arXiv ScienceSearch

arXiv subjects

Thomas Hanneforth

Publications and source records attributed to Thomas Hanneforth.

2 recordsLinked to original sources

Pushing for weighted tree automata

A weight normalization procedure, commonly called pushing, is introduced for weighted tree automata (wta) over commutative semifields. The normalization preserves the recognized weighted tree language even for nondeterministic wta, but it is most useful for bottom-up deterministic wta, where it can be used for minimization and equivalence testing. In both applications a careful selection of the weights to be redistributed followed by normalization allows a reduction of the general problem to the corresponding problem for bottom-up deterministic unweighted tree automata. This approach was already successfully used by Mohri and Eisner for the minimization of deterministic weighted string automata. Moreover, the new equivalence test for two wta $M$ and $M'$ runs in time $\mathcal O((\lvert M \rvert + \lvert M'\rvert) \cdot \log {(\lvert Q\rvert + \lvert Q'\rvert)})$, where $Q$ and $Q'$ are the states of $M$ and $M'$, respectively, which improves the previously best run-time $\mathcal O(\lvert M \rvert \cdot \lvert M'\rvert)$.

cs.FL

Random Generation of Nondeterministic Finite-State Tree Automata

Algorithms for (nondeterministic) finite-state tree automata (FTAs) are often tested on random FTAs, in which all internal transitions are equiprobable. The run-time results obtained in this manner are usually overly optimistic as most such generated random FTAs are trivial in the sense that the number of states of an equivalent minimal deterministic FTA is extremely small. It is demonstrated that nontrivial random FTAs are obtained only for a narrow band of transition probabilities. Moreover, an analytic analysis yields a formula to approximate the transition probability that yields the most complex random FTAs, which should be used in experiments.

cs.FL