
 
  Reference: <https://www.ijcai.org/proceedings/2021/390>
 
- .. py:method:: forward(pred_cost, true_sol, reduction='mean')
+ .. py:method:: forward(pred_cost, true_sol)
 
  Forward pass
 
 
 
-.. py:class:: contrastiveMAP(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: contrastiveMAP(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -226,13 +251,13 @@ Classes
 
  Reference: <https://www.ijcai.org/proceedings/2021/390>
 
- .. py:method:: forward(pred_cost, true_sol, reduction='mean')
+ .. py:method:: forward(pred_cost, true_sol)
 
  Forward pass
 
 
 
-.. py:class:: listwiseLTR(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: listwiseLTR(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -248,13 +273,13 @@ Classes
 
  Reference: <https://proceedings.mlr.press/v162/mandi22a.html>
 
- .. py:method:: forward(pred_cost, true_cost, reduction='mean')
+ .. py:method:: forward(pred_cost, true_cost)
 
  Forward pass
 
 
 
-.. py:class:: pairwiseLTR(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: pairwiseLTR(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -270,13 +295,13 @@ Classes
 
  Reference: <https://proceedings.mlr.press/v162/mandi22a.html>
 
- .. py:method:: forward(pred_cost, true_cost, reduction='mean')
+ .. py:method:: forward(pred_cost, true_cost)
 
  Forward pass
 
 
 
-.. py:class:: pointwiseLTR(optmodel, processes=1, solve_ratio=1, dataset=None)
+.. py:class:: pointwiseLTR(optmodel, processes=1, solve_ratio=1, reduction='mean', dataset=None)
 
 
  Bases: :py:obj:`pyepo.func.abcmodule.optModule`
@@ -293,7 +318,7 @@ Classes
 
  Reference: <https://proceedings.mlr.press/v162/mandi22a.html>
 
- .. py:method:: forward(pred_cost, true_cost, reduction='mean')
+ .. py:method:: forward(pred_cost, true_cost)
 
  Forward pass