Active Learning for Interactive Neural Machine Translation of Data Streams

Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018), pages 151–160 Brussels, Belgium, October 31 - November 1, 2018. c⃝2018 Association for Computational Linguistics 151 Active Learning for Interactive Neural Machine Translation of Data Streams ´Alvaro Peris and Francisco Casacuberta Pattern Recognition and Human Language Technology Research Center Universitat Polit`ecnica de Val`encia, Val`encia, Spain {lvapeab, fcn}@prhlt.upv.es Abstract We study the application of active learning techniques to the translation of unbounded data streams via interactive neural machine translation. The main idea is to select, from an unbounded stream of source sentences, those worth to be supervised by a human agent. The user will interactively translate those samples. Once validated, these data is useful for adapt- ing the neural machine translation model. We propose two novel methods for selecting the samples to be validated. We exploit the information from the attention mechanism of a neural machine translation system. Our ex- periments show that the inclusion of active learning techniques into this pipeline allows to reduce the effort required during the pro- cess, while increasing the quality of the trans- lation system. Moreover, it enables to balance the human effort required for achieving a cer- tain translation quality. Moreover, our neural system outperforms classical approaches by a large margin. 1 Introduction The translation industry is a high-demand field. Large amounts of data must be translated on a reg- ular basis. Machine translation (MT) techniques greatly boost the productivity of the translation agencies (Arenas, 2008). However, despite the re- cent advances achieved in this field, MT systems are still far to be perfect and make errors. The correction of such errors is usually done in a post- processing step, called post-editing. This requires a great effort, as it needs from expert human su- pervisors. The requirements of the translation industry have increased in the last years. We live in a global world, in which large amounts of data must be pe- riodically translated. This is the case of the Euro- pean Parliament, whose proceedings must be reg- ularly translated; or the Project Syndicate1 plat- form, which translates editorials from newspapers to several languages. In these scenarios, the sen- tences to be translated can be seen as unbounded streams of data (Levenberg et al., 2010). When dealing with such massive volumes of data, it is prohibitively expensive to manually re- vise all the translations. Therefore, it is manda- tory to spare human effort, at the expense of some translation quality. Hence, when facing this sit- uation, we have a twofold objective: on the one hand, we aim to obtain translations with the high- est quality possible. On the other hand, we are constrained by the amount of human effort spent in the supervision and correction process of the translations proposed by an MT system. The active learning (AL) framework is well- suited for these objectives. The application of AL techniques to MT involve to ask a human oracle to supervise a fraction of the incoming data (Blood- good and Callison-Burch, 2010). Once …

Content truncated to 500 words