Online learning via dynamic reranking for Computer Assisted Translation

Abstract

New techniques for online adaptation in computer assisted translation are explored and compared to previously existing approaches. Under the online adaptation paradigm, the translation system needs to adapt itself to real-world changing scenarios, where training and tuning may only take place once, when the system is set-up for the first time. For this purpose, post-edit information, as described by a given quality measure, is used as valuable feedback within a dynamic reranking algorithm. Two possible approaches are presented and evaluated. The first one deals with the translation probabilities within the translation model, whereas the second one modifies the weights of the log-linear model dynamically. Our experimental results show that such algorithms are able to improve translation quality by learning from the errors produced by the system on a sentence-by-sentence basis.