Bernoulli factory
A probabilistic algorithmic framework that can exactly simulate one Bernoulli distribution while only being able to sample from a related distribution. It is commonly used to convert decisions into accept/reject decisions in rejection sampling.
Timeline (last 90 days)
SIRS uses the Bernoulli factory proposed by Chen et al. (2026) to convert Poisson-distributed clipped estimates into per-token accept/reject decisions.
A Bernoulli factory is an algorithmic framework that converts random samples from one probability distribution into samples from another, and can construct a Bernoulli distribution with exact parameter p using only black-box queries to p.
All Facts (2)
SIRS uses the Bernoulli factory proposed by Chen et al. (2026) to convert Poisson-distributed clipped estimates into per-token accept/reject decisions.
50%UnverifiedA Bernoulli factory is an algorithmic framework that converts random samples from one probability distribution into samples from another, and can construct a Bernoulli distribution with exact parameter p using only black-box queries to p.
50%