University of Cambridge > Talks.cam > Cambridge Algorithms talks > A Near-Optimal Adaptive Algorithm for Estimating Binary Mixtures of Unknown Coins

A Near-Optimal Adaptive Algorithm for Estimating Binary Mixtures of Unknown Coins

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Dr Thomas Sauerwald.

Given a mixture between two populations of coins: “positive” coins that have (unknown and potentially different) probabilities of heads at least 1/2+Δ and “negative” coins with probabilities at most 1/2-Δ, we consider the task of estimating the fraction ρ of coins of each type to within additive error ε, with a constant probability of success.

In this talk, we will focus on the construction of an algorithm with sample complexity O((ρ/ε²Δ²)(1+ρ log 1/ε)), which matches the fully-adaptive lower bound of Ω(ρ/ε²Δ²) shown in our paper, except for the regime where ρ = ω(1/log 1/ε).

The fine-grained adaptive flavour of both our algorithm and lower bound contrasts with much previous work in distributional testing and learning.

Joint work with Paul Valiant, available on arXiv at https://arxiv.org/abs/1904.09228

This talk is part of the Cambridge Algorithms talks series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2024 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity