[−][src]Trait believer::channel::BinaryChannel

pub trait BinaryChannel: Sync {
    fn intrinsic_likelyhood(&self, output: GF2) -> f64;

    fn send(&self, input: GF2) -> GF2;

    fn message_likelyhood(&self, output: &[GF2]) -> Vec<f64> { ... }

    fn sample(&self, inputs: &[GF2]) -> Vec<GF2> { ... }

    fn sample_uniform(&self, input: GF2, n_inputs: usize) -> Vec<GF2> { ... }
}

A trait that represent a binary transmission channel. It takes an GF2 element and (randomly) map it to an other GF2 element. It has an intrinsic likelyhood for each output.

Example

// Create a bsc with error prob of 0.2.
let bsc = channel::BinarySymmetricChannel::new(0.2);
// Sample the channel by always sending 0.
let received = bsc.sample_uniform(GF2::B0, 1000);
let number_of_one = received.iter()
    .filter(|&x| x == &GF2::B1)
    .collect::<Vec<_>>()
    .len();
println!("{}", number_of_one); // Should be around 200.

Required methods

fn intrinsic_likelyhood(&self, output: GF2) -> f64

For a given output, compute log(p(output|input = 1) / p(output|input = 0)).

Example

let bsc = channel::BinarySymmetricChannel::new(0.2);
assert_eq!(bsc.intrinsic_likelyhood(GF2::B0), -2.0);
assert_eq!(bsc.intrinsic_likelyhood(GF2::B1), 2.0);

fn send(&self, input: GF2) -> GF2

For a given input, returns an output according to some probability distribution depending on the channel.

Loading content...

Provided methods

fn message_likelyhood(&self, output: &[GF2]) -> Vec<f64>

fn sample(&self, inputs: &[GF2]) -> Vec<GF2>

Returns a Vec of outputs as long as the inputs where each output is sample using the send method.

fn sample_uniform(&self, input: GF2, n_inputs: usize) -> Vec<GF2>

Returns a Vec of outputs where the input is send n_inputs times.

Loading content...

Implementors

impl BinaryChannel for BinarySymmetricChannel[src]

Loading content...