VerticalFederatedBenchmark<T>

Models & Types · AiDotNet.FederatedLearning.Vertical

Provides benchmarking utilities for evaluating VFL implementations.

Inheritance: FederatedLearningComponentBase<T> → VerticalFederatedBenchmark<T>

For Beginners

When developing VFL systems, it’s important to test with controlled datasets to verify that the system works correctly and to measure performance. This class generates synthetic benchmark datasets and runs standardized evaluation suites.

How It Works

Benchmark scenarios:

Reference: VertiBench (ICLR 2024): “Feature distribution diversity matters for fair and accurate VFL evaluation.”

Constructors

Constructor Summary
VerticalFederatedBenchmark()

Methods

Method Summary
VflBenchmarkDataset<T> GenerateDataset(int totalEntities, int totalFeatures, int numberOfParties, double overlapRatio, int? seed) Generates a synthetic vertically-partitioned dataset for benchmarking.
VflBenchmarkResult RunBenchmark(VerticalFederatedLearningOptions options, VflBenchmarkDataset<T> dataset) Runs a standardized benchmark suite on a VFL implementation.