scientific article; zbMATH DE number 4043288
[For the entire collection see Zbl 0624.00016.] In connection with the design and implementation of an algorithm that generates a formal translation system by analyzing a set of only a few examples of the underlying translation process the new model of a segment translation system has been developed, which is very suitable for formalizing translation processes of comparatively low complexity (e.g. braille, phonology, hyphenation). Each segment translation system consists of an ordered set of rules \(u\to v[x,y]\) with the meaning: the segment u of a word w should be translated into the segment v if u has x as left context and y as right context in w. After a precise definition of the model of a segment translation system and its application as a translation algorithm the paper shows, that the power of this model is strictly below that of a finite transducer. Furthermore it turns out that the equivalence problem for the subclass of character translation systems is decidable, this question being still open for the classes of segment translation systems and deterministic finite transducers, respectively.
- Finitely generated relations and their application to the translation problem in one class of languages
- A mathematical model for translation of natural languages
- The many faces of a translation
- The many faces of a translation
- scientific article; zbMATH DE number 977102 (Why is no real title available?)
- Rule-restricted automaton-grammar transducers: power and linguistic applications
- Grid model of text translation and structure of data base for Chuvash-Russian translation
- Unit completion for a computer-aided translation typing system
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3780475)