All for One": A Hybrid Translation Synthesis Framework
Authors: Malavazos, Christos 
Triantafyllou, Ioannis 
Antonopoulos, Vassilios 
Karagiannis, Georgios 
Piperidis, Stelios 
Issue Date: 1-Jan-2004
Journal: WSEAS Transactions on Acoustics and Music 
Volume: 1
Issue: 1
Keywords: Machine Translation/MT, Example Based Machine Translation/EBMT, Analogical Modelling, Translation by Analogy
Abstract: 
This paper introduces a dynamic approach in enhancing the effectiveness of a typical translation memory system towards providing a more flexible translation framework. The system core consists of three basic modules: a) a self-modelling, incremental learning module for extracting translation rules from existing parallel corpora, b) a statistical module for the dynamic extraction of translation units from the same corpora and c) a translation module that efficiently utilizes various levels of available information for dealing with new input. Even though all modules do not rely on heavy linguistic processing, they are designed in such a way as to allow for additional information to be easily incorporated and improve system performance.
URI: https://uniwacris.uniwa.gr/handle/3000/478
Type: Article
Department: Department of Archival, Library and Information Studies 
School: School of Administrative, Economics and Social Sciences 
Affiliation: University of West Attica (UNIWA) 
Appears in Collections:Articles / Άρθρα

CORE Recommender
Show full item record

Page view(s)

34
checked on Nov 5, 2024

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.