Limit-average automata are not PAC-learnable passively (NP-complete to fit samples) but can be learned almost-exactly in polynomial time via active queries under uniform distribution.
An efficient membership-query algorithm for learning dnf with respect to the uniform distribution
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.FL 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Approximate Learning of Limit-Average Automata
Limit-average automata are not PAC-learnable passively (NP-complete to fit samples) but can be learned almost-exactly in polynomial time via active queries under uniform distribution.