|
GEGELATI
|
Class to extract metrics from either the agent or the environment. More...
#include <classificationSelectionMetrics.h>
Public Member Functions | |
| ClassificationSelectionMetrics ()=default | |
| Default constructor. | |
| ClassificationSelectionMetrics (const std::vector< double > &scorePerClass, const std::vector< size_t > &nbEvalPerClass) | |
| Constructor with score and utility initialization. | |
| virtual const std::vector< double > & | getScorePerClass () const |
| virtual const std::vector< size_t > & | getNbEvalPerClassPerClass () const |
| void | initMetrics (const TPG::TPGVertex *agent, const Learn::LearningEnvironment &learningEnvironment) override |
| Specialization of the initialisation of the metrics. | |
| void | extractMetricsEpisode (const TPG::TPGVertex *agent, size_t nbStepsExecuted, const Learn::LearningEnvironment &learningEnvironment) override |
| Specialization of the extraction of the metrics at the end of an episode. | |
| virtual void | weightedSum (std::shared_ptr< SelectionMetrics > other, size_t nbEvaluation, size_t nbEvaluationOther) override |
| Specialization of weightedSum method to add the score per class and nbEvalPerClass. | |
Public Member Functions inherited from Selector::SelectionMetrics | |
| SelectionMetrics ()=default | |
| Default constructor. | |
| virtual | ~SelectionMetrics ()=default |
| Default destructor. | |
| SelectionMetrics (double score, double utility=0) | |
| Constructor with score and utility initialization. | |
| virtual double | getScore () const |
| virtual double | getUtility () const |
| virtual void | extractMetricsStep (const TPG::TPGVertex *agent, std::vector< double > actionValues, const Learn::LearningEnvironment &learningEnvironment) |
| Extract metrics from the agent in the learning environment. | |
Protected Attributes | |
| std::vector< double > | scorePerClass |
| Vector storing a double score per class (i.e. per Action) of a classification LearningEnvironment. | |
| std::vector< size_t > | nbEvalPerClass |
| Vector storing a size_t value per class representing the number of evaluation per class. | |
Protected Attributes inherited from Selector::SelectionMetrics | |
| double | score = 0 |
| double | utility = 0 |
Additional Inherited Members | |
Static Public Member Functions inherited from Selector::SelectionMetrics | |
| template<class T > | |
| static T | weightedSum (T value, T valueOther, size_t nbEvaluation, size_t nbEvaluationOther) |
| Perform a weighted sum between 2 values. | |
Class to extract metrics from either the agent or the environment.
This metrics can be used to specify the selection of the selector. This class does not implement any metrics, it need to be override by the different selection methods.
|
inline |
Constructor with score and utility initialization.
| [in] | scorePerClass | the vector of score obtained by the agent per class. |
| [in] | nbEvalPerClass | the vector of number of evaluation per class. |
|
overridevirtual |
Specialization of the extraction of the metrics at the end of an episode.
Reimplemented from Selector::SelectionMetrics.
|
virtual |
Return the number of evaluation per class of the agent.
|
virtual |
Return the score per class of the agent.
|
overridevirtual |
Specialization of the initialisation of the metrics.
Copyright or © or Copr. IETR/INSA - Rennes (2025) :
Quentin Vacher qvach.nosp@m.er@i.nosp@m.nsa-r.nosp@m.enne.nosp@m.s.fr (2025)
GEGELATI is an open-source reinforcement learning framework for training artificial intelligence based on Tangled Program Graphs (TPGs).
This software is governed by the CeCILL-C license under French law and abiding by the rules of distribution of free software. You can use, modify and/ or redistribute the software under the terms of the CeCILL-C license as circulated by CEA, CNRS and INRIA at the following URL "http://www.cecill.info".
As a counterpart to the access to the source code and rights to copy, modify and redistribute granted by the license, users are provided only with a limited warranty and the software's author, the holder of the economic rights, and the successive licensors have only limited liability.
In this respect, the user's attention is drawn to the risks associated with loading, using, modifying and/or developing or reproducing the software by the user in light of its specific status of free software, that may mean that it is complicated to manipulate, and that also therefore means that it is reserved for developers and experienced professionals having in-depth computer knowledge. Users are therefore encouraged to load and test the software's suitability as regards their requirements in conditions enabling the security of their systems and/or data to be ensured and, more generally, to use and operate it in the same conditions as regards security.
The fact that you are presently reading this means that you have had knowledge of the CeCILL-C license and that you accept its terms.
Reimplemented from Selector::SelectionMetrics.
|
overridevirtual |
Specialization of weightedSum method to add the score per class and nbEvalPerClass.
Reimplemented from Selector::SelectionMetrics.