![]() “The major challenge is really, how do we take the translation systems we have, and then actually meet the demand of people around the world, Angela Fan, a research associate at Facebook AI, told Engadget. That’s why Facebook AI has developed a new MT model that can bidirectionally translate directly between two languages (Chinese to French and French to Chinese) without ever using English as a crutch - and which outperforms the English-centric model by 10 points on BLEU metrics. This is done because data sets of translations to and from English are massive and widely available but putting English in the middle reduces the overall translation accuracy while making the entire process more complex and cumbersome than it needs to be. However these systems typically use English as an intermediary step - that is, translating from Chinese to French actually goes Chinese to English to French. In fact, Facebook provides around 20 billion translations everyday for its News Feed alone. Finally, the last chapter summarized the whole study, detailed the findings and ended with recommendations.Whether you’re logging on from the US, Brazil, Borneo, or France, Facebook can translate virtually any written content published on its platform into the local language using automated machine translation. It explains the errors made with reference to the SMT/RBMT theoretical considerations embedded in the major theoretical framework (artificial intelligence) in order to confirm or disconfirm the hypotheses stated. lexical, semantic and syntactic as well as combinations of these errors.Ĭhapter four deals with a comparative analysis of the translation results from the two paradigms. The errors made were equally analyzed at different linguistic levels i.e. Chapter three presents the same data used in chapter two that was collected from grammar books, aligned with its corresponding machine translation from Yahoo! Babelfish Translate (RBMn. lexical, semantic and syntactic as well as combinations of these errors. The errors made were analyzed at different linguistic levels i.e. It also establishes the objectives, hypotheses as well as the methodological framework of the research.Ĭhapter two presents the data (embedded clauses) that was collected from grammar books aligned with its corresponding machine translation from Google Translate (SMT). The chapter states, limits and defines the problem that has necessitated the present research. It was also revealed that RBMT is more of a literal translator compared to SMT which is more of a free translator.Ĭhapter one laid the groundwork of the study by anchoring it into the artificial intelligence framework that was intertwined with statistical machine translation (SMT) and rule-based machine translation (RBMT) paradigms. On the overall, RBMT was found to be more accurate than SMT with its strength lying in the syntactic aspect attributed to the grammatical rules applied in the process of translation. that> which, does not affect the translation and that the fog index is not the main source of errors in machine translation. The study revealed too that changing the complementizer which retain the meaning in the clause i.e. It also established that omitting the complementizer could be a source of errors in SMT. ![]() The analysis of the results led to the confirmation of the major hypothesis: machine translation cannot translate embedded clauses accurately since almost 70% of the clauses were mistranslated by each of the MT tools. ![]() The main purpose of this study was to investigate whether machine translation (with specific reference to Google Translate and Yahoo! Babelfish Translate) is able to translate embedded clauses from English to French accurately and also to establish whether statistical machine translation (SMT) renders a more accurate translation compared to rule-based machine translation (RBMT).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |