44
1
v1v2 (latest)

Reconstructing decision trees

Abstract

We give the first {\sl reconstruction algorithm} for decision trees: given queries to a function ff that is opt\mathrm{opt}-close to a size-ss decision tree, our algorithm provides query access to a decision tree TT where: \circ TT has size S=sO((logs)2/ε3)S = s^{O((\log s)^2/\varepsilon^3)}; \circ dist(f,T)O(opt)+ε\mathrm{dist}(f,T)\le O(\mathrm{opt})+\varepsilon; \circ Every query to TT is answered with poly((logs)/ε)logn\mathrm{poly}((\log s)/\varepsilon)\cdot \log n queries to ff and in poly((logs)/ε)nlogn\mathrm{poly}((\log s)/\varepsilon)\cdot n\log n time. This yields a {\sl tolerant tester} that distinguishes functions that are close to size-ss decision trees from those that are far from size-SS decision trees. The polylogarithmic dependence on ss in the efficiency of our tester is exponentially smaller than that of existing testers. Since decision tree complexity is well known to be related to numerous other boolean function properties, our results also provide a new algorithms for reconstructing and testing these properties.

View on arXiv
Comments on this paper