Noisy population recovery in polynomial time
This solves a computational learning theory problem for distributions with bounded support, representing a significant improvement over prior work.
The paper tackles the noisy population recovery problem by improving the sample and algorithmic complexity from poly(k^(log log k), n, 1/ε) to poly(k, n, 1/ε) for μ > 0, achieving polynomial bounds in k, n, and 1/ε.
In the noisy population recovery problem of Dvir et al., the goal is to learn an unknown distribution $f$ on binary strings of length $n$ from noisy samples. For some parameter $μ\in [0,1]$, a noisy sample is generated by flipping each coordinate of a sample from $f$ independently with probability $(1-μ)/2$. We assume an upper bound $k$ on the size of the support of the distribution, and the goal is to estimate the probability of any string to within some given error $\varepsilon$. It is known that the algorithmic complexity and sample complexity of this problem are polynomially related to each other. We show that for $μ> 0$, the sample complexity (and hence the algorithmic complexity) is bounded by a polynomial in $k$, $n$ and $1/\varepsilon$ improving upon the previous best result of $\mathsf{poly}(k^{\log\log k},n,1/\varepsilon)$ due to Lovett and Zhang. Our proof combines ideas from Lovett and Zhang with a \emph{noise attenuated} version of Möbius inversion. In turn, the latter crucially uses the construction of \emph{robust local inverse} due to Moitra and Saks.