#Apresentação
#Sumário
#Ficha bibliográfica
#Ficha técnica
#Comentários
#Pesquisa
Universidade do Porto Faculdade de Letras da Universidade do Porto Serviços de Documentação e Informação
#
 #Menu Inicial#Publicações#Outras publicações/coleções #A linguística em diálogo : volume comemorativo dos 40 anos do Centro de Linguística da Universidade do Porto
# #
#Sumário
A linguística em diálogo
#Systèmes de traduction automatique et levée d’ambiguïté : étude comparée de systèmes de TABR, TAS et TAN
Françoise Bacquelaine
#Documento (.pdf)    
#Abstract
Machine Translation (MT) has been a lively field of research ever since the invention of the computer. Rule-Based Machine Translation (RBMT) was the first option back in the 1940s–1950s; Statistical Machine Translation (SMT) appeared a few decades later and Neural Machine Translation (NMT) in the 21st century. This distinction is not strict since most MT systems are now hybrid, but natural language ambiguity is a well-known pitfall, be it in human or machine translation. Two types of ambiguity can arise when using the rather common English word issue: “grammatical ambiguity” (noun or verb?), on the one hand, and “homographic and polysemic ambiguity (one word form with different senses in the source language)” (Hutchins 2005: 17), on the other hand. The scope of this research is limited to three senses of the noun issue (1. An important topic or problem for debate or discussion; 2. The action of supplying or distributing an item for use, sale, or official; 3. (formal or law) Children of one’s own) and two senses of the verb to issue (1. [WITH OBJECT] Supply or distribute (something) for use or sale; 2. [NO OBJECT] (issue from) Come, go, or flow out from). A sample of sentences containing at least one example of usage was selected from the British National Corpus in order to test and compare four English-French MT systems: SYSTRAN (free online RBMT), Google Translate (free online SMT), MT@EC (restricted access SMT) and free online Neural Machine Translation by LISA (University of Montreal). The outputs were compared to a human translation model based on translation memories (parallel corpora) in order to evaluate weaknesses and strengths of each system, compare the results and find out possible ways of improving MT output through hybridisation. Research results in cases like these are not just useful for theoretical linguistics but can also be used to heighten awareness in human translators and demonstrate that translators who are trained in computational linguistics can also work together with experts in artificial intelligence and machine translation.
#Voltar
Data da última atualização: 2021-02-24

DestaForma, Design e Multimédia
Biblioteca Central © 2006 - Todos os direitos reservados - FLUP