IDiversityStrategy<T, TInput, TOutput>
Interfaces · AiDotNet.ActiveLearning.Interfaces
Interface for diversity-based sampling strategies in active learning.
Known implementations: CoreSetStrategy<T, TInput, TOutput>, DiversityStrategyBase<T, TInput, TOutput>
For Beginners
Diversity strategies select samples that are diverse or representative of the data distribution. This helps ensure the model learns from a variety of examples rather than similar ones.
How It Works
Common Diversity Strategies:
Properties
| Property | Summary |
|---|---|
string Name { get; } |
Gets the name of the diversity strategy. |
Methods
| Method | Summary |
|---|---|
Matrix<T> ComputeDistanceMatrix(IDataset<T, TInput, TOutput> dataset) |
Computes pairwise distances between samples. |
Vector<T> ComputeDiversityScores(IDataset<T, TInput, TOutput> unlabeledData, IDataset<T, TInput, TOutput> labeledData) |
Computes diversity scores for samples in the unlabeled pool. |
Vector<T> GetFeatureRepresentation(TInput input) |
Gets the feature representation for a sample. |
int[] SelectDiverseSamples(IDataset<T, TInput, TOutput> unlabeledData, IDataset<T, TInput, TOutput> labeledData, int numSamples) |
Selects diverse samples from the unlabeled pool. |