AIS_Kriging_S contrib method fails

[RomainGoussault]

AIS_Kriging_S contrib method fails.
It now removed from the end to end tests.

How to reproduce:

    scenario = Scenario(2, [0.25, 0.75], epoch_count=4, minibatch_count=2, dataset_name='mnist',
                                    contributivity_methods=["AIS_Kriging_S"], dataset_proportion=0.1)
    exp = Experiment(scenarios_list=[scenario])
    exp.run()

Stacktrace:

      ---------------------------------------------------------------------------
      LinAlgError                               Traceback (most recent call last)
      <ipython-input-5-8d35b47f50db> in <module>
            2                                 contributivity_methods=["AIS_Kriging_S"], dataset_proportion=0.1)
            3 exp = Experiment(scenarios_list=[scenario])
    ----> 4 exp.run()
    
    ~/code/distributed-learning-contributivity/mplc/experiment.py in run(self)
        162                 else:
        163                     scenario = blank_scenario.copy(repeat_count=repeat_idx)
    --> 164                 scenario.run()
        165 
        166                 # Save scenario results
    
    ~/code/distributed-learning-contributivity/mplc/scenario.py in run(self)
        572             logger.info(f"{method}")
        573             contrib = contributivity.Contributivity(scenario=self)
    --> 574             contrib.compute_contributivity(method)
        575             self.append_contributivity(contrib)
        576             logger.info(f"Evaluating contributivity with {method}: {contrib}")
    
    ~/code/distributed-learning-contributivity/mplc/contributivity.py in compute_contributivity(self, method_to_compute, sv_accuracy, alpha, truncation, update)
       1165         elif method_to_compute == "AIS_Kriging_S":
       1166             # Contributivity 7: Adaptative importance sampling with Kriging model
    -> 1167             self.AIS_Kriging(sv_accuracy=sv_accuracy, alpha=alpha, update=update)
       1168         elif method_to_compute == "SMCS":
       1169             # Contributivity 8:  Stratified Monte Carlo
    
    ~/code/distributed-learning-contributivity/mplc/contributivity.py in AIS_Kriging(self, sv_accuracy, alpha, update)
        669             if t % update == 0:  # renew the importance density g
        670                 j = t // update
    --> 671                 make_models()
        672                 # ## compute the renormalization constant of the new importance density for all datatsets
        673                 renorms = []
    
    ~/code/distributed-learning-contributivity/mplc/contributivity.py in make_models()
        642             for k in range(n):
        643                 model_k = KrigingModel(2, cov[k])
    --> 644                 model_k.fit(datasets[k], outputs[k])
        645                 models.append(model_k)
        646             all_models.append(models)
    
    ~/code/distributed-learning-contributivity/mplc/contributivity.py in fit(self, X, Y)
         45         self.invK = np.linalg.inv(K)
         46         Ht_invK_H = H.transpose().dot(self.invK).dot(H)
    ---> 47         self.beta = np.linalg.inv(Ht_invK_H).dot(H.transpose()).dot(self.invK).dot(self.Y)
         48 
         49     def predict(self, x):
    
    <__array_function__ internals> in inv(*args, **kwargs)
    
    ~/Library/Python/3.8/lib/python/site-packages/numpy/linalg/linalg.py in inv(a)
        544     signature = 'D->D' if isComplexType(t) else 'd->d'
        545     extobj = get_linalg_error_extobj(_raise_linalgerror_singular)
    --> 546     ainv = _umath_linalg.inv(a, signature=signature, extobj=extobj)
        547     return wrap(ainv.astype(result_t, copy=False))
        548 
    
    ~/Library/Python/3.8/lib/python/site-packages/numpy/linalg/linalg.py in _raise_linalgerror_singular(err, flag)
         86 
         87 def _raise_linalgerror_singular(err, flag):
    ---> 88     raise LinAlgError("Singular matrix")
         89 
         90 def _raise_linalgerror_nonposdef(err, flag):
    
    LinAlgError: Singular matrix

@bowni @arthurPignet @Thomas-Galtier

[RomainGoussault]

We have decided to remove this method.
To be done by @RomainGoussault