Feat/logeuclid mean and logeuclid distance to convex hull

Dockerfile

@@ -24,16 +24,16 @@ RUN mkdir /root/mne_data
 RUN mkdir /home/mne_data
 
 ## Workaround for firestore
-RUN pip install protobuf==4.25.2
-RUN pip install google_cloud_firestore==2.14.0
+RUN pip install protobuf==4.25.3
+RUN pip install google_cloud_firestore==2.15.0
 ### Missing __init__ file in protobuf
-RUN touch /usr/local/lib/python3.9/site-packages/protobuf-4.25.2-py3.9.egg/google/__init__.py
+RUN touch /usr/local/lib/python3.9/site-packages/protobuf-4.25.3-py3.9.egg/google/__init__.py
 ## google.cloud.location is never used in these files, and is missing in path.
-RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.14.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/client.py'
-RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.14.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/base.py'
-RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.14.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/grpc.py'
-RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.14.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/grpc_asyncio.py'
-RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.14.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/rest.py'
-RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.14.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/async_client.py'
+RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.15.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/client.py'
+RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.15.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/base.py'
+RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.15.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/grpc.py'
+RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.15.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/grpc_asyncio.py'
+RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.15.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/transports/rest.py'
+RUN sed -i 's/from google.cloud.location import locations_pb2//g' '/usr/local/lib/python3.9/site-packages/google_cloud_firestore-2.15.0-py3.9.egg/google/cloud/firestore_v1/services/firestore/async_client.py'
 
 ENTRYPOINT [ "python", "/examples/ERP/classify_P300_bi.py" ]

benchmarks/light_benchmark.py

@@ -11,7 +11,7 @@
 # Modified from plot_classify_P300_bi.py of pyRiemann
 # License: BSD (3-clause)
 
-from pyriemann.estimation import XdawnCovariances
+from pyriemann.estimation import XdawnCovariances, Shrinkage
 from pyriemann.tangentspace import TangentSpace
 from sklearn.pipeline import make_pipeline
 from sklearn.model_selection import train_test_split
@@ -78,15 +78,17 @@
 )
 
 pipelines["RG_VQC"] = QuantumClassifierWithDefaultRiemannianPipeline(
-    shots=100, spsa_trials=5, two_local_reps=2, params={"seed": 42}
+    shots=100, spsa_trials=1, two_local_reps=2, params={"seed": 42}
 )
 
 pipelines["QMDM_mean"] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="mean", quantum=True
+    metric={"mean": "euclid_cpm", "distance": "euclid"},
+    quantum=True,
+    regularization=Shrinkage(shrinkage=0.9),
 )
 
 pipelines["QMDM_dist"] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="distance", quantum=True
+    metric={"mean": "logeuclid", "distance": "logeuclid_cpm"}, quantum=True
 )
 
 pipelines["RG_LDA"] = make_pipeline(

doc/api.rst

@@ -96,7 +96,8 @@ Mean
 .. autosummary::
     :toctree: generated/
 
-    fro_mean_convex
+    mean_euclid_cpm
+    mean_logeuclid_cpm
 
 Distance
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -106,7 +107,7 @@ Distance
 .. autosummary::
     :toctree: generated/
 
-    logeucl_dist_convex
+    distance_logeuclid_cpm
 
 Docplex
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~

doc/requirements.txt

@@ -10,6 +10,7 @@ imbalanced-learn==0.11.0
 joblib
 pandas
 cvxpy==1.4.1
+qiskit==0.45.0
 qiskit_machine_learning==0.6.1
 qiskit-ibm-provider==0.7.3
 qiskit-optimization==0.5.0

examples/ERP/classify_P300_bi_illiteracy.py

@@ -172,19 +172,19 @@ def placeholder(key):
 placeholder(PIP.xDAWN_LDA.value)
 
 pipelines[PIP.ERPCov_CvxMDM_Dist.value] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="distance", quantum=False
+    metric="distance", quantum=False
 )
 placeholder(PIP.ERPCov_CvxMDM_Dist.value)
 
