Collaborative Work?ow for Crowdsourcing Translation Vamshi Ambati, Stephan Vogel, Jaime Carbonell Language Technologies Institute, Carnegie Mellon University 5000 Forbes Avenue, Pittsburgh, PA 15213 {vamshi,vogel,jgc }***@ ACM Classi?cation Keywords Impacts: Computer-supported collaborative work] General Terms Design; Human Factors. Author Keywords Language translation, Collaborative work?ow, Crowdsourcing, Amazon Mechanical Turk ABSTRACT In this paper, we explore the challenges involved in crowd- sourcing the task of translation over the web, where remotely located translators work on providing translations indepen- dent of each other. We then, propose a collaborative work?ow for crowdsourcing translation to address some of these chal- lenges. In our pipeline model, the translators are working in phases where output from earlier phases can be enhanced in the subsequent phases. We also highlight some of the novel contributions of the pipeline model like assistive translation and translation synthesis that can leverage monolingual and bilingual speakers alike. We evaluate our approach by elicit- ing translations for both a minority-to-majority language-pair and a minority-to-minority language-pair. We observe that in both scenarios, our work?ow produces better quality transla- tions in a cost-effective manner, pared to the tradi- tional crowdsourcing work?
