Unsupervised Machine Translation: How Machines Learn to Understand across Languages

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3A10511539" target="_blank" >RIV/00216208:11320/25:10511539 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Unsupervised Machine Translation: How Machines Learn to Understand across Languages

Popis výsledku v původním jazyce

For decades, machine translation between natural languages has relied in a fundamental way on so-called parallel texts, i.e. texts that have been translated by humans. For a long time, the idea that machine translation systems could be trained on non-parallel texts, i.e. texts in the source and target languages that are completely independent of each other, was rather illusory. In our monograph, we discuss the motivation for research on machine translation based on monolingual texts, organizing the methods used into two main classes: methods based on data manipulation (typically finding pairs of sentences that, although not mutually translated, could be because they carry similar content) and methods based on designing and modifying machine translation models so that the models can find linguistic equivalence between source and target language expressions on their own. The focus of the monograph is on a series of experiments with each method and combinations of methods, which demonstrate the practical

Název v anglickém jazyce

Unsupervised Machine Translation: How Machines Learn to Understand across Languages

Popis výsledku anglicky

For decades, machine translation between natural languages has relied in a fundamental way on so-called parallel texts, i.e. texts that have been translated by humans. For a long time, the idea that machine translation systems could be trained on non-parallel texts, i.e. texts in the source and target languages that are completely independent of each other, was rather illusory. In our monograph, we discuss the motivation for research on machine translation based on monolingual texts, organizing the methods used into two main classes: methods based on data manipulation (typically finding pairs of sentences that, although not mutually translated, could be because they carry similar content) and methods based on designing and modifying machine translation models so that the models can find linguistic equivalence between source and target language expressions on their own. The focus of the monograph is on a series of experiments with each method and combinations of methods, which demonstrate the practical

Druh

B - Odborná kniha

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GX19-26934X" target="_blank" >GX19-26934X: Neuronové reprezentace v multimodálním a mnohojazyčném modelování</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

ISBN

978-80-246-6084-4

Počet stran knihy

180

Název nakladatele

Karolinum Press

Místo vydání

Praha, Czechia

Kód UT WoS knihy

—