Total: 1
Hypothesis selection asks for a distribution close to an unknown $P$, given samples from $P$ and access to $n$ candidate hypotheses. We study the computational complexity of achieving statistically optimal sample complexity and approximation constants.