We present a new algorithm to construct a deterministic Rabin automaton for an LTL formula φ. The automaton is the product of a master automaton and an array of slave automata, one for each Gsubformula of φ. The slave automaton for Gψ is in charge of recognizing whether FGψ holds. As opposed to standard determinization procedures, the states of all our automata have a clear logical structure, w...

The goal of this work was to develop an algorithm for the integration of the lexicon with the language model which would be computationally efficient in terms of memory requirements, even in the case of large trigram models. Two specialized versions of the algorithm for transducer composition were implemented. The first one is basically a composition algorithm that uses the precomputed set of t...

We show that the neighbourhood of a regular language L with respect to an additive quasi-distance can be recognized by an additive weighted finite automaton (WFA). The size of the WFA is the same as the size of an NFA (nondeterministic finite automaton) for L and the construction gives an upper bound for the state complexity of a neighbourhood of a regular language with respect to a quasi-dista...

We propose to use the knowledge that an ω-regular property is stutter insensitive to construct potentially smaller deterministic ω-automata for such a property, e.g. using Safra’s determinization construction. This knowledge allows us to skip states that are redundant under stuttering, which can reduce the size of the generated automaton. In order to use this technique even for automata that ar...

We study the class of relations implemented by nested word to word transducers (also known as visibly pushdown transducers). We show that any such relation can be uniformized by a functional relation from the same class, implemented by an unambiguous transducer. We give an exponential upper bound on the state complexity of the uniformization, improving a previous doubly exponential upper bound....

Processing Mehryar Mohri, Fernando Pereira and Michael Riley AT&T Research 600 Mountain Avenue Murray Hill, 07974 NJ fmohri,pereira,[email protected] Abstract. Finite-state automata are a very e ective tool in natural language processing. However, in a variety of applications and especially in speech precessing, it is necessary to consider more general machines in which arcs are assigned ...

ABSTRACT Speech recognition requires solving many space and time problems that can have a critical effect on the overall system performance. We describe the use of two general new algorithms [5] that transform recognition networks into equivalent ones that require much less time and space in large-vocabulary speech recognition. The new algorithms generalize classical automata determinization an...

Finite-state machines have been used in various domains of natural language processing. We consider here the use of a type of transducers that supports very efficient programs: sequential transducers. We recall classical theorems and give new ones characterizing sequential string-tostring transducers. Transducers that output weights also play an important role in language and speech processing....

Abstract: A general way to find repetitions of factors in a given text is shown. We start with a classification of repetitions. The general models for finding exact repetitions in one string and in a finite set of strings are introduced. It is shown that d-subsets created during determinization of nondeterministic factor automata contain all information concerning repetitions of factors. The pr...

We show that regarding finite automata (FA) as discrete, time-invariant linear systems over semimodules, allows to: (1) express FA minimization and FA determinization as particular observability and reachability transformations of FA, respectively; (2) express FA pumping as a property of the FA’s reachability matrix; (3) derive canonical forms for FAs. These results are to our knowledge new, an...