Enum ContinualLearningStrategyType

Namespace

AiDotNet.Models.Options

Assembly

AiDotNet.dll

Specifies the continual learning strategy to use for preventing catastrophic forgetting.

public enum ContinualLearningStrategyType

Fields

AveragedGEM = 6

Averaged GEM - efficient variant using single averaged constraint.

EWC = 0

Elastic Weight Consolidation - penalizes changes to important weights using Fisher information.

ExperienceReplay = 7

Experience Replay - stores and replays examples from previous tasks.

GEM = 5

Gradient Episodic Memory - constrains gradients to not hurt stored examples.

GenerativeReplay = 8

Generative Replay - uses generative model to create pseudo-examples for rehearsal.

LearningWithoutForgetting = 4

Learning without Forgetting - uses knowledge distillation to preserve old predictions.

MAS = 3

Memory Aware Synapses - unsupervised importance estimation using output sensitivity.

OnlineEWC = 1

Online EWC - memory-efficient variant that maintains running Fisher estimate.

PackNet = 9

PackNet - isolates parameters per task through pruning and freezing.

ProgressiveNeuralNetworks = 10

Progressive Neural Networks - adds new columns with lateral connections for each task.

SynapticIntelligence = 2

Synaptic Intelligence - tracks weight importance online during training.

VCL = 11

Variational Continual Learning - Bayesian approach using posterior as prior.

Remarks

For Beginners: Continual learning strategies help neural networks learn new tasks without forgetting previously learned ones. Different strategies use different approaches to balance learning new knowledge while preserving old knowledge.