From Two-Way Transducers to Regular Function Expressions

Abstract

Transducers constitute a fundamental extension of automata. The class of regular word functions has recently emerged as an important class of word-to-word functions, characterized by means of (functional, or unambiguous, or deterministic) two-way transducers, copyless streaming string transducers, and MSO-definable graph transformations. A fundamental result in language theory is Kleene’s Theorem, relating finite state automata and regular expressions. In [3], the authors introduced a set of regular function expressions and proved a similar result for regular word functions, by showing the equivalence with copyless streaming string transducers. In this paper, we propose a direct, simplified and effective translation from unambiguous two-way transducers to regular function expressions extending the Brzozowski and McCluskey algorithm. In addition, we identify a subset of regular function expressions characterizing the (strict) subclass of functional sweeping transducers.