A simple neural network with a fixed architecture: 3 fully connected layers, each followed by a square activation layer. More...

Public Member Functions
SimpleNeuralNet (HeContext &he)
Construct a network. More...

std::streamoff save (std::ostream &stream) const
Save network to binary stream. More...

std::streamoff load (std::istream &stream)
Load network from binary stream. More...

void initFromNet (const SimpleNeuralNetPlain &net, int baseChainIndex=-1)
Init network from a plain network. More...

void predict (const CipherMatrix &input, CipherMatrix &output) const
Run prediction. More...

Detailed Description

A simple neural network with a fixed architecture: 3 fully connected layers, each followed by a square activation layer.

It works on encrypted data: both inputs and weights, stored as CipherMatrix.

The encrypted weights of this network are either loaded from file, or encrypted from a SimpleNeuralNetPlain.

Constructor & Destructor Documentation

helayers::SimpleNeuralNet::SimpleNeuralNet ( HeContext & he )

Construct a network.

Parameters

[in] he the underlying context.

void helayers::SimpleNeuralNet::initFromNet	(	const SimpleNeuralNetPlain &	net,
		int	baseChainIndex = `-1`
	)

Init network from a plain network.

Parameters

[in]	net	network to init from.
[in]	baseChainIndex	first chain index to use (where applicable).

streamoff helayers::SimpleNeuralNet::load ( std::istream & stream )

Load network from binary stream.

Parameters

[in] stream output stream to read from

void helayers::SimpleNeuralNet::predict	(	const CipherMatrix &	input,
		CipherMatrix &	output
	)		const

Run prediction.

Parameters

[in]	input	input data
[out]	output	output prediction

streamoff helayers::SimpleNeuralNet::save ( std::ostream & stream ) const

Save network to binary stream.

Parameters

[in] stream output stream to write to

The documentation for this class was generated from the following files: