In today’s society, machine translation systems are taking over and gradually reshaping the way people communicate across different languages. But how precisely are these tools capable of generating publication-ready contents? To what extent can they adapt to the target audiences’ cultures and conventions? The main goal of this dissertation has been to investigate these aspects while also demonstrating how a corpus-based translation approach can successfully address recurring issues found in neural machine translation outputs. The following dissertation includes three key chapters. Firstly, chapter 1 provides a theoretical framework for machine translation, tracing its evolution up to contemporary AI-based systems, while simultaneously introducing corpora and their wide-ranging applications. Chapter 2 outlines the methodological approach, specifically highlighting the most widespread practices adopted to analyze corpora within the field of translation studies, while also emphasizing the pivotal role of evaluating machine translation output to ensure the continued advancement of the field. Finally, chapter 3 applies the theoretical and methodological insights from the preceding chapters by focusing on an authentic example of machine-translated tourism text, rendered from Italian into English. The conclusions drawn from this study suggest that the ideal of “pure machine translation” is yet to be achieved, highlighting instead the strategic role of corpora in the post-editing phase.
Il presente elaborato si propone di dimostrare in che misura gli strumenti di traduzione automatica neurale siano in grado di generare contenuti corretti, comunicativamente efficaci e conformi alle convenzioni linguistiche e culturali di un pubblico internazionale. Parallelamente, si intende verificare quanto un approccio traduttivo basato sui corpora possa rivelarsi una strategia metodologicamente solida per colmare lacune ancora presenti nei moderni sistemi di traduzione neurale. L’elaborato si compone di tre capitoli principali. Il primo è di natura teorica concettuale e fornisce un quadro generale sull’evoluzione della “machine translation”, fino ai più recenti sistemi di intelligenza artificiale. Parallelamente vengono introdotti i corpora e le relative applicazioni nei diversi ambiti di utilizzo. Il secondo capitolo assume invece una prospettiva metodologica, descrivendo la fenomenologia dei corpora, le principali pratiche di interrogazione di tali risorse, e conclude con una breve panoramica sull’importanza della valutazione dei contenuti generati da tools automatici. Il terzo capitolo si prefigge infine di applicare a livello pratico gli assunti teorici e metodologici dei capitoli precedenti, esaminando il caso concreto di una traduzione automatica nella combinazione italiano-inglese di un testo turistico locale. L’analisi degli output di Google Traduttore e altri strumenti analoghi, ha evidenziato quanto l’ideale di “Pure Machine Translation” non sia ancora una realtà concreta, sottolineando al contempo il ruolo essenziale dei corpora nel processo di post-editing.
L’approccio traduttivo corpus-based: una risposta efficace ai limiti di Google Traduttore.
GHIELMI, MARCO
2024/2025
Abstract
In today’s society, machine translation systems are taking over and gradually reshaping the way people communicate across different languages. But how precisely are these tools capable of generating publication-ready contents? To what extent can they adapt to the target audiences’ cultures and conventions? The main goal of this dissertation has been to investigate these aspects while also demonstrating how a corpus-based translation approach can successfully address recurring issues found in neural machine translation outputs. The following dissertation includes three key chapters. Firstly, chapter 1 provides a theoretical framework for machine translation, tracing its evolution up to contemporary AI-based systems, while simultaneously introducing corpora and their wide-ranging applications. Chapter 2 outlines the methodological approach, specifically highlighting the most widespread practices adopted to analyze corpora within the field of translation studies, while also emphasizing the pivotal role of evaluating machine translation output to ensure the continued advancement of the field. Finally, chapter 3 applies the theoretical and methodological insights from the preceding chapters by focusing on an authentic example of machine-translated tourism text, rendered from Italian into English. The conclusions drawn from this study suggest that the ideal of “pure machine translation” is yet to be achieved, highlighting instead the strategic role of corpora in the post-editing phase.| File | Dimensione | Formato | |
|---|---|---|---|
|
Ghielmi.Marco.pdf
accesso aperto
Dimensione
2.71 MB
Formato
Adobe PDF
|
2.71 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14251/3939