MHRN-Space / src /learning /__init__.py
ThomasHeisig's picture
Sync GitHub main af692537c27175c8410bf57a747be0caf23718e2 (part 8)
9d98328 verified
Raw History Blame Contribute Delete
2.35 kB
"""Learning and plasticity for MHRN.
This package provides:
- Pair-based STDP (isolated and production variants)
- Reward-modulated plasticity
- Independent prediction-error modulation
- Eligibility traces
- Learning engine for network integration
- Guarded learning-preparation contracts
"""
from .contract import (
LEARNING_CONTRACT_ID,
LEARNING_CONTRACT_STATUS,
LEARNING_CPU_REFERENCE,
LEARNING_CUDA_STATUS,
LearningContractState,
learning_contract_check,
learning_contract_descriptor,
learning_contract_hash,
learning_contract_status,
)
from .eligibility import EligibilityTrace, create_eligibility_trace
from .learning_engine import LearningEngine, LearningParameters, LearningStats
from .prediction_error import (
PredictionErrorPlasticity,
PredictionErrorPlasticityConfig,
PredictionErrorPlasticityError,
PredictionErrorPlasticityStats,
PredictionErrorSignal,
)
from .preparation import (
LearningDataPartition,
LearningObjective,
LearningPlanOrigin,
LearningPreparationGuard,
LearningPreparationProposal,
LearningPreparationService,
LearningSourceRef,
PreparedLearningPlan,
)
from .reward import RewardSignal, create_reward
from .stdp_plugin import STDPParameters, STDPSynapse, create_stdp_synapse
__all__ = [
"LEARNING_CONTRACT_ID",
"LEARNING_CONTRACT_STATUS",
"LEARNING_CPU_REFERENCE",
"LEARNING_CUDA_STATUS",
"LearningContractState",
"learning_contract_check",
"learning_contract_descriptor",
"learning_contract_hash",
"learning_contract_status",
# STDP
"STDPParameters",
"STDPSynapse",
"create_stdp_synapse",
# Reward
"RewardSignal",
"create_reward",
# Prediction Error
"PredictionErrorPlasticity",
"PredictionErrorPlasticityConfig",
"PredictionErrorPlasticityError",
"PredictionErrorPlasticityStats",
"PredictionErrorSignal",
# Learning Engine
"LearningEngine",
"LearningParameters",
"LearningStats",
# Eligibility
"EligibilityTrace",
"create_eligibility_trace",
# Learning Preparation
"LearningDataPartition",
"LearningObjective",
"LearningPlanOrigin",
"LearningPreparationGuard",
"LearningPreparationProposal",
"LearningPreparationService",
"LearningSourceRef",
"PreparedLearningPlan",
]