The European Union, aware of the importance of communication between the member countries and their citizens, has decided to finance a project known as Molto, to the tune of 2.3 million euros, that will develop a tool offering translations of a high quality for a large number of language pairs on the Internet.
The University of Gothenburg (Sweden), and specifically the Department of Science and Engineering, will be responsible for managing this project. The director of the Department, Aarne Ranta, asserts that the prototypes will cover the 23 official languages of the European Union with the goal being that all citizens of the community be able to access the full base of knowledge on the Internet with out any advantage for any particular language.
This project is designed to solve some of the problems currently present in automatic translators on the Internet, which have not been able to correctly address certain issues of grammar. As Ranta notes, as of now there is no translation tool that fully covers languages; even though there are several translation tools available on the Internet, such as the highly popular Google Translator, which improves the quality of its translators over time through automatic learning and information recovery mechanisms that help it learn from its own errors, they do not have any explicit grammatical rules that help their translations reach a level that can be considered perfect.
The goal of the MOLTO Project is to develop a system of grammar rules and standards. Molto will use a grammatical framework (GF) where different languages are related through a common abstract syntax. The GF is applied in several fields on a small or medium scale, generally to work with up to ten languages; MOLTO is looking to expand this aspect for greater productivity and applicability, which will consist of increasing the number of fields and languages. One important thing will be having the technology available for experts in the field who do not have experience with the GF and thereby minimizing the effort necessary for creating the translator.
The experience will be based mostly on the interoperability between Web Ontology standards (OWL) and the GF and the extension of translations based on rules with statistical methods. Such interoperability shall allow for multilingual interaction based on natural language with knowledge that is valid for machines.
Even though, as with all occasions when we collect information on potential automatic translation tools, we cannot help but be incredulous about the real quality that this type of programs can offer, mainly for technical and grammatically complex texts since, generally, the result is more often than not lacking.
The MOLTO Project began on March 1, 2010, and shall continue until February 2013. The following timeline lays out the main milestones for the project:
June 2010: online demo at www.molto-project.eu
November 2010: infrastructure of knowledge representation.
March 2011: GF grammar compiler.
(Spanish version: https://www.trustedtranslations.com/el-proyecto-molto-2010-07-14.html)
