ISecretSharingScheme<T>

Interfaces · AiDotNet.FederatedLearning.MPC

Defines the contract for a secret sharing scheme that splits and recombines tensor values.

Known implementations: ArithmeticSecretSharing<T>

For Beginners

Secret sharing lets you split a secret number into multiple “shares” that individually look random. The original secret can only be recovered when enough shares are combined. This is the foundation of MPC.

How It Works

Types of secret sharing:

Extends: This generic interface works with tensors and complements the existing ShamirSecretSharing class (which operates on raw byte arrays for crypto-level operations).

Properties

Property Summary
int ReconstructionThreshold { get; } Gets the minimum number of shares required to reconstruct the secret.
string SchemeName { get; } Gets the name of this secret sharing scheme.

Methods

Method Summary
Tensor<T> Combine(Tensor<T>[] shares) Reconstructs the original tensor from a set of shares.
Tensor<T>[] Split(Tensor<T> secret, int numberOfParties) Splits a tensor value into secret shares for the specified number of parties.