magistrsko delo
Miha Šemen (Author), Špela Vintar (Mentor)

Abstract

Številne raziskave so pokazale, da so strojni prevajalniki naredili velik korak naprej na področju literarnega prevajanja. Kar se je nekoč zdelo kot nepremostljiva ovira, se je izkazalo za cilj, h kateremu danes stremijo vsi strojni prevajalniki. Čeprav so bile izvedene že številne raziskave na področju strojnega prevajanja literarnih besedil iz angleščine v slovenščino, se še nobena ni izrecno ukvarjala s strojnim prevajanjem osebnih in zemljepisnih imen v fantazijski literaturi v tem jezikovnem paru. V tem magistrskem delu analiziramo strojne in človeške prevode osebnih in zemljepisnih imen v paru angleščina-slovenščina. Podnapise serije Igra prestolov smo najprej prevedli z nevronskima strojnima prevajalnikoma DeepL in Google Prevajalnik, nato pa preučili in primerjali uporabljene ter prepoznane prevajalske strategije ter analizirali njihovo doslednost pri prevajanju. Analiza je pokazala, da se število uporabljenih prevajalskih strategij razlikuje glede na prevajalca ter da DeepL v primerjavi z Google Prevajalnikom nudi nižjo raven doslednosti. Prav tako pa smo ugotovili, da se v večini primerov strojni prevodi razlikujejo od človeških ter da Google Prevajalnik nudi več napačnih prevodov kot DeepL.

Keywords

strojno prevajanje;nevronski prevajalniki;književno prevajanje;analiza napak;prevajanje lastnih imen;magistrska dela;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UL FF - Faculty of Arts
Publisher: [M. Šemen]
UDC: 81'322.4:81'367.622.12(043.2)
COBISS: 169699075 Link will open in a new window
Views: 38
Downloads: 9
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Comparative analysis of machine translations of personal and geographical names in the Game of thrones series
Secondary abstract: Numerous studies have shown that machine translation systems have made great progress in the field of literary translation. Once an insurmountable challenge, turned out to be the main goal for all machine translation systems strive for today. Although research has been done in the field of machine translation of literary texts from English to Slovene, none has explicitly dealt with machine translation of personal and geographical names in fantasy literature for this language pair. Thus, in this thesis, we analyse these names' machine and human translations in the English-Slovenian pair. The subtitles of the Game of Thrones series were first translated using two neural machine translation systems DeepL and Google Translate, followed by an examination and comparison of the used and identified translation strategies and the consistency of the translations. The data shows that the number of translation strategies used varies by translator and that DeepL provides less consistent translations than Google Translate. However, we also found out that in most cases machine translations differ from human translations and that Google Translate provides more mistranslations compared to DeepL.
Secondary keywords: machine translation;neural translation systems;literary translation;error analysis;translation of named entities;master's theses;
Type (COBISS): Master's thesis/paper
Study programme: 0
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univerza v Ljubljani, Filozofska fak., Odd. za prevajalstvo
Pages: 71 f.
ID: 20253349