We study the problem of reducing the size of a language model while preserving recognition performance (accuracy and speed). A successful approach has been to represent language models by weighted nite-state automata (WFAs). Analogues of classical automata determinization and minimization algorithms then provide a general method to produce smaller but equivalent WFAs. We extend this approach by...

In this paper, we propose a fixed point theory to solve games of imperfect information. The fixed point theory is defined on the lattice of antichains of sets of states. Contrary to the classical solution proposed by Reif [Rei84], our new solution does not involve determinization. As a consequence, it is readily applicable to classes of systems that do not admit determinization. Notable example...

This paper revisits the notions of observer and diagnoser, and adapts them to probabilistic automata, in a setting of weighted automata computations. In the non stochastic case, observers and diagnosers are obtained by standard elementary steps, as state augmentation, epsilon-reduction and determinization. It is shown that these steps can be adapted to probabilistic automata, and algorithms to ...

In this paper we propose a new composition algorithm for weighted finite-states transducers that are more and more used for speech and pattern recognition applications. Composition joins multiple transducers into one. We have implemented an embedded speech based dialog system for steering applications. Therefore regular grammars are very useful, but they may enlarge strongly by determinization....

In this contribution, we reconsider the notion of determinism for max-plus automata by including quantitative aspect in its definition. A determinization procedure adapted to so-called weight-determinism is proposed. The motivation generalize results based on automata.

The synthesis of reactive systems requires the solution of two-player games on graphs with ω-regular objectives. When the objective is specified by a linear temporal logic formula or nondeterministic Büchi automaton, then previous algorithms for solving the game require the construction of an equivalent deterministic automaton. However, determinization for automata on infinite words is extremel...

A central result in the theory of ω-automata is McNaughton's Theorem which states that every Buchi automaton is equivalent to a deterministic Muller automaton. The complexity of McNaughton's algorithm is huge. Safra provided an optimal construction. Safra's algorithm converts a Buchi automaton with n states to a deterministic Rabin automaton with n states. Much e orts have been invested to redu...

We introduce a family of MDP reduced models characterized by two parameters: the maximum number of primary outcomes per action that are fully accounted for and the maximum number of occurrences of the remaining exceptional outcomes that are planned for in advance. Reduced models can be solved much faster using heuristic search algorithms such as LAO*, benefiting from the dramatic reduction in t...