Personalized Machine Translation: Preserving Original Author Traits

Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers, pages 1074–1084, Valencia, Spain, April 3-7, 2017. c⃝2017 Association for Computational Linguistics Personalized Machine Translation: Preserving Original Author Traits Ella Rabinovich Shachar Mirkin Raj Nath Patel Department of Computer Science IBM Research - Haifa C-DAC Mumbai University of Haifa, Israel Mount Carmel, Haifa Gulmohar Cross Road No. 9, Juhu & IBM Research - Haifa 31905, Israel Mumbai-400049, India ellarabi@gmail.com shacharm@il.ibm.com patelrajnath@gmail.com Lucia Specia Shuly Wintner Department of Computer Science Department of Computer Science University of Sheffield, United Kingdom University of Haifa, Israel l.specia@sheffield.ac.uk shuly@cs.haifa.ac.il Abstract The language that we produce reflects our personality, and various personal and de- mographic characteristics can be detected in natural language texts. We focus on one particular personal trait of the author, gender, and study how it is manifested in original texts and in translations. We show that author’s gender has a powerful, clear signal in originals texts, but this signal is obfuscated in human and machine trans- lation. We then propose simple domain- adaptation techniques that help retain the original gender traits in the translation, without harming the quality of the trans- lation, thereby creating more personalized machine translation systems. 1 Introduction Among many factors that mold the makeup of a text, gender and other authorial traits play a major role in our perception of the content we face. Many studies have shown that these traits can be identified by means of automatic classifi- cation methods. Classical examples include gen- der identification (Koppel et al., 2002), and au- thorship attribution and profiling (Seroussi et al., 2014). Most research, however, addressed texts in a single language, typically English. We investigate a related but different question: we are interested in understanding what happens to personality and demographic textual markers during the translation process. It is generally agreed that good translation goes beyond transfor- mation of the original content, by preserving more subtle and implicit characteristics inferred by au- thor’s personality, as well as era, geography, and various cultural and sociological aspects. In this work we explore whether translations preserve the stylistic characteristic of the author and, further- more, whether the prominent signals of the source are retained in the target language. As a first step, we focus on gender as a demo- graphic trait (partially due to the absence of par- allel data annotated for other traits). We evalu- ate the accuracy of automatic gender classification on original texts, on their manual translations and on their automatic translations generated through statistical machine translation (SMT). We show that while gender has a strong signal in originals, this signal is obfuscated in human and machine translation. Surprisingly, determining gender over manual translation is even harder than over SMT; this may be an artifact of the translation process itself or the human translators involved in it. Mirkin et al. (2015) were the first to show that authorial gender signals tend to vanish through both manual and automatic translation, using a …

Content truncated to 500 words