IClientContributionEvaluator<T>
Interfaces · AiDotNet.FederatedLearning.Fairness
Evaluates how much each client contributed to the federated global model.
Known implementations: DataShapleyEvaluator<T>, LightweightShapleyEvaluator<T>, PrototypicalContributionEvaluator<T>, ShapleyValueEvaluator<T>
For Beginners
In a group project, some students do more work than others. This interface measures each client’s “fair share” of the global model improvement. High-contribution clients provided valuable data, while low-contribution clients may be free-riding (benefiting without contributing). This is essential for fair compensation and for detecting poisoned or low-quality updates.
How It Works
Usage:
Properties
| Property | Summary |
|---|---|
string MethodName { get; } |
Gets the name of this contribution evaluation method. |
Methods
| Method | Summary |
|---|---|
Dictionary<int, double> EvaluateContributions(Dictionary<int, Tensor<T>> clientModels, Tensor<T> globalModel, Dictionary<int, List<Tensor<T>>> clientHistories) |
Evaluates the contribution of each client to the global model. |
HashSet<int> IdentifyFreeRiders(Dictionary<int, double> contributionScores) |
Identifies free-rider clients whose contribution falls below the threshold. |