Speech processing requires very efficient methods and algorithms. Finite-state transducers have been shown recently both to constitute a very useful abstract model and to lead to highly efficient time and space algorithms in this field. We present these methods and algorithms and illustrate them in the case of speech recognition. In addition to classical techniques, we describe many new algorit...

In this paper we address the synthesis problem for specifications given in linear temporal single-agent epistemic logic, KLTL (or KL1), over single-agent systems having imperfect information of the environment state. [18] have shown that this problem is 2Exptime complete. However, their procedure relies on complex automata constructions that are notoriously resistant to efficient implementation...

All the components used in the search stage of speech recognition systems – language model, pronunciation dictionary, context-dependent network, HMM model – can be represented by finite-state labeled networks. To construct real-time recognition systems, it is important to optimize these networks and to efficiently combine them. We present new methods that substantially improve these steps. We s...

Communication protocols can be formally described by the Communicating Finite-State Machines (CFSM) model. This model is expressive, but not expressive enough to deal with complex protocols that involve structured messages encapsulating integers or lists of integers. This is the reason why we propose an extension of this model : the Symbolic Communicating Machines (SCM). We also propose an appr...

Computationally feasible representations of models of rst-order formulas are of interest in various elds like e.g. in automated model building, automated deduction, deductive databases, etc. For representing a Herbrand model symbolically we only need to represent the true ground atoms of the model. Many mechanisms for representing such (term) sets appear as candidates for model representations....

We present a new disambiguation algorithm for finite automata and functional finite-state transducers. We give a full description of the algorithm, including a detailed pseudocode and analysis, and several illustrating examples. Our algorithm is often more efficient and the result dramatically smaller than the one obtained using determinization for finite automata or an existing disambiguation ...

Guarded strings are like ordinary strings over a finite alphabet P, except that atoms of the free Boolean algebra on a set of atomic tests B alternate with the symbols of P. The regular sets of guarded strings play the same role in Kleene algebra with tests as the regular sets of ordinary strings do in Kleene algebra. In this paper we develop the elementary theory of finite automata on guarded ...

Regular cost functions have been introduced recently as an extension of the notion of regular languages with counting capabilities, which retains strong closure, equivalence, and decidability properties. Cost functions are functions ranging over N∪{∞}, and considered modulo an equivalence which preserves the existence of bounds over each subset of the domain. A central result in the theory of r...

This paper describes a non-conventional method for compiling (phonological or morpho-syntactic) context restriction (CR) constraints into non-deterministic automata in finite-state tools and surface parsing systems. The method reduces any CR into a simple one that constraints the occurrences of the empty string and represents right contexts with co-determististic states. In cases where a fully ...

Recently a new method directly translating linear temporal logic (LTL) formulas to deterministic (generalized) Rabin automata was proposed by Javier Esparza and Jan Kret́ınský [1]. Compared to the existing approaches of constructing a non-deterministic Buechi-automaton in the first step and then applying a determinization procedure (e.g. some variant of Safra’s construction) in a second step, th...