User-Friendly Text Prediction For Translators

User-Friendly Text Prediction for Translators George Foster and Philippe Langlais and Guy Lapalme RALI, Universit´e de Montr´eal {foster,felipe,lapalme}@iro.umontreal.ca Abstract Text prediction is a form of interactive machine translation that is well suited to skilled translators. In principle it can as- sist in the production of a target text with minimal disruption to a translator’s nor- mal routine. However, recent evaluations of a prototype prediction system showed that it significantly decreased the produc- tivity of most translators who used it. In this paper, we analyze the reasons for this and propose a solution which consists in seeking predictions that maximize the ex- pected benefit to the translator, rather than just trying to anticipate some amount of upcoming text. Using a model of a “typ- ical translator” constructed from data col- lected in the evaluations of the prediction prototype, we show that this approach has the potential to turn text prediction into a help rather than a hindrance to a translator. 1 Introduction The idea of using text prediction as a tool for trans- lators was first introduced by Church and Hovy as one of many possible applications for “crummy” machine translation technology (Church and Hovy, 1993). Text prediction can be seen as a form of in- teractive MT that is well suited to skilled transla- tors. Compared to the traditional form of IMT based on Kay’s original work (Kay, 1973)—in which the user’s role is to help disambiguate the source text— prediction is less obtrusive and more natural, allow- ing the translator to focus on and directly control the contents of the target text. Predictions can benefit a translator in several ways: by accelerating typing, by suggesting translations, and by serving as an im- plicit check against errors. The first implementation of a predictive tool for translators was described in (Foster et al., 1997), in the form of a simple word-completion system based on statistical models. Various enhancements to this were carried out as part of the TransType project (Langlais et al., 2000), including the addition of a re- alistic user interface, better models, and the capabil- ity of predicting multi-word lexical units. In the fi- nal TransType prototype for English to French trans- lation, the translator is presented with a short pop- up menu of predictions after each character typed. These may be incorporated into the text with a spe- cial command or rejected by continuing to type nor- mally. Although TransType is capable of correctly antic- ipating over 70% of the characters in a freely-typed translation (within the domain of its training cor- pus), this does not mean that users can translate in 70% less time when using the tool. In fact, in a trial with skilled translators, the users’ rate of text pro- duction declined by an average of 17% as a result of using TransType (Langlais et al., 2002). There are two main reasons for this. First, it takes time to read the system’s proposals, so that in cases where they are wrong or too short, the …

Content truncated to 500 words