 # Quantum Pipelines
 
 pipelines[PIP.ERPCov_QMDM_Dist.value] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="distance", quantum=True
+    metric="distance", quantum=True
 )
 placeholder(PIP.ERPCov_QMDM_Dist.value)
 
 pipelines[PIP.ERPCov_QMDM_Dist.value] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="distance", quantum=True
+    metric="distance", quantum=True
 )
 placeholder(PIP.ERPCov_QMDM_Dist.value)
 

examples/ERP/classify_P300_bi_quantum_mdm.py

@@ -6,8 +6,8 @@
 The mean and the distance in MDM algorithm are formulated as
 optimization problems. These optimization problems are translated
 to Qiskit using Docplex and additional glue code. These optimizations
-are enabled when we use convex mean or convex distance. This is set
-using the 'convex_metric' parameter of the QuantumMDMWithRiemannianPipeline.
+are enabled when we use cpm mean or cpm distance. This is set
+using the 'metric' parameter of the QuantumMDMWithRiemannianPipeline.
 
 Classification can be run either on emulation or real quantum computer.
 
@@ -43,7 +43,7 @@
 from moabb.evaluations import WithinSessionEvaluation
 from moabb.paradigms import P300
 
-# inject convex distance and mean to pyriemann (if not done already)
+# inject cpm distance and mean to pyriemann (if not done already)
 from pyriemann_qiskit.utils import distance, mean  # noqa
 from pyriemann_qiskit.pipelines import (
     QuantumMDMVotingClassifier,
@@ -107,15 +107,15 @@
 
 pipelines = {}
 
-pipelines["mean=convex/distance=euclid"] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="mean", quantum=quantum
+pipelines["mean=logeuclid_cpm/distance=logeuclid"] = QuantumMDMWithRiemannianPipeline(
+    metric="mean", quantum=quantum
 )
 
-pipelines["mean=logeuclid/distance=convex"] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="distance", quantum=quantum
+pipelines["mean=logeuclid/distance=logeuclid_cpm"] = QuantumMDMWithRiemannianPipeline(
+    metric="distance", quantum=quantum
 )
 
-pipelines["Voting convex"] = QuantumMDMVotingClassifier(quantum=quantum)
+pipelines["Voting logeuclid_cpm"] = QuantumMDMVotingClassifier(quantum=quantum)
 
 ##############################################################################
 # Run evaluation

examples/MI/classify_alexmi_with_quantum_pipeline.py

@@ -20,7 +20,7 @@
 from moabb.evaluations import WithinSessionEvaluation
 from moabb.paradigms import MotorImagery
 
-# inject convex distance and mean to pyriemann (if not done already)
+# inject cpm distance and mean to pyriemann (if not done already)
 from pyriemann_qiskit.utils import distance, mean  # noqa
 from pyriemann_qiskit.pipelines import (
     QuantumMDMWithRiemannianPipeline,
@@ -68,8 +68,8 @@
 pipelines = {}
 
 # Will run QAOA under the hood
-pipelines["mean=logeuclid/distance=convex"] = QuantumMDMWithRiemannianPipeline(
-    convex_metric="distance", quantum=True
+pipelines["mean=logeuclid/distance=cpm"] = QuantumMDMWithRiemannianPipeline(
+    metric="distance", quantum=True
 )
 
 # Classical baseline for evaluation

pyriemann_qiskit/classification.py

@@ -21,6 +21,7 @@
 from qiskit_ibm_provider import IBMProvider, least_busy
 from qiskit_machine_learning.algorithms import QSVC, VQC, PegasosQSVC
 from qiskit_machine_learning.kernels.quantum_kernel import QuantumKernel
+from qiskit_optimization.algorithms import CobylaOptimizer
 from sklearn.base import BaseEstimator, ClassifierMixin
 from sklearn.svm import SVC
 
@@ -582,7 +583,7 @@ class QuanticMDM(QuanticClassifierBase):
 
     """Quantum-enhanced MDM classifier
 
-    This class is a convex implementation of the Minimum Distance to Mean (MDM)
+    This class is a quantic implementation of the Minimum Distance to Mean (MDM)
     [1]_, which can run with quantum optimization.
     Only log-Euclidean distance between trial and class prototypes is supported
     at the moment, but any type of metric can be used for centroid estimation.
@@ -593,11 +594,13 @@ class QuanticMDM(QuanticClassifierBase):
     .. versionchanged:: 0.1.0
         Fix: copy estimator not keeping base class parameters.
     .. versionchanged:: 0.2.0
-        Add seed parameter
+        Add seed parameter.
+        Add regularization parameter.
+        Add classical_optimizer parameter.
 
     Parameters
     ----------
-    metric : string | dict, default={"mean": 'logeuclid', "distance": 'convex'}
+    metric : string | dict, default={"mean": 'logeuclid', "distance": 'cpm'}
         The type of metric used for centroid and distance estimation.
         see `mean_covariance` for the list of supported metric.
         the metric could be a dict with two keys, `mean` and `distance` in
@@ -606,7 +609,7 @@ class QuanticMDM(QuanticClassifierBase):
         the mean in order to boost the computional speed and 'riemann' for the
         distance in order to keep the good sensitivity for the classification.
     quantum : bool (default: True)
-        Only applies if `metric` contains a convex distance or mean.
+        Only applies if `metric` contains a cpm distance or mean.
 
         - If true will run on local or remote backend
           (depending on q_account_token value),
@@ -624,6 +627,10 @@ class QuanticMDM(QuanticClassifierBase):
         Random seed for the simulation
     upper_bound : int (default: 7)
         The maximum integer value for matrix normalization.
+    regularization: MixinTransformer (defulat: None)
+        Additional post-processing to regularize means.
+    classical_optimizer : OptimizationAlgorithm
+        An instance of OptimizationAlgorithm [3]_
 
     See Also
     --------
@@ -642,26 +649,32 @@ class QuanticMDM(QuanticClassifierBase):
         A. Barachant, S. Bonnet, M. Congedo and C. Jutten. 9th International
         Conference Latent Variable Analysis and Signal Separation
         (LVA/ICA 2010), LNCS vol. 6365, 2010, p. 629-636.
+    .. [3] \
+        https://qiskit-community.github.io/qiskit-optimization/stubs/qiskit_optimization.algorithms.OptimizationAlgorithm.html#optimizationalgorithm
     """
 
     def __init__(
         self,
-        metric={"mean": "logeuclid", "distance": "convex"},
+        metric={"mean": "logeuclid", "distance": "logeuclid_cpm"},
         quantum=True,
         q_account_token=None,
         verbose=True,
         shots=1024,
         seed=None,
         upper_bound=7,
+        regularization=None,
+        classical_optimizer=CobylaOptimizer(rhobeg=2.1, rhoend=0.000001),
     ):
         QuanticClassifierBase.__init__(
             self, quantum, q_account_token, verbose, shots, None, seed
         )
         self.metric = metric
         self.upper_bound = upper_bound
+        self.regularization = regularization
+        self.classical_optimizer = classical_optimizer
 
     def _init_algo(self, n_features):
-        self._log("Convex MDM initiating algorithm")
+        self._log("Quantic MDM initiating algorithm")
         classifier = MDM(metric=self.metric)
         if self.quantum:
             self._log("Using NaiveQAOAOptimizer")
@@ -670,10 +683,17 @@ def _init_algo(self, n_features):
             )
         else:
             self._log("Using ClassicalOptimizer (COBYLA)")
-            self._optimizer = ClassicalOptimizer()
+            self._optimizer = ClassicalOptimizer(self.classical_optimizer)
         set_global_optimizer(self._optimizer)
         return classifier
 
+    def _train(self, X, y):
+        QuanticClassifierBase._train(self, X, y)
+        if self.regularization is not None:
+            self._classifier.covmeans_ = self.regularization.fit_transform(
+                self._classifier.covmeans_
+            )
+
     def predict(self, X):
         """Calculates the predictions